Ivy Bridge can fix it quite easily. If they drop the power usage of the quad down to say 30W, they can get a Radeon GPU in there.
They are all hyper-threaded now so dual-cores perform very well in multi-core tasks.
All machines become obsolete too fast. If you bought a quad-core Mac Pro in 2009, it's already being outperformed by the quad-core Mini. That's 2 years for the Pro to drop to what you'd call 'entry-level'.
If you look at any of the generic parts used in the 2009 model (say everything but the case and logic board design) none of it was high end hardware even if it was priced that way. It used the xeon version of the i7 920 which was one of their volume processors. The basic graphics card option went for around $100 on the pc side if I recall correctly. There wasn't anything special about it really. There's just nothing that is terribly amazing about the engineering or hardware of the machine.
Regarding the mini. If it got usb3 (already on PCs in that price range), a second thunderbolt port, and decent graphics (intel graphics have always sucked, remember how Apple used to prefer nvidia chipsets before intel sued them?), we could have an ok machine. I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop, I'd use it as one. Considering the lack of internal storage bays, a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth. If you look at what I care about personally, it's a system without severe bottlenecks. I tend to prefer the ability for PCI expansion, and I think it would help widen the appeal, but if it had more than a single thunderbolt port it would be less of an issue for me personally.
Right now Apple's retreat is kind of awkward. The mac pro has kind of an uncertain future, and they haven't shown it much love, but some of the video editors and stuff don't really have the option of even looking at a mini if they need huge disk IO bandwidth or realtime playback. You can actually get pci expansion even on cheap PCs for this kind of thing, but then you wouldn't be able to run OSX. Some accessories of this sort have started to debut in thunderbolt versions which somewhat solves this issue (they're still extremely expensive but they could come down in time). The problem is if you're limited to one for display and data, you can saturate it too easily. 120Hz displays are becoming more common as are SSDs. With changes like that and newer panel generations coming in at higher resolutions, thunderbolt will become incredibly tight on bandwidth.
You must understand it's frustrating at times. The lower end is missing features. You get to the imac and performance picks up a little, gpu is still kind of weak and it's not exactly cheap, then go up a bit more in price and go down in performance as it's older and cheaper hardware at its core (aside from the gpu if you order the upgrade). It's not a very balanced line in general. Most of us just choose a compromise and suck it up until we need a new computer again. As for the imac screen, it's not a very good one. It's ok but not great. I wouldn't ever buy the cinema display and it has nothing to do with the price. As of right now Apple doesn't make a single computer that's truly engineered for performance under $3500 before ram, storage upgrades, etc. The rest are engineered to be compact or portable and look pretty. This year they haven't had anything where I wanted to accept the compromise. Hopefully they will next year.
Edit: by the way, the mini has never been one of their top sellers. I think if they did turn it into a slightly more complete/robust system it might go somewhere. Right now it lacks the portability of a laptop and the performance of a desktop. It also lacks an easily added display as the cinema display is too costly for such a machine. Overall it's just in an awkward spot.
Ivy Bridge can fix it quite easily. If they drop the power usage of the quad down to say 30W, they can get a Radeon GPU in there.
Really it can't if Apples intentions are to purposefully limit the Mini's capability relative to the rest of the line up. Again it isn't a question of the value of a low end machine, Apple needs one, it is the problem of the step up being so limited. Limited to the extent that it won't have a long functional life.
Quote:
They are all hyper-threaded now so dual-cores perform very well in multi-core tasks.
Sometimes yes most of the time no.
Quote:
All machines become obsolete too fast. If you bought a quad-core Mac Pro in 2009, it's already being outperformed by the quad-core Mini. That's 2 years for the Pro to drop to what you'd call 'entry-level'.
The hardware can become outdated too fast but in Apples case they make sure of that with the Mini. In any event I think you are still missing the point, it is possible these days to buy hardware that will remain viable for much longer than would have been the case in the past. It requires that the hardware have a few features to help keep it viable for a few years.
Quote:
Everything it seems. You want lots of internal storage, fast processors, fast GPU, more RAM slots, PCI expansion slots - you basically want a cheaper Mac Pro and if they put one on sale, you'd still want an extra power supply.
Again I'm apparently not communicating well. That list represented an aggregate of features an XMac would need to have and no it is not a Mac Pro. In reality this is a mid range machine and is commonly represented by any number of PC systems as desktop PCs. One would have to have their head awfully deep in the sand not to recognize that simply implementing desktop parts gives you most of this for free. Apples job is simply to build a box.
As to the extra power supply, that has a lot to do with Apple trying to market the current hardware as servers. Not everybody needs such a capability, but if a box is to be marketed as a server it really ought to support the feature.
Quote:
You can't get it all and still have it affordable because it has to sell in volume and the things that matter to you don't matter to the vast majority of people.
That is BS. Especially in this economic climate, people are far more careful with their spending. You imply that it would be expensive but that flies in the face of what motherboards and the various chips on board cost. Seriously if Apple can't produce a desktop machine in the $1000 to $1500 range then they need to hire some new engineers. In that price range they should be able to deliver a decent box that will sell well. That box would include a processor in the 45-65 watt range, a GPU, RAM and some bays/slots for storage. This sort of machine could easily be a quarter of the size of the Mac Pro.
Quote:
The 27" iMac is a better value proposition because you get a 27" IPS screen included worth $1000.
The iMac isn't even in the running here and in fact is a terrible value.
Quote:
Performance doubles every 2 years so about 50% every year. Right now, the quad is 30-50% faster than the dual-core so Ivy Bridge dual-cores will match the Sandy Bridge quad i7 and the Radeon 7000 series should double in performance over the current models.
It would be nice if any of the above was true, but sadly it isn't. For a single app the value of a quad processor is software dependent, it is very possible for an app to use all of those cores. But a single app seldom defines computer usage these days, those cores can be effectively used when multiple apos are running.
You seem to be extremely optimistic with respect to Ivy Bridge and its performance. I really don't think it is justified. More cores is a fundamental advantage.
Quote:
Basically, I expect the Ivy Bridge version of the middle Mac Mini to match the current $2500 Mac Pro in CPU and GPU.
Well that is very optimistic. However we already know that the Ivy Bridge GPU only gets about a 30% increase in performance, to bump the discreet GPU you really need to bump the Video RAM also. In the end, even with Ivy Bridges thermals, you still are trying to stuff more into the box than is practical.
In any event an Ivy Bridge Mini, with just two to four cores won't come close to the performance of a Mac Pro. This should be obvious as the Pro has far more cores available to it.
Quote:
With USB 3 and the Thunderbolt port, that should be enough for expansion.
USB 3 and Thunderbolt combined will be a significant I/O improvement for the Mini. However that means nothing in the context of desktop machines. TB is totally unsuitable for one of the most common uses for slots in industry, that is the addition of additional communications ports. Here Ethernet is the most common port added to a machine though there are many others. Of course there are a host of other cards that need something better than TB.
Beyond that Apple currently only implements a single TB port, due to the limited bandwidth that is a serious problem.
Quote:
I think it's the way to go and while I expect the desire for a mid-range tower to continue on ad nauseam, one day it will stop.
Not likely. In fact I see demand growing. Plus you use the term mid-range tower which frankly I don't remember anybody here asking for. They could put the XMac in a pizza box for all I care though I believe there are far better form factors.
Quote:
If it takes 5 years or even 10 years, one day in the near future, a machine will arrive on the low-end that will put an end to it.
Well in ten years just about anything could happen including a nuclear war that would destroy all of our "I" devices, Macs and what have you. The problem today is solving real issues that users have. The quest for an XMac doesn't go away because Apple does have this gaping hole in their line up. It does little good to try to promote the Mini as a fill in for that hole as it will always be limited in capability.
I suspect the problem here is that when you hear about XMac or other stated desires, you immediately start thinking of a PC tower. A tower that size wise isn't much different than a Mac Pro. I really don't think that is the goal most of us have. Rather we want modern technology applied to the problem to build a platform that takes us into the future. That means dropping legacy devices and breaching new ground. You might of noticed that I often talk about secondary storage in terms of bays/slots, that is because a modern form factor needs to acknowledge that solid state storage is nothing more than another printed circuit card. As such economics should drive such storage on to low cost plug in cards.
In a nut shell the desktop is ripe for the same sort of innovation that Apple is famous for elsewhere. Frankly Apple, in tandem with Intel, is the only manufacture capable of driving such innovation right now. HP is screwed up, Dell is well Dell and there are few others that could even try.
I say that because they need to get TB into the platform. However it could simply be a transitional machine.
This is all true but you know the so called Pros will make a big stink.
I'm not sure how you get $200 it if a $20 drive but I agree with the rest. For the same reasons I'd like to see the 3.5" disk slots leave also.
Smaller is good. However I think you need to reconsider the issue of expansion. Expansion is the whole point of a Pro machine other wise people would buy iMacs or Minis. Here I'm talking expansion in all the various ways be it RAM, PCI Express slots, "storage bays" or what have you. Without the ability to significantly out do the other machines with expansion there is little reason for the Mac Pro to exist. Simply putting a bunch of CPU power in a box does little good if it isn't supported properly.
The MXM card can go as those have little draw when it comes to Pros. You did not mention RAM expansion which is a valued part of the current Pro.
Ivy Bridge holds a lot of promise for Laptops. I'm doing everything I can to resist buying before that hardware shows up. However I'm not so certain such chips will make sense in a Mac Pro. I'm actually hoping that AMD has really good luck with Bulldozer, especially bulldozer cores integrated into Fusion chips. Bulldozer is so different that Apple will likely have to do a bunch of testing but I see great potential for common workloads on Mac OS. Of course this means Apple would need TB hardware that would work with AMD systems but I'm under the impression that will not be a problem.
Sounds about right but this could happen earlier, again mainly to support the TB initiative.
Maybe there is something that I missed, but Ivy Bridge should make sense for most any platform. The new gate process and reduced power consumption alone should be worthwhile. I don't recall the specifics about the on board L3 cache specs of Ivy Bridge, but the trend is to increase L3 cache because it has such a worthwhile effect upon performance. With all these cores, even the faster bus speeds to the RAM are limiting, although dual channel RAM and lots of RAM appears to be working well.
There are some PCIe based SSD systems which, because of price, are presently limited to the server market for the most part, but they illustrate the need for faster systems to feed data to all those cores.
For example, the 15 inch MacBook Pro top-of-the-line model uses the 2.3 GHz CPU rather than the 2.2 GHz CPU. That is not much of a difference in clock speed for a $250 premium, but it has 8 MB L3 cache instead of 6 MB L3 cache of the 2.2 GHz unit. L3 cache can make a difference in overall performance in applications such as Photoshop, Final Cut Pro or any of the various video applications.
The Mac Pro community has not been getting much love from Apple for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server. When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
If you look at any of the generic parts used in the 2009 model (say everything but the case and logic board design) none of it was high end hardware even if it was priced that way. It used the xeon version of the i7 920 which was one of their volume processors. The basic graphics card option went for around $100 on the pc side if I recall correctly. There wasn't anything special about it really. There's just nothing that is terribly amazing about the engineering or hardware of the machine.
I suspect the problem with the Mac Pro is that it was or is unprofitable for Apple. Thus they turned towards the cheaper end ofIntels hardware line up and raised prices. Unfortunately this is a common mistake in industry as it results in a downward spiral in sales. The Pro effectively becomes harder to justify, due to the cost issues. Interestingly I suspect this biases sales towards the high end configurations. After all the people buying Mac Pros are now more likely to need the power of the high end configurations anyways.
Quote:
Regarding the mini. If it got usb3 (already on PCs in that price range), a second thunderbolt port, and decent graphics (intel graphics have always sucked, remember how Apple used to prefer nvidia chipsets before intel sued them?), we could have an ok machine.
I don't really see it that way. The Mini is a very OK machine if your needs are within it's limits. Unfortunately those limits are pretty tight and inflexible. So if we acknowledge that there is a hole in Apples line up, is it smart to try to fill it with something that never was designed to fill that hole. Sort of like the proverbial round peg in the square hole, or does Apple make a peg to fit.
Quote:
I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop, I'd use it as one. Considering the lack of internal storage bays, a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth. If you look at what I care about personally, it's a system without severe bottlenecks. I tend to prefer the ability for PCI expansion, and I think it would help widen the appeal, but if it had more than a single thunderbolt port it would be less of an issue for me personally.
I must confess to needing to read up on TB more but can data really impact video bandwidth? There is some confusion in the documents I've seen so far but I was under the impression that TB would dedicate one channel to video data. That may be a mistake on my part because other documents seem to imply multiplexing of data with video.
One day I need to sit down and develop a firmer grasp of TB. It is just very hard for me to get excited about the interface.
Quote:
Right now Apple's retreat is kind of awkward. The mac pro has kind of an uncertain future, and they haven't shown it much love, but some of the video editors and stuff don't really have the option of even looking at a mini if they need huge disk IO bandwidth or realtime playback. You can actually get pci expansion even on cheap PCs for this kind of thing, but then you wouldn't be able to run OSX.
I'm not sure where this idea that expansion has to imply expensive came from. At the same time people promote TB even though it will always be relatively expensive. I just don't get it.
Quote:
Some accessories of this sort have started to debut in thunderbolt versions which somewhat solves this issue (they're still extremely expensive but they could come down in time). The problem is if you're limited to one for display and data, you can saturate it too easily. 120Hz displays are becoming more common as are SSDs. With changes like that and newer panel generations coming in at higher resolutions, thunderbolt will become incredibly tight on bandwidth.
A serious issue! Having just one TB port on a Mini does not significantly alleviate it's built in limitations. In the end the Mini is still a low end solution. How they will address this on the Mac Pros replacement is an open question.
Note I said Mac Pros replacement. I really believe that the next Pro will be vastly different
Quote:
You must understand it's frustrating at times. The lower end is missing features. You get to the imac and performance picks up a little, gpu is still kind of weak and it's not exactly cheap, then go up a bit more in price and go down in performance as it's older and cheaper hardware at its core (aside from the gpu if you order the upgrade).
It isn't a universal drop in performance, it does require that software can use all of those cores though.
Quote:
It's not a very balanced line in general. Most of us just choose a compromise and suck it up until we need a new computer again. As for the imac screen, it's not a very good one. It's ok but not great. I wouldn't ever buy the cinema display and it has nothing to do with the price. As of right now Apple doesn't make a single computer that's truly engineered for performance under $3500 before ram, storage upgrades, etc. The rest are engineered to be compact or portable and look pretty. This year they haven't had anything where I wanted to accept the compromise. Hopefully they will next year.
Heck I'd be happy with midrange performance. It is a rather pathetic reality but one does get a better deal out of Apples laptops if performance is one significant measure.
Quote:
Edit: by the way, the mini has never been one of their top sellers. I think if they did turn it into a slightly more complete/robust system it might go somewhere.
I'm not sure the above is true. It often sneaks onto Apples top sellers list and is very popular with the Internet sellers. Given that Apple hardly promotes the machine as they seem to be obsessed with the notebook market.
Quote:
Right now it lacks the portability of a laptop and the performance of a desktop. It also lacks an easily added display as the cinema display is too costly for such a machine. Overall it's just in an awkward spot.
Again I don't believe the Mini is all that bad, intact it has many positives going for it. It just isn't the machine to invest in as a workstation. Off the top of my head good points with respect to the Mini:
Low power as in watts. This makes the Mini very usable in a number of use cases where a more conventional machine would not work.
As configured it is a very good home theater PC.
Compact is good where space is limited.
It is a very good to excellent network connected node.
It isn't an iMac with a built in display.
The current revision is somewhat more serviceable. In any event it is a hands down winner over the iMac.
It is well constructed and with SSD one might in a stretch call it rugged. Maybe semi rugged is a better phrase.
The problem for me and I suspect you, is that it really comes up short as a workstation. Maybe others don't agree with our desires but I don't see them as irrational, nor expensive to implement.
One more thing (I used that phrase long before I heard Steve say it) even the ipad and iphone have dual core processors these days. Quad is basically the norm for desktop computing. It's not exactly a high end feature. I'm trying to think of something more to add here to make the discussion interesting, but it seems we're just at odds on a lot of these points.
So did Columbo. "Oh, just one more thing" and then he sprung the trap.
My problem is that there seems to be excessive expectations with respect to performance. For example the GPU has gone through limited public benchmarking and is not all that great of an improvement. Something like 30% to 60% in a few specific cases. Actual CPU performance increases seem to be unclear on a clock by clock basis.
Quote:
Originally Posted by RBR
Maybe there is something that I missed, but Ivy Bridge should make sense for most any platform. The new gate process and reduced power consumption alone should be worthwhile. I don't recall the specifics about the on board L3 cache specs of Ivy Bridge, but the trend is to increase L3 cache because it has such a worthwhile effect upon performance. With all these cores, even the faster bus speeds to the RAM are limiting, although dual channel RAM and lots of RAM appears to be working well.
This is correct but there is one problem. If that Ivy Bridge installation uses the on board GPU then graphics puts a lot of pressure on the cache. AMD has tried to address this in some novel ways so I'm sure Intel has at least studied it.
Note to there is a positive aspect to the GPU using the cache, it greatly enhances communications between the CPU and the integrated GPU.
Quote:
There are some PCIe based SSD systems which, because of price, are presently limited to the server market for the most part, but they illustrate the need for faster systems to feed data to all those cores.
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
Quote:
For example, the 15 inch MacBook Pro top-of-the-line model uses the 2.3 GHz CPU rather than the 2.2 GHz CPU. That is not much of a difference in clock speed for a $250 premium, but it has 8 MB L3 cache instead of 6 MB L3 cache of the 2.2 GHz unit. L3 cache can make a difference in overall performance in applications such as Photoshop, Final Cut Pro or any of the various video applications.
There is no doubt what so ever that more cache can help significantly with many workloads. However at some point you need to up core clock rate or add cores to get speed boosts. What will be interesting with Ivy Bridge is just what the maximum clock rates will be for various power levels. I still suspect we will see a wide spread in performance values from AIR suitable chips to laptop chips to desktop chips. Thus the need for that mid-range desktop machine.
Quote:
The Mac Pro community has not been getting much love from Apple
I'd have to say the entire desktop line is getting ignored. If you compare it to the love the laptops get you would have to wonder how they get any desktop sales at all.
Quote:
for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server.
The server market is slow to change so getting traction in that market with a Mini is tough. In some regards I see these sorts of ultra small machines as the wave of the future in the data center. It probably won't be the Mini that makes the big dent there though. Most likely it will be an ARM based system that is maybe 1/6 the width of a rack based server.
Quote:
When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
My problem is that there seems to be excessive expectations with respect to performance. For example the GPU has gone through limited public benchmarking and is not all that great of an improvement. Something like 30% to 60% in a few specific cases. Actual CPU performance increases seem to be unclear on a clock by clock basis.
This is correct but there is one problem. If that Ivy Bridge installation uses the on board GPU then graphics puts a lot of pressure on the cache. AMD has tried to address this in some novel ways so I'm sure Intel has at least studied it.
Note to there is a positive aspect to the GPU using the cache, it greatly enhances communications between the CPU and the integrated GPU.
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
There is no doubt what so ever that more cache can help significantly with many workloads. However at some point you need to up core clock rate or add cores to get speed boosts. What will be interesting with Ivy Bridge is just what the maximum clock rates will be for various power levels. I still suspect we will see a wide spread in performance values from AIR suitable chips to laptop chips to desktop chips. Thus the need for that mid-range desktop machine.
I'd have to say the entire desktop line is getting ignored. If you compare it to the love the laptops get you would have to wonder how they get any desktop sales at all.
The server market is slow to change so getting traction in that market with a Mini is tough. In some regards I see these sorts of ultra small machines as the wave of the future in the data center. It probably won't be the Mini that makes the big dent there though. Most likely it will be an ARM based system that is maybe 1/6 the width of a rack based server.
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
The laptop market has been an increasingly larger segment of all PC sales, not just Apple's, and so I understand devoting a lot of resources to it...personally I wish all of the manufacturers would make matte/non-glare screens available on all models..., but there is still a substantial market for desktops. Apple abandoned the XServe because it represented such a small portion of their Mac sales (note that Mac sales are only something in the neighborhood of 25% of Apple revenue), but it got the proverbial foot in the door at major edu sites and at least some businesses (usually smaller ones).
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
There were many good ideas in the XServe. Some of them were ridiculously simple, but practical, such as the flashing light on the unit with the drive problem so that the system admin knew which one in the bank of XServes to go to.
Graphics are very important in some scientific applications. I am aware of Mac Pros being used for medical research visualizing drug and cell interactions at the molecular level which takes some "grunt" to accomplish in a useful time frame. Those people are doing important work with some OS X apps and some UNIX apps. Sooner than later they will have to look for other solutions if Apple does not devote the resources to keep current with these things.
The laptop market has been an increasingly larger segment of all PC sales, not just Apple's, and so I understand devoting a lot of resources to it...personally I wish all of the manufacturers would make matte/non-glare screens available on all models..., but there is still a substantial market for desktops. Apple abandoned the XServe because it represented such a small portion of their Mac sales (note that Mac sales are only something in the neighborhood of 25% of Apple revenue), but it got the proverbial foot in the door at major edu sites and at least some businesses (usually smaller ones).
My problem is this, if Apple doesn't do something about Mac Pros soon they will go the way of the XServe. On top of that Apple has a lot of desktop sales they could steal from the PC industry. I realize the laptop market is strong but the desktop market is not dead and shouldn't be ignored.
Quote:
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
Apple buying Sun? I think that would have been ugly.
Quote:
There were many good ideas in the XServe. Some of them were ridiculously simple, but practical, such as the flashing light on the unit with the drive problem so that the system admin knew which one in the bank of XServes to go to.
They where good servers but also very expensive for what you got. Apple screwed themselves just as they have with the Mac Pro. Most people don't mind paying a little extra for their Macs but bolt at highway robbery. In a data center this is even more important as no one cares about the manufacture it is all about cost.
In any event those server related features are cheap to implement and could easily be added to a desktop Mac.
Quote:
Graphics are very important in some scientific applications. I am aware of Mac Pros being used for medical research visualizing drug and cell interactions at the molecular level which takes some "grunt" to accomplish in a useful time frame. Those people are doing important work with some OS X apps and some UNIX apps. Sooner than later they will have to look for other solutions if Apple does not devote the resources to keep current with these things.
Cheers
Well those people buying such hardware are generally better in formed than the mainstream. The thing is if you max out a Mac Pro it really isn't that bad of a machine. After all how many desktops are built to implement multi core dual socket Xeon's? The hardware to upgrade the Pro will come soon. So I don't see it as being neglected in the context of the high end customer. However Apple does have to make that upgrade compelling.
The Mac Pro community has not been getting much love from Apple for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server. When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
Yeah and since it's Apple they don't truly acknowledge their direction for the line due to their culture of secrecy. Even when new processors come out, it's not a guarantee of updates anytime soon. Recall how long it took just to get the weak westmere refresh?
Quote:
Originally Posted by wizard69
I suspect the problem with the Mac Pro is that it was or is unprofitable for Apple. Thus they turned towards the cheaper end ofIntels hardware line up and raised prices. Unfortunately this is a common mistake in industry as it results in a downward spiral in sales. The Pro effectively becomes harder to justify, due to the cost issues. Interestingly I suspect this biases sales towards the high end configurations. After all the people buying Mac Pros are now more likely to need the power of the high end configurations anyways.
The higher end configurations are still priced high relative to similar configurations PC side and not necessarily much better (really Apple's engineering time seems to go into other lines). It's just that at that point the difference generally isn't completely excessive. At that level the machine will be expensive no matter where you buy it. My issue was that the configurations are terribly awkward. In 2009 when they went to four slots for ram in the single socket machine, 4GB dimms were still quite costly. This meant unless you were willing to spend significantly more these machines were limited to 8GB of ram, which is okay for many people but not everyone. That's just one example, but in 2010 the configurations became even more awkward. They don't balance clock speed to core count very well in several cases. For the cost of the starting eight core the processor choice could have been a better one in both 2009 and 2010, as they resulted in an overall slower machine when you weren't stressing all available cores.
Quote:
Originally Posted by RBR
So did Columbo. "Oh, just one more thing" and then he sprung the trap.
Columbo quote = automatically awesome
Quote:
Originally Posted by wizard69
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
Your posts are fun to read. You were the first that even brought up PCIe SSD storage. I'm wondering what the preferred form will be for storage enclosures and external drives. You mentioned the cost. Wouldn't those allow for a simpler manufacturing process? I realize card prices are all over the place depending on type and other factors.
Quote:
Originally Posted by wizard69
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
There was that rumor about Apple developing a rackmountable mac pro, but the only available "photo" just looked like something photoshopped. Remember back in June or so? If they're trying to get in the server market it could make sense. I've been saying for a while that they need to do something with it to prevent its following the Xserve's fate. If they can come up with a good strategy for a new machine, they obviously have the funds to develop it. I'd really like them to develop a successor to that line that's actually functional in the sense that it remains up to date and is priced in a way that isn't completely insane (I've made points on this before). No one wants a 2009 design in late 2011 for the same price. It shouldn't surprise Apple either.
Quote:
Originally Posted by RBR
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
Cheers
Hahahaha ahhhh that would have been so bad. They don't have an exceptional reputation in managing independent product lines after acquisition when it comes to things designed for commercial usage. Look at what they did to Shake after they purchased that. I've noticed everyone has a different idea of what Apple should buy with their massive cash pile, but many of the suggestions don't seem synergetic with Apple the way they are today.
Quote:
Originally Posted by wizard69
My problem is this, if Apple doesn't do something about Mac Pros soon they will go the way of the XServe. On top of that Apple has a lot of desktop sales they could steal from the PC industry. I realize the laptop market is strong but the desktop market is not dead and shouldn't be ignored.
Well those people buying such hardware are generally better in formed than the mainstream. The thing is if you max out a Mac Pro it really isn't that bad of a machine. After all how many desktops are built to implement multi core dual socket Xeon's? The hardware to upgrade the Pro will come soon. So I don't see it as being neglected in the context of the high end customer. However Apple does have to make that upgrade compelling.
There are a few workstation manufacturers like that. I agree with you on the parallel to the Xserve. I've been saying that for some time. Letting it sit really doesn't inspire confidence. You've probably noticed but a lot of people still own either the original mac pro or the 2008 configuration. There's not a lot of compelling reason to update from the 2008 model unless your needs demand one of the top machines which is kind of weird three years later. The 2006 one also aged decently for how old it is. Sure it's slower than the new ones by a wide margin, but I would suggest that it probably held out better than these newer ones will over the next five years.
I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop
The Mini functions just fine as a desktop. The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
Quote:
Originally Posted by hmm
a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
You can run a Pegasus RAID drive and a display from the one Thunderbolt port and stick another display on the HDMI port.
Quote:
Originally Posted by wizard69
Really it can't if Apples intentions are to purposefully limit the Mini's capability relative to the rest of the line up.
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
Quote:
Originally Posted by wizard69
Apples job is simply to build a box.
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
Quote:
Originally Posted by wizard69
This sort of machine could easily be a quarter of the size of the Mac Pro.
Not with PCI slots.
Quote:
Originally Posted by wizard69
In any event an Ivy Bridge Mini, with just two to four cores won't come close to the performance of a Mac Pro. This should be obvious as the Pro has far more cores available to it.
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
Quote:
Originally Posted by wizard69
Here Ethernet is the most common port added to a machine though there are many others.
No reason you can't get a USB 3 adaptor for this and way easier to install.
Quote:
Originally Posted by wizard69
you use the term mid-range tower which frankly I don't remember anybody here asking for. They could put the XMac in a pizza box
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
The Mini functions just fine as a desktop. The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
I will try to think of a test for this. As I've stated my opinion on the mac pro was that it was a downgrade of the 2008 machine from the processor aspect. The 2010 one had a decent gpu option though. Now when you do run longer tasks on it, that mac pro really isn't very fast, and it's a horrendous value. The gpu if you take the cto is about the only nice thing in it. The hard drive bays aren't even well designed in terms of cooling unless they fixed that on the 2010 one. The 2009 it was still really uneven.
Quote:
Originally Posted by Marvin
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth. It seems pretty high considering the peak bandwidth of displayport 1.2 is roughly 18Gb/s total. I'd like to know what kind of specs they were getting for throughput + what the display resolution was, and if this setup supports ddc based calibration as many people who would purchase such a display need that. On mine it goes through a shielded usb cable but not many operate that way.
Quote:
Originally Posted by Marvin
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it. I also remember how long Apple retained the Core2s in one of their laptops to avoid intel chipsets. I don't think that'll happen again, but it would be easy for them to go the mac pro route on this. Raise the price again (they've adjusted mini price points before) and downgrade their relative choices of processor grades for the machine. Mac pro used to start with mid range dual socket versions. Now they have the cheapest budget model. This is Apple. They've done this kind of thing before because they figure that people will grumble and buy it anyway.
Quote:
Originally Posted by Marvin
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
I'd like to comment on this. The mac pro is one of the larger towers of its type. This design was originally to accommodate the G5 processors with their massive heatsinks and the accompanying logic board design. In 2011 it looks like a dinosaur. Regarding aluminum it's not that expensive of a material. They aren't using a terribly high grade here so it doesn't contribute as much to the cost as you anticipate. In fact much of the cost of aluminum is often from processing it. The macbook airs and pros use a cnc process which should in theory be more labor intensive than the mac pro case. Beyond that they've provided machines using this form factor at $1500 and $2000 during the G5 era into the first generation of the mac pro. Right now the top macbook pro processor available to the 15" in a macbook pro is $568 retail. That machine is still cheaper than a mac pro with the inclusion of the thunderbolt chip. The 27" imac has the cost of a 27" ips panel (not the most expensive one of all but none of the 27" ones are dirt cheap) and uses a processor around $300.
That mac pro you mentioned is artificially inflated. Doing anything to turn it into a better volume machine including a redesigned form factor and up to date setup would encourage volume and potentially a much better deal on the machine. Remember what Apple charged for the original macbook air? It was grossly overpriced relative to its capability, and ended up with dismal sales. Also what do you really gain from single socket Xeons anyway? Going with normal desktop i7s that are updated more frequently would be better for a single socket machine.
Quote:
Originally Posted by Marvin
Not with PCI slots.
It's already been suggested that there are newer more compact methods of implementation for this stuff. It's just a matter of what Apple wants to use. Really a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work (I'm serious, if anyone questions it I'll bring up links to companies, pictures of capture carts, etc). Some use laptops for lighter jobs but they suck. A lighter form factor here would be amazing for these kinds of guys. Rackmountable configurations could be good if there are any Xserves still in use. I realize it wouldn't be the most efficient 2U kind of setup, but it could add more market potential to such a machine.
Quote:
Originally Posted by Marvin
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
It's quite a difference from the mac mini with a 45W processor, but the quad core mini isn't the same thing. Like I said it's not even close to the imac processor which is around 90W. By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
I think we need to get away from the idea that Apple built a performance machine in that 2010 baseline mac pro. For lighter computing it feels fast enough, but anything does these days if you've used computers for more than a decade. You have some really good choices at 65-95W and a couple okay ones at 45W. If Ivy Bridge surprises me and intel does actually come through, we may see see some great choices in smaller thermal envelopes. Regarding power, the mini still has slower hardware than what is available on the macbook pros and imacs. The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
Quote:
Originally Posted by Marvin
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
Configurations like this have been fit into mini towers for years. Some of them really do suck, but not all are that way. In terms of heat Apple has quite a bit of experience trimming things down. Let's talk about thunderbolt for a minute here. It has a pretty limited ability to supply power. Currently it still lacks the throughput of solutions like SAS cards. I'd really like to see that change to the point of where it could start to edge out legacy ports. Even then PCI is still there. There are still some ways it could be implemented in the future. As mentioned SSD storage cards at some point in the future could provide a better means of internal storage. Presently a move like going for 2.5" bays could cut a minor amount of space.
If Apple is planning to completely ditch optical drives, that would cut further room needed. That's one I've always been a bit hesitant on because I still use it for a few things even though I abhor optical media. Downloads just aren't truly instantaneous and a lot of offices still have crappy methods for intra office file sharing of large documents. So anyway I'd like to see them either roll the low end of the mac pro line into a more cost effective line that can at least perform in line with the imac, or bring up the performance on the low end of the mac pro to a truly suitable level for its price point.
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth.
The director of Thunderbolt engineering says 40Gbps per port (20 up, 20 down):
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
Quote:
Originally Posted by hmm
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it.
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
Quote:
Originally Posted by hmm
a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
Quote:
Originally Posted by hmm
By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
Quote:
Originally Posted by hmm
The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
For some users yes it functions as a desktop. For the majority of Apple users it doesn't. That should be pretty clear as this conversation wouldn't be happening otherwise. Further Mini sales would be exploding considering Apples recent market successes.
Quote:
The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
Every single response here you keep falling back onto the processors speed as a key argument. First the quad doesn't even come close to Mac Pro performance, more importantly you are stuck with the built in GPU. All of this is running on a laptop like memory bus.
Taken as a whole though the Mini comes up short capability wise. A computer is made up of parts, or provisioned for parts beyond the CPU.
Quote:
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
You can run a Pegasus RAID drive and a display from the one Thunderbolt port and stick another display on the HDMI port.
That HDMI port does make the Mini very useful for certain applications. However the point about the TB port is still valid.
Quote:
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
BS they have crippled the Mini with every other release they can do it again with the Ivy Bridge (IB) machine. In fact I pretty much expect that they will debut the IB Mini in such a way that the iMac is a far better machine performance wise. It would be shocking to find a Mini effectively running as fast as it's contemporary iMac, with the same number of cores.
Quote:
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
Have you even looked at a modern computer lately? There is no reason for a desktop mid range machine to be as large as an iMac these days. There is no need for the large bays, no need even for the 3.5" bays.
Remember we are going into 2012 all of the old requirements of the past decades are pretty much legacy items. Motherboards are comparatively small these days, there is no need for bays to support optical drives, solid state storage for the most part comes in laptop drive format or some sort of printed circuit card. There is no reason to suggest a massive box.
Quote:
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
All I know or need to know is that similar hardware is available and at reasonable prices. No I don't expect Apple to hit those low prices, but it clearly indicates that my price range is viable. Apple can come close though by doing what they always do, that is distill the machine down to a minimum required to deliver the required functionality.
Quote:
Not with PCI slots.
What is it with PCI slots that is a problem. Engineers can put three of them into a 1U server box these days. it isn't a big deal.
Quote:
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
I really don't know why you are so intent on selling dual core IB chips. It is a step backwards especially if your software is heavily threaded or you are in the habit of running lots of threads. You have convinced yourself that the performance is there but we have nothing to support that. Further history has shown that multiple cores are more important than single core performance.
Quote:
No reason you can't get a USB 3 adaptor for this and way easier to install.
I've stopped laughing now. The response above obviously indicates that you don't get it.
Quote:
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
You are so damn convinced that this machine can only go into a mid range tower that you can't see the other possibilities. They could build the machine into a round ball for all I care. The reason for that statement, about the pizza box, is to get people to think outside the box so to speak. XMac is not the type of machine that is heavily constrained by yesterday's technology. Think of it as a machine deserving the engineering time of an Apple laptop focusing on desktop needs.
I will try to think of a test for this. As I've stated my opinion on the mac pro was that it was a downgrade of the 2008 machine from the processor aspect. The 2010 one had a decent gpu option though. Now when you do run longer tasks on it, that mac pro really isn't very fast, and it's a horrendous value. The gpu if you take the cto is about the only nice thing in it. The hard drive bays aren't even well designed in terms of cooling unless they fixed that on the 2010 one. The 2009 it was still really uneven.
These days a Mac Pro is really of value only when you can employ all of the possible cores or exploit it's bandwidth capabilities. Unfortunately for many users not all software can do that these days, thus an iMac can look really good if an all in one is viable.
Quote:
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth. It seems pretty high considering the peak bandwidth of displayport 1.2 is roughly 18Gb/s total. I'd like to know what kind of specs they were getting for throughput + what the display resolution was, and if this setup supports ddc based calibration as many people who would purchase such a display need that. On mine it goes through a shielded usb cable but not many operate that way.
You get two channels supporting 10GB/S each way. What I'm not clear on is the transmission of the Display Port signal and data on the same channel. It does look like bandwidth is shared in some Intel graphics I've seen while others imply Display Port being switched onto a channel.
More reading is in order but it does look like multiplexing of data is going on to some extent.
Quote:
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it. I also remember how long Apple retained the Core2s in one of their laptops to avoid intel chipsets. I don't think that'll happen again, but it would be easy for them to go the mac pro route on this. Raise the price again (they've adjusted mini price points before) and downgrade their relative choices of processor grades for the machine. Mac pro used to start with mid range dual socket versions. Now they have the cheapest budget model. This is Apple. They've done this kind of thing before because they figure that people will grumble and buy it anyway.
I fully believe Ivy Bridge will be better. The problem I have is that Apple will cripple the processor in the Mini just like they have with every other Mini release. Well that and I have no confidence in dual core technology, if the Mini goes IB it really needs a quad core as a minimal implementation.
Quote:
I'd like to comment on this. The mac pro is one of the larger towers of its type. This design was originally to accommodate the G5 processors with their massive heatsinks and the accompanying logic board design. In 2011 it looks like a dinosaur. Regarding aluminum it's not that expensive of a material. They aren't using a terribly high grade here so it doesn't contribute as much to the cost as you anticipate. In fact much of the cost of aluminum is often from processing it. The macbook airs and pros use a cnc process which should in theory be more labor intensive than the mac pro case. Beyond that they've provided machines using this form factor at $1500 and $2000 during the G5 era into the first generation of the mac pro. Right now the top macbook pro processor available to the 15" in a macbook pro is $568 retail. That machine is still cheaper than a mac pro with the inclusion of the thunderbolt chip. The 27" imac has the cost of a 27" ips panel (not the most expensive one of all but none of the 27" ones are dirt cheap) and uses a processor around $300.
A Dinosaur it is!!!! It is way to big to implement a mid range computer in. It has it's place but it isn't a place many people go to these days.
Quote:
That mac pro you mentioned is artificially inflated. Doing anything to turn it into a better volume machine including a redesigned form factor and up to date setup would encourage volume and potentially a much better deal on the machine. Remember what Apple charged for the original macbook air? It was grossly overpriced relative to its capability, and ended up with dismal sales. Also what do you really gain from single socket Xeons anyway? Going with normal desktop i7s that are updated more frequently would be better for a single socket machine.
It's already been suggested that there are newer more compact methods of implementation for this stuff. It's just a matter of what Apple wants to use.
Whatever they use needs to be an industry accepted standard.
Quote:
Really a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work (I'm serious, if anyone questions it I'll bring up links to companies, pictures of capture carts, etc).
I'm glad I'm not in that business!
However we do use PCs on carts. Sometimes it is just easier to have everything on a cart that passes as a mobile workstation. It is a place for cables, connectors, cameras, I/O boxes and the like.
Quote:
Some use laptops for lighter jobs but they suck. A lighter form factor here would be amazing for these kinds of guys. Rackmountable configurations could be good if there are any Xserves still in use. I realize it wouldn't be the most efficient 2U kind of setup, but it could add more market potential to such a machine.
One day, if it ever takes off, TB supporting Macs may allow for that compact system. It is just that that day is a ways off from what I can see. In the end though I don't think your carts will go away. The computer might be smaller but the cart solves other issues. At work we often use the same sorts of carts to support laptops.
Quote:
It's quite a difference from the mac mini with a 45W processor, but the quad core mini isn't the same thing. Like I said it's not even close to the imac processor which is around 90W. By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
I'm convinced that power levels will drop enough to significantly upgrade the Mini. It really isn't a point to be argued as that is what the industry has done for years. The problem is Apple as they seem to purposely under power the Mini to the point that it doesn't even keep up with their own laptops.
Even if intel comes up a bit short with IB and the rumored and not so rumored massive power reductions we will still see an improvement. The other problem Intel has is with the GPU, they may have to start allocating a lot more transistors to that half of the chip. This will likely be driven by ant success AMD has with Bulldozer based Fusion chips.
Quote:
I think we need to get away from the idea that Apple built a performance machine in that 2010 baseline mac pro. For lighter computing it feels fast enough, but anything does these days if you've used computers for more than a decade. You have some really good choices at 65-95W and a couple okay ones at 45W. If Ivy Bridge surprises me and intel does actually come through, we may see see some great choices in smaller thermal envelopes. Regarding power, the mini still has slower hardware than what is available on the macbook pros and imacs. The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
It is a performance machine just not in the way you or I want.
Quote:
Configurations like this have been fit into mini towers for years. Some of them really do suck, but not all are that way. In terms of heat Apple has quite a bit of experience trimming things down. Let's talk about thunderbolt for a minute here. It has a pretty limited ability to supply power. Currently it still lacks the throughput of solutions like SAS cards. I'd really like to see that change to the point of where it could start to edge out legacy ports. Even then PCI is still there. There are still some ways it could be implemented in the future. As mentioned SSD storage cards at some point in the future could provide a better means of internal storage. Presently a move like going for 2.5" bays could cut a minor amount of space.
It is all about cooling. More importantly a tighter enclosure can put any cooling capability it has to better use. Far less of your air flow gets used to move dead air. moving away from legacy devices and concepts further enhances cooling.
If cooling was a critical issue for Apple they could build upon the concepts used for Compact PCI and other card based systems. There is no reason to maintain old PC approaches to the age old issue of thermal design.
Quote:
If Apple is planning to completely ditch optical drives, that would cut further room needed. That's one I've always been a bit hesitant on because I still use it for a few things even though I abhor optical media. Downloads just aren't truly instantaneous and a lot of offices still have crappy methods for intra office file sharing of large documents. So anyway I'd like to see them either roll the low end of the mac pro line into a more cost effective line that can at least perform in line with the imac, or bring up the performance on the low end of the mac pro to a truly suitable level for its price point.
No matter what they make, an XMac or new Mac Pro they need to build for the future not the past. Building for the future permits innovation.
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
I'd like you to first consider that the guy in that video has the most awesome job title ever created.
Everything I read previously suggested it was shared bandwidth. That would make it considerably more awesome. I do understand that the display is downstream only. We have a couple things that have been emerging over the past couple years. Display panels have been increasing slowly in resolution per inch and 10 bit out is starting to see software and gpu support. Regarding SSDs the newest ones are extremely fast. That's probably encouraged the emergence of them in the PCIe form factor as just a couple can quickly saturate a SATA channel. The SATA standard overall has been around for quite some time. I wouldn't be surprised to see it replaced by the next thing if prices fall on solid state technology. The SATA/HDD form factor was really never ideal for mobile technology although it's been fine in practice. The 8x 1080p streams comment is a case of where you are too trusting of manufacturers . 120Hz panels do consume more bandwidth. 10 bit out which will see increasing adoption rates is more of a bandwidth hog than 8. You can't really go solely on what they tell you there, but yeah it should be fine.
Quote:
Originally Posted by Marvin
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
It hasn't always seemed this way, and at times the release cycles seem a bit weird. Westmere focused more on ultra high end chips and now we haven't seen a replacement in the Xeon line for well over a year (counting from the time of intel's release, not Apple's). Just going by benchmarks never helped me. There are just too many factors that require more in depth testing.
Quote:
Originally Posted by Marvin
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
It'll be interesting to see. There are some definite plateaus to this stuff when it comes to justifying upgrades. When tasks are able to run over a lunch break instead of overnight (say they would have taken 3 hours instead of 1) or overnight rather than over the weekend, or even when something decreases the amount of dedicated server hardware required, it can make a difference. For things involving responsiveness it's just a matter of whether you feel like you're having to stop and wait for the system, or required to taper down settings to avoid this.
Quote:
Originally Posted by Marvin
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
They haven't so far. They come closer in some tests than others, such as on lighter tasks, but they're still behind both the imacs and laptops Regardless I'd figure out some kind of test to run on the best one available. I've read up on every test I can find. I've run some very basic testing on each in the Apple store. I expect that with more ram the difference would only increase.
Quote:
Originally Posted by Marvin
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
The 12 core is a totally different beast from the others. For the price points Apple sets, the starting machine in that line should have used something like a w3670. Sure it's more expensive than what it is in it currently, but it would have maintained a clear demarcation in performance when you move up to a mac pro and it's well within reason for a machine of this price range. Right now it's like you take a moderate dip in performance then have to pay more just to get back to where you were, and nothing about the machine warrants it. Big heavy aluminum cases don't account for the vast majority of that price. nor do the internals.
The 6970M (even the 2GB version) is still significantly slower than the 5870 in all possible uses from gaming to rendering (After Effects, maya, etc). Anyway the 6970M is basically a compromise for the enclosure style of the imac, but it's better than what many of the older imacs received.
May I ask what suggested the current cpu power of the mini as being close to the i7 2600 (can't recall if they use the k version) in the cto imac? Everything I've found suggests outside of lighter computing that doesn't actually saturate the bus, the unit in the imac is pretty far ahead.
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
I really need to find the Intel documents because my understanding is that the data is multiplexed onto the channels. This would make sense due to the requirement that the monitor is the last item on the chain.
Quote:
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
Baloney. At best Intel has been indicating around 20 to 25% speed ups over the last couple of years. Speed ups not always realized in real life. I don't ever recall Intel saying their chips are 50% faster from one model year to the next. Not overall at least though some obscure functions might be. Some years you are lucky to get 10% speed increase in general performance.
Quote:
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
TB is all well and good but right now the world is stuck with PCI cards. On top of that TB adoption in the PC world is spotty at best so many of those cards will stay PCI for a very long time. Beyond that if the laptop needs a disk array for storage you might as well haul around a Mac Pro.
Quote:
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
History does not support your position. I really don't see how you can keep saying things like this.
Quote:
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
I don't think anybody here dismisses the idea that the iMac is a good value. For many it is an excellent platform. A limited platform it is though as you again focus on CPU performance as the only metric by which to judge a machines value.
I really need to find the Intel documents because my understanding is that the data is multiplexed onto the channels. This would make sense due to the requirement that the monitor is the last item on the chain.
I looked for data on this. I couldn't find anything conclusive.
Quote:
Originally Posted by wizard69
Baloney. At best Intel has been indicating around 20 to 25% speed ups over the last couple of years. Speed ups not always realized in real life. I don't ever recall Intel saying their chips are 50% faster from one model year to the next. Not overall at least though some obscure functions might be. Some years you are lucky to get 10% speed increase in general performance.
If you're comparing comparable chips from one year to the next, I haven't seen any truly amazing speed gains from one year to the next in a while. As mentioned the mac pro if you concentrate on roughly the same price point has been in stasis for a long time. I'm not going to compare what you got for $2800 (less from some of the online retailer) in 2008 to what the 2011 12 core monster at $5000. Aside from the gpu, performance at the sub 3k range hasn't budged in these machines since 2008. Slow sales shouldn't be a surprise at that point.
Quote:
Originally Posted by wizard69
TB is all well and good but right now the world is stuck with PCI cards. On top of that TB adoption in the PC world is spotty at best so many of those cards will stay PCI for a very long time. Beyond that if the laptop needs a disk array for storage you might as well haul around a Mac Pro.
I've seen some things being built into new form factors. As long as it's a well developed solution it could work. I'd still like to see the performance levels of a stress tested thunderbolt connection.
Quote:
Originally Posted by wizard69
History does not support your position. I really don't see how you can keep saying things like this.
Westmere was only a moderate gain over the original Nehalem in similar configurations. It was also a die shrink year. Lowered power consumption could allow for a beefier chip within the same thermal spec in the mini, but that doesn't seem to be suggested here. I've been quoting cpu specs because they're something easy to identify and can represent one of the significant cost factors in construction and a significant factor in performance assuming they aren't bottlenecked by something else.
Quote:
Originally Posted by wizard69
I don't think anybody here dismisses the idea that the iMac is a good value. For many it is an excellent platform. A limited platform it is though as you again focus on CPU performance as the only metric by which to judge a machines value.
They have a good cpu, an ok gpu, four slots for ram, and now thunderbolt which should be more useful by next year (too little stuff made for it currently). I dislike having to buy the display as one unit. It can make sense but it's a very expensive forced purchase. I do think that for a lot of people it's still a better choice than the mac pro, and it works out pretty well for someone who would have otherwise purchased an external Apple display. I've never been a fan of Apple displays. They have a nice aesthetic but they've always lacked quality control. I'm still waiting to hear what is so great about glossy. If you hate the sparkle seen on some ips displays it's an LG thing. They use an obnoxiously strong anti reflective coating although it tends to help against color bias from weak reflections. On a consumer display a weaker AR coating can be applied which won't have the sparkle issue but also will not show reflections unless they're taking a lot of direct lighting from a lower angle.
They show 30-60% faster for the top iMac vs the top Mini, which Ivy Bridge will make up next year. The entry Mac Pro is only 5-20% faster. My point is that if people can use these machines as desktops, the Mini is perfectly capable of the same tasks.
Quote:
Originally Posted by wizard69
No I don't expect Apple to hit those low prices, but it clearly indicates that my price range is viable.
Other manufacturers ship desktop towers in much larger volumes.
Quote:
Originally Posted by wizard69
What is it with PCI slots that is a problem. Engineers can put three of them into a 1U server box these days. it isn't a big deal.
You wouldn't put GPUs in a server box though and people would try to in a desktop and it won't handle the heat in a 1U space or the power usage.
Quote:
Originally Posted by wizard69
I really don't know why you are so intent on selling dual core IB chips. It is a step backwards especially if your software is heavily threaded or you are in the habit of running lots of threads.
They work like 4-core chips. The dual-core 2.7GHz Mini performs around the same as the quad 2.5GHz iMac.
Quote:
Originally Posted by wizard69
I've stopped laughing now. The response above obviously indicates that you don't get it.
I don't see why you want to have PCI slots in a machine for ethernet and other comms ports when you can plug a tiny adaptor inline and get the same functionality. You can have a USB modem for fax, TV tuner, audio capture device, ethernet using USB but you'd rather have 4 giant PCI slots making the whole machine 30% bigger. In that respect, your'e right, I don't get it. I don't get why you want something in a PCI form-factor instead of something the size of your thumb.
Quote:
Originally Posted by hmm
The 6970M (even the 2GB version) is still significantly slower than the 5870
The 5870 is under 50% faster. You both seem to consider something 50% faster to be significantly faster. I would say that 100% faster is significant. If you are playing a video game, 50% is the difference between running at 20FPS and 30FPS. All you have to do on the lower one is reduce anti-aliasing or resolution.
You also both keep mentioning the capabilities of the highest-end parts but we're talking about a mid-range machine. Yes, they can take the BTO quad i7 from the 2nd iMac and put it in a small enclosure along with 4 RAM slots and 2-4 HDD bays but they can't get a double-wide PCI slot in it supporting a 200W 5870 GPU too.
This stuff as you said will plateau like all computer hardware. Just look at games consoles. Now that they've hit a certain performance level, nobody cares about the upgrades any more, it's all about software.
I agree that there are more powerful designs now but Apple doesn't build new products knowing they have no future. They build products that start out low and have growth.
For years, I wanted the same kind of mid-range machine described in this thread and I've seen the underpowered hardware come and go but for me, the Ivy Bridge Mini brings it to an end. If the middle one breaks 9k in Geekbench and bundles a Radeon 7000 series GPU and has USB 3, I just don't see the reasons for building a bigger machine.
If people want good price/performance, they need to get an over-clocked PC dedicated to rendering and/or gaming. If they want a stylish, stable, powerful, everyday machine, get a Mini.
Comments
Ivy Bridge can fix it quite easily. If they drop the power usage of the quad down to say 30W, they can get a Radeon GPU in there.
They are all hyper-threaded now so dual-cores perform very well in multi-core tasks.
All machines become obsolete too fast. If you bought a quad-core Mac Pro in 2009, it's already being outperformed by the quad-core Mini. That's 2 years for the Pro to drop to what you'd call 'entry-level'.
If you look at any of the generic parts used in the 2009 model (say everything but the case and logic board design) none of it was high end hardware even if it was priced that way. It used the xeon version of the i7 920 which was one of their volume processors. The basic graphics card option went for around $100 on the pc side if I recall correctly. There wasn't anything special about it really. There's just nothing that is terribly amazing about the engineering or hardware of the machine.
Regarding the mini. If it got usb3 (already on PCs in that price range), a second thunderbolt port, and decent graphics (intel graphics have always sucked, remember how Apple used to prefer nvidia chipsets before intel sued them?), we could have an ok machine. I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop, I'd use it as one. Considering the lack of internal storage bays, a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth. If you look at what I care about personally, it's a system without severe bottlenecks. I tend to prefer the ability for PCI expansion, and I think it would help widen the appeal, but if it had more than a single thunderbolt port it would be less of an issue for me personally.
Right now Apple's retreat is kind of awkward. The mac pro has kind of an uncertain future, and they haven't shown it much love, but some of the video editors and stuff don't really have the option of even looking at a mini if they need huge disk IO bandwidth or realtime playback. You can actually get pci expansion even on cheap PCs for this kind of thing, but then you wouldn't be able to run OSX. Some accessories of this sort have started to debut in thunderbolt versions which somewhat solves this issue (they're still extremely expensive but they could come down in time). The problem is if you're limited to one for display and data, you can saturate it too easily. 120Hz displays are becoming more common as are SSDs. With changes like that and newer panel generations coming in at higher resolutions, thunderbolt will become incredibly tight on bandwidth.
You must understand it's frustrating at times. The lower end is missing features. You get to the imac and performance picks up a little, gpu is still kind of weak and it's not exactly cheap, then go up a bit more in price and go down in performance as it's older and cheaper hardware at its core (aside from the gpu if you order the upgrade). It's not a very balanced line in general. Most of us just choose a compromise and suck it up until we need a new computer again. As for the imac screen, it's not a very good one. It's ok but not great. I wouldn't ever buy the cinema display and it has nothing to do with the price. As of right now Apple doesn't make a single computer that's truly engineered for performance under $3500 before ram, storage upgrades, etc. The rest are engineered to be compact or portable and look pretty. This year they haven't had anything where I wanted to accept the compromise. Hopefully they will next year.
Edit: by the way, the mini has never been one of their top sellers. I think if they did turn it into a slightly more complete/robust system it might go somewhere. Right now it lacks the portability of a laptop and the performance of a desktop. It also lacks an easily added display as the cinema display is too costly for such a machine. Overall it's just in an awkward spot.
Ivy Bridge can fix it quite easily. If they drop the power usage of the quad down to say 30W, they can get a Radeon GPU in there.
Really it can't if Apples intentions are to purposefully limit the Mini's capability relative to the rest of the line up. Again it isn't a question of the value of a low end machine, Apple needs one, it is the problem of the step up being so limited. Limited to the extent that it won't have a long functional life.
They are all hyper-threaded now so dual-cores perform very well in multi-core tasks.
Sometimes yes most of the time no.
All machines become obsolete too fast. If you bought a quad-core Mac Pro in 2009, it's already being outperformed by the quad-core Mini. That's 2 years for the Pro to drop to what you'd call 'entry-level'.
The hardware can become outdated too fast but in Apples case they make sure of that with the Mini. In any event I think you are still missing the point, it is possible these days to buy hardware that will remain viable for much longer than would have been the case in the past. It requires that the hardware have a few features to help keep it viable for a few years.
Everything it seems. You want lots of internal storage, fast processors, fast GPU, more RAM slots, PCI expansion slots - you basically want a cheaper Mac Pro and if they put one on sale, you'd still want an extra power supply.
Again I'm apparently not communicating well. That list represented an aggregate of features an XMac would need to have and no it is not a Mac Pro. In reality this is a mid range machine and is commonly represented by any number of PC systems as desktop PCs. One would have to have their head awfully deep in the sand not to recognize that simply implementing desktop parts gives you most of this for free. Apples job is simply to build a box.
As to the extra power supply, that has a lot to do with Apple trying to market the current hardware as servers. Not everybody needs such a capability, but if a box is to be marketed as a server it really ought to support the feature.
You can't get it all and still have it affordable because it has to sell in volume and the things that matter to you don't matter to the vast majority of people.
That is BS. Especially in this economic climate, people are far more careful with their spending. You imply that it would be expensive but that flies in the face of what motherboards and the various chips on board cost. Seriously if Apple can't produce a desktop machine in the $1000 to $1500 range then they need to hire some new engineers. In that price range they should be able to deliver a decent box that will sell well. That box would include a processor in the 45-65 watt range, a GPU, RAM and some bays/slots for storage. This sort of machine could easily be a quarter of the size of the Mac Pro.
The 27" iMac is a better value proposition because you get a 27" IPS screen included worth $1000.
The iMac isn't even in the running here and in fact is a terrible value.
Performance doubles every 2 years so about 50% every year. Right now, the quad is 30-50% faster than the dual-core so Ivy Bridge dual-cores will match the Sandy Bridge quad i7 and the Radeon 7000 series should double in performance over the current models.
It would be nice if any of the above was true, but sadly it isn't. For a single app the value of a quad processor is software dependent, it is very possible for an app to use all of those cores. But a single app seldom defines computer usage these days, those cores can be effectively used when multiple apos are running.
You seem to be extremely optimistic with respect to Ivy Bridge and its performance. I really don't think it is justified. More cores is a fundamental advantage.
Basically, I expect the Ivy Bridge version of the middle Mac Mini to match the current $2500 Mac Pro in CPU and GPU.
Well that is very optimistic. However we already know that the Ivy Bridge GPU only gets about a 30% increase in performance, to bump the discreet GPU you really need to bump the Video RAM also. In the end, even with Ivy Bridges thermals, you still are trying to stuff more into the box than is practical.
In any event an Ivy Bridge Mini, with just two to four cores won't come close to the performance of a Mac Pro. This should be obvious as the Pro has far more cores available to it.
With USB 3 and the Thunderbolt port, that should be enough for expansion.
USB 3 and Thunderbolt combined will be a significant I/O improvement for the Mini. However that means nothing in the context of desktop machines. TB is totally unsuitable for one of the most common uses for slots in industry, that is the addition of additional communications ports. Here Ethernet is the most common port added to a machine though there are many others. Of course there are a host of other cards that need something better than TB.
Beyond that Apple currently only implements a single TB port, due to the limited bandwidth that is a serious problem.
I think it's the way to go and while I expect the desire for a mid-range tower to continue on ad nauseam, one day it will stop.
Not likely. In fact I see demand growing. Plus you use the term mid-range tower which frankly I don't remember anybody here asking for. They could put the XMac in a pizza box for all I care though I believe there are far better form factors.
If it takes 5 years or even 10 years, one day in the near future, a machine will arrive on the low-end that will put an end to it.
Well in ten years just about anything could happen including a nuclear war that would destroy all of our "I" devices, Macs and what have you. The problem today is solving real issues that users have. The quest for an XMac doesn't go away because Apple does have this gaping hole in their line up. It does little good to try to promote the Mini as a fill in for that hole as it will always be limited in capability.
I suspect the problem here is that when you hear about XMac or other stated desires, you immediately start thinking of a PC tower. A tower that size wise isn't much different than a Mac Pro. I really don't think that is the goal most of us have. Rather we want modern technology applied to the problem to build a platform that takes us into the future. That means dropping legacy devices and breaching new ground. You might of noticed that I often talk about secondary storage in terms of bays/slots, that is because a modern form factor needs to acknowledge that solid state storage is nothing more than another printed circuit card. As such economics should drive such storage on to low cost plug in cards.
In a nut shell the desktop is ripe for the same sort of innovation that Apple is famous for elsewhere. Frankly Apple, in tandem with Intel, is the only manufacture capable of driving such innovation right now. HP is screwed up, Dell is well Dell and there are few others that could even try.
I say that because they need to get TB into the platform. However it could simply be a transitional machine.
This is all true but you know the so called Pros will make a big stink.
I'm not sure how you get $200 it if a $20 drive but I agree with the rest. For the same reasons I'd like to see the 3.5" disk slots leave also.
Smaller is good. However I think you need to reconsider the issue of expansion. Expansion is the whole point of a Pro machine other wise people would buy iMacs or Minis. Here I'm talking expansion in all the various ways be it RAM, PCI Express slots, "storage bays" or what have you. Without the ability to significantly out do the other machines with expansion there is little reason for the Mac Pro to exist. Simply putting a bunch of CPU power in a box does little good if it isn't supported properly.
The MXM card can go as those have little draw when it comes to Pros. You did not mention RAM expansion which is a valued part of the current Pro.
Ivy Bridge holds a lot of promise for Laptops. I'm doing everything I can to resist buying before that hardware shows up. However I'm not so certain such chips will make sense in a Mac Pro. I'm actually hoping that AMD has really good luck with Bulldozer, especially bulldozer cores integrated into Fusion chips. Bulldozer is so different that Apple will likely have to do a bunch of testing but I see great potential for common workloads on Mac OS. Of course this means Apple would need TB hardware that would work with AMD systems but I'm under the impression that will not be a problem.
Sounds about right but this could happen earlier, again mainly to support the TB initiative.
Maybe there is something that I missed, but Ivy Bridge should make sense for most any platform. The new gate process and reduced power consumption alone should be worthwhile. I don't recall the specifics about the on board L3 cache specs of Ivy Bridge, but the trend is to increase L3 cache because it has such a worthwhile effect upon performance. With all these cores, even the faster bus speeds to the RAM are limiting, although dual channel RAM and lots of RAM appears to be working well.
There are some PCIe based SSD systems which, because of price, are presently limited to the server market for the most part, but they illustrate the need for faster systems to feed data to all those cores.
For example, the 15 inch MacBook Pro top-of-the-line model uses the 2.3 GHz CPU rather than the 2.2 GHz CPU. That is not much of a difference in clock speed for a $250 premium, but it has 8 MB L3 cache instead of 6 MB L3 cache of the 2.2 GHz unit. L3 cache can make a difference in overall performance in applications such as Photoshop, Final Cut Pro or any of the various video applications.
The Mac Pro community has not been getting much love from Apple for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server. When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
If you look at any of the generic parts used in the 2009 model (say everything but the case and logic board design) none of it was high end hardware even if it was priced that way. It used the xeon version of the i7 920 which was one of their volume processors. The basic graphics card option went for around $100 on the pc side if I recall correctly. There wasn't anything special about it really. There's just nothing that is terribly amazing about the engineering or hardware of the machine.
I suspect the problem with the Mac Pro is that it was or is unprofitable for Apple. Thus they turned towards the cheaper end ofIntels hardware line up and raised prices. Unfortunately this is a common mistake in industry as it results in a downward spiral in sales. The Pro effectively becomes harder to justify, due to the cost issues. Interestingly I suspect this biases sales towards the high end configurations. After all the people buying Mac Pros are now more likely to need the power of the high end configurations anyways.
Regarding the mini. If it got usb3 (already on PCs in that price range), a second thunderbolt port, and decent graphics (intel graphics have always sucked, remember how Apple used to prefer nvidia chipsets before intel sued them?), we could have an ok machine.
I don't really see it that way. The Mini is a very OK machine if your needs are within it's limits. Unfortunately those limits are pretty tight and inflexible. So if we acknowledge that there is a hole in Apples line up, is it smart to try to fill it with something that never was designed to fill that hole. Sort of like the proverbial round peg in the square hole, or does Apple make a peg to fit.
I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop, I'd use it as one. Considering the lack of internal storage bays, a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth. If you look at what I care about personally, it's a system without severe bottlenecks. I tend to prefer the ability for PCI expansion, and I think it would help widen the appeal, but if it had more than a single thunderbolt port it would be less of an issue for me personally.
I must confess to needing to read up on TB more but can data really impact video bandwidth? There is some confusion in the documents I've seen so far but I was under the impression that TB would dedicate one channel to video data. That may be a mistake on my part because other documents seem to imply multiplexing of data with video.
One day I need to sit down and develop a firmer grasp of TB. It is just very hard for me to get excited about the interface.
Right now Apple's retreat is kind of awkward. The mac pro has kind of an uncertain future, and they haven't shown it much love, but some of the video editors and stuff don't really have the option of even looking at a mini if they need huge disk IO bandwidth or realtime playback. You can actually get pci expansion even on cheap PCs for this kind of thing, but then you wouldn't be able to run OSX.
I'm not sure where this idea that expansion has to imply expensive came from. At the same time people promote TB even though it will always be relatively expensive. I just don't get it.
Some accessories of this sort have started to debut in thunderbolt versions which somewhat solves this issue (they're still extremely expensive but they could come down in time). The problem is if you're limited to one for display and data, you can saturate it too easily. 120Hz displays are becoming more common as are SSDs. With changes like that and newer panel generations coming in at higher resolutions, thunderbolt will become incredibly tight on bandwidth.
A serious issue! Having just one TB port on a Mini does not significantly alleviate it's built in limitations. In the end the Mini is still a low end solution. How they will address this on the Mac Pros replacement is an open question.
Note I said Mac Pros replacement. I really believe that the next Pro will be vastly different
You must understand it's frustrating at times. The lower end is missing features. You get to the imac and performance picks up a little, gpu is still kind of weak and it's not exactly cheap, then go up a bit more in price and go down in performance as it's older and cheaper hardware at its core (aside from the gpu if you order the upgrade).
It isn't a universal drop in performance, it does require that software can use all of those cores though.
It's not a very balanced line in general. Most of us just choose a compromise and suck it up until we need a new computer again. As for the imac screen, it's not a very good one. It's ok but not great. I wouldn't ever buy the cinema display and it has nothing to do with the price. As of right now Apple doesn't make a single computer that's truly engineered for performance under $3500 before ram, storage upgrades, etc. The rest are engineered to be compact or portable and look pretty. This year they haven't had anything where I wanted to accept the compromise. Hopefully they will next year.
Heck I'd be happy with midrange performance. It is a rather pathetic reality but one does get a better deal out of Apples laptops if performance is one significant measure.
Edit: by the way, the mini has never been one of their top sellers. I think if they did turn it into a slightly more complete/robust system it might go somewhere.
I'm not sure the above is true. It often sneaks onto Apples top sellers list and is very popular with the Internet sellers. Given that Apple hardly promotes the machine as they seem to be obsessed with the notebook market.
Right now it lacks the portability of a laptop and the performance of a desktop. It also lacks an easily added display as the cinema display is too costly for such a machine. Overall it's just in an awkward spot.
Again I don't believe the Mini is all that bad, intact it has many positives going for it. It just isn't the machine to invest in as a workstation. Off the top of my head good points with respect to the Mini:
- Low power as in watts. This makes the Mini very usable in a number of use cases where a more conventional machine would not work.
- As configured it is a very good home theater PC.
- Compact is good where space is limited.
- It is a very good to excellent network connected node.
- It isn't an iMac with a built in display.
- The current revision is somewhat more serviceable. In any event it is a hands down winner over the iMac.
- It is well constructed and with SSD one might in a stretch call it rugged. Maybe semi rugged is a better phrase.
The problem for me and I suspect you, is that it really comes up short as a workstation. Maybe others don't agree with our desires but I don't see them as irrational, nor expensive to implement.<snip>
One more thing (I used that phrase long before I heard Steve say it) even the ipad and iphone have dual core processors these days. Quad is basically the norm for desktop computing. It's not exactly a high end feature. I'm trying to think of something more to add here to make the discussion interesting, but it seems we're just at odds on a lot of these points.
So did Columbo. "Oh, just one more thing" and then he sprung the trap.
Maybe there is something that I missed, but Ivy Bridge should make sense for most any platform. The new gate process and reduced power consumption alone should be worthwhile. I don't recall the specifics about the on board L3 cache specs of Ivy Bridge, but the trend is to increase L3 cache because it has such a worthwhile effect upon performance. With all these cores, even the faster bus speeds to the RAM are limiting, although dual channel RAM and lots of RAM appears to be working well.
This is correct but there is one problem. If that Ivy Bridge installation uses the on board GPU then graphics puts a lot of pressure on the cache. AMD has tried to address this in some novel ways so I'm sure Intel has at least studied it.
Note to there is a positive aspect to the GPU using the cache, it greatly enhances communications between the CPU and the integrated GPU.
There are some PCIe based SSD systems which, because of price, are presently limited to the server market for the most part, but they illustrate the need for faster systems to feed data to all those cores.
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
For example, the 15 inch MacBook Pro top-of-the-line model uses the 2.3 GHz CPU rather than the 2.2 GHz CPU. That is not much of a difference in clock speed for a $250 premium, but it has 8 MB L3 cache instead of 6 MB L3 cache of the 2.2 GHz unit. L3 cache can make a difference in overall performance in applications such as Photoshop, Final Cut Pro or any of the various video applications.
There is no doubt what so ever that more cache can help significantly with many workloads. However at some point you need to up core clock rate or add cores to get speed boosts. What will be interesting with Ivy Bridge is just what the maximum clock rates will be for various power levels. I still suspect we will see a wide spread in performance values from AIR suitable chips to laptop chips to desktop chips. Thus the need for that mid-range desktop machine.
The Mac Pro community has not been getting much love from Apple
I'd have to say the entire desktop line is getting ignored. If you compare it to the love the laptops get you would have to wonder how they get any desktop sales at all.
for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server.
The server market is slow to change so getting traction in that market with a Mini is tough. In some regards I see these sorts of ultra small machines as the wave of the future in the data center. It probably won't be the Mini that makes the big dent there though. Most likely it will be an ARM based system that is maybe 1/6 the width of a rack based server.
When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
My problem is that there seems to be excessive expectations with respect to performance. For example the GPU has gone through limited public benchmarking and is not all that great of an improvement. Something like 30% to 60% in a few specific cases. Actual CPU performance increases seem to be unclear on a clock by clock basis.
This is correct but there is one problem. If that Ivy Bridge installation uses the on board GPU then graphics puts a lot of pressure on the cache. AMD has tried to address this in some novel ways so I'm sure Intel has at least studied it.
Note to there is a positive aspect to the GPU using the cache, it greatly enhances communications between the CPU and the integrated GPU.
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
There is no doubt what so ever that more cache can help significantly with many workloads. However at some point you need to up core clock rate or add cores to get speed boosts. What will be interesting with Ivy Bridge is just what the maximum clock rates will be for various power levels. I still suspect we will see a wide spread in performance values from AIR suitable chips to laptop chips to desktop chips. Thus the need for that mid-range desktop machine.
I'd have to say the entire desktop line is getting ignored. If you compare it to the love the laptops get you would have to wonder how they get any desktop sales at all.
The server market is slow to change so getting traction in that market with a Mini is tough. In some regards I see these sorts of ultra small machines as the wave of the future in the data center. It probably won't be the Mini that makes the big dent there though. Most likely it will be an ARM based system that is maybe 1/6 the width of a rack based server.
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
The laptop market has been an increasingly larger segment of all PC sales, not just Apple's, and so I understand devoting a lot of resources to it...personally I wish all of the manufacturers would make matte/non-glare screens available on all models..., but there is still a substantial market for desktops. Apple abandoned the XServe because it represented such a small portion of their Mac sales (note that Mac sales are only something in the neighborhood of 25% of Apple revenue), but it got the proverbial foot in the door at major edu sites and at least some businesses (usually smaller ones).
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
There were many good ideas in the XServe. Some of them were ridiculously simple, but practical, such as the flashing light on the unit with the drive problem so that the system admin knew which one in the bank of XServes to go to.
Graphics are very important in some scientific applications. I am aware of Mac Pros being used for medical research visualizing drug and cell interactions at the molecular level which takes some "grunt" to accomplish in a useful time frame. Those people are doing important work with some OS X apps and some UNIX apps. Sooner than later they will have to look for other solutions if Apple does not devote the resources to keep current with these things.
Cheers
The laptop market has been an increasingly larger segment of all PC sales, not just Apple's, and so I understand devoting a lot of resources to it...personally I wish all of the manufacturers would make matte/non-glare screens available on all models..., but there is still a substantial market for desktops. Apple abandoned the XServe because it represented such a small portion of their Mac sales (note that Mac sales are only something in the neighborhood of 25% of Apple revenue), but it got the proverbial foot in the door at major edu sites and at least some businesses (usually smaller ones).
My problem is this, if Apple doesn't do something about Mac Pros soon they will go the way of the XServe. On top of that Apple has a lot of desktop sales they could steal from the PC industry. I realize the laptop market is strong but the desktop market is not dead and shouldn't be ignored.
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
Apple buying Sun? I think that would have been ugly.
There were many good ideas in the XServe. Some of them were ridiculously simple, but practical, such as the flashing light on the unit with the drive problem so that the system admin knew which one in the bank of XServes to go to.
They where good servers but also very expensive for what you got. Apple screwed themselves just as they have with the Mac Pro. Most people don't mind paying a little extra for their Macs but bolt at highway robbery. In a data center this is even more important as no one cares about the manufacture it is all about cost.
In any event those server related features are cheap to implement and could easily be added to a desktop Mac.
Graphics are very important in some scientific applications. I am aware of Mac Pros being used for medical research visualizing drug and cell interactions at the molecular level which takes some "grunt" to accomplish in a useful time frame. Those people are doing important work with some OS X apps and some UNIX apps. Sooner than later they will have to look for other solutions if Apple does not devote the resources to keep current with these things.
Cheers
Well those people buying such hardware are generally better in formed than the mainstream. The thing is if you max out a Mac Pro it really isn't that bad of a machine. After all how many desktops are built to implement multi core dual socket Xeon's? The hardware to upgrade the Pro will come soon. So I don't see it as being neglected in the context of the high end customer. However Apple does have to make that upgrade compelling.
The Mac Pro community has not been getting much love from Apple for some time and the only "server" is the Mac Mini and while useful for a lot of things, it isn't quite the same as a really big server. When the existing server units give out one has to wonder what institutional users are going to do. It might be that Sun/Oracle will be knocking on their door.
Yeah and since it's Apple they don't truly acknowledge their direction for the line due to their culture of secrecy. Even when new processors come out, it's not a guarantee of updates anytime soon. Recall how long it took just to get the weak westmere refresh?
I suspect the problem with the Mac Pro is that it was or is unprofitable for Apple. Thus they turned towards the cheaper end ofIntels hardware line up and raised prices. Unfortunately this is a common mistake in industry as it results in a downward spiral in sales. The Pro effectively becomes harder to justify, due to the cost issues. Interestingly I suspect this biases sales towards the high end configurations. After all the people buying Mac Pros are now more likely to need the power of the high end configurations anyways.
The higher end configurations are still priced high relative to similar configurations PC side and not necessarily much better (really Apple's engineering time seems to go into other lines). It's just that at that point the difference generally isn't completely excessive. At that level the machine will be expensive no matter where you buy it. My issue was that the configurations are terribly awkward. In 2009 when they went to four slots for ram in the single socket machine, 4GB dimms were still quite costly. This meant unless you were willing to spend significantly more these machines were limited to 8GB of ram, which is okay for many people but not everyone. That's just one example, but in 2010 the configurations became even more awkward. They don't balance clock speed to core count very well in several cases. For the cost of the starting eight core the processor choice could have been a better one in both 2009 and 2010, as they resulted in an overall slower machine when you weren't stressing all available cores.
So did Columbo. "Oh, just one more thing" and then he sprung the trap.
Columbo quote = automatically awesome
Yes, currently expensive but realize that everybody and their brother is working on doing low cost high performance PCI - Exprees storage solutions. OCZ in fact just bought PLX most likely for their PCI - Express experience. I suspect that SSD's on old hard drive form factors are a thing of the past. Once reliable and low cost silicon can be had SATA will due a rapid death. This is actually a good thing for Apples laptops as it means more logic can be dumped reducing thermals yet again.
Your posts are fun to read. You were the first that even brought up PCIe SSD storage. I'm wondering what the preferred form will be for storage enclosures and external drives. You mentioned the cost. Wouldn't those allow for a simpler manufacturing process? I realize card prices are all over the place depending on type and other factors.
Possibly. This is another reason that I promote the XMac as a flexible platform. It wouldn't be all that difficult to design an XMac that could be ganged up to fit into a standard rack space but yet be fully suitable for the desktop. For many users the Mini won't be accepted as a server no matter how sensible it may be to some. With a couple of key refinements an XMac could. Apple really should address the issue and I believe a flexible platform is the best avenue. In many ways a server is often nothing more than a mid-range machine with a couple of enhancements.
There was that rumor about Apple developing a rackmountable mac pro, but the only available "photo" just looked like something photoshopped. Remember back in June or so? If they're trying to get in the server market it could make sense. I've been saying for a while that they need to do something with it to prevent its following the Xserve's fate. If they can come up with a good strategy for a new machine, they obviously have the funds to develop it. I'd really like them to develop a successor to that line that's actually functional in the sense that it remains up to date and is priced in a way that isn't completely insane (I've made points on this before). No one wants a 2009 design in late 2011 for the same price. It shouldn't surprise Apple either.
The one thing Apple has resisted for far too many years in my view is some sort of association with SUN, now SUN/Oracle, to provide integrated solutions for business, education and so on. Larry Ellison even described SUN/Oracle as Apple for business recently. Its just too bad that Apple didn't buy them and make it the Mac Business Unit.
Cheers
Hahahaha ahhhh that would have been so bad. They don't have an exceptional reputation in managing independent product lines after acquisition when it comes to things designed for commercial usage. Look at what they did to Shake after they purchased that. I've noticed everyone has a different idea of what Apple should buy with their massive cash pile, but many of the suggestions don't seem synergetic with Apple the way they are today.
My problem is this, if Apple doesn't do something about Mac Pros soon they will go the way of the XServe. On top of that Apple has a lot of desktop sales they could steal from the PC industry. I realize the laptop market is strong but the desktop market is not dead and shouldn't be ignored.
Well those people buying such hardware are generally better in formed than the mainstream. The thing is if you max out a Mac Pro it really isn't that bad of a machine. After all how many desktops are built to implement multi core dual socket Xeon's? The hardware to upgrade the Pro will come soon. So I don't see it as being neglected in the context of the high end customer. However Apple does have to make that upgrade compelling.
There are a few workstation manufacturers like that. I agree with you on the parallel to the Xserve. I've been saying that for some time. Letting it sit really doesn't inspire confidence. You've probably noticed but a lot of people still own either the original mac pro or the 2008 configuration. There's not a lot of compelling reason to update from the 2008 model unless your needs demand one of the top machines which is kind of weird three years later. The 2006 one also aged decently for how old it is. Sure it's slower than the new ones by a wide margin, but I would suggest that it probably held out better than these newer ones will over the next five years.
I'd like to see them use desktop cpus like the imac. As mentioned the price to performance ratio is better. Basically if it had just enough to function as a desktop
The Mini functions just fine as a desktop. The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
a second thunderbolt port is kind of a big deal. This allows non throttled use of SSDs and/or storage without throttling display bandwidth
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
You can run a Pegasus RAID drive and a display from the one Thunderbolt port and stick another display on the HDMI port.
Really it can't if Apples intentions are to purposefully limit the Mini's capability relative to the rest of the line up.
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
Apples job is simply to build a box.
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
This sort of machine could easily be a quarter of the size of the Mac Pro.
Not with PCI slots.
In any event an Ivy Bridge Mini, with just two to four cores won't come close to the performance of a Mac Pro. This should be obvious as the Pro has far more cores available to it.
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
Here Ethernet is the most common port added to a machine though there are many others.
No reason you can't get a USB 3 adaptor for this and way easier to install.
you use the term mid-range tower which frankly I don't remember anybody here asking for. They could put the XMac in a pizza box
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
The Mini functions just fine as a desktop. The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
I will try to think of a test for this. As I've stated my opinion on the mac pro was that it was a downgrade of the 2008 machine from the processor aspect. The 2010 one had a decent gpu option though. Now when you do run longer tasks on it, that mac pro really isn't very fast, and it's a horrendous value. The gpu if you take the cto is about the only nice thing in it. The hard drive bays aren't even well designed in terms of cooling unless they fixed that on the 2010 one. The 2009 it was still really uneven.
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth. It seems pretty high considering the peak bandwidth of displayport 1.2 is roughly 18Gb/s total. I'd like to know what kind of specs they were getting for throughput + what the display resolution was, and if this setup supports ddc based calibration as many people who would purchase such a display need that. On mine it goes through a shielded usb cable but not many operate that way.
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it. I also remember how long Apple retained the Core2s in one of their laptops to avoid intel chipsets. I don't think that'll happen again, but it would be easy for them to go the mac pro route on this. Raise the price again (they've adjusted mini price points before) and downgrade their relative choices of processor grades for the machine. Mac pro used to start with mid range dual socket versions. Now they have the cheapest budget model. This is Apple. They've done this kind of thing before because they figure that people will grumble and buy it anyway.
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
I'd like to comment on this. The mac pro is one of the larger towers of its type. This design was originally to accommodate the G5 processors with their massive heatsinks and the accompanying logic board design. In 2011 it looks like a dinosaur. Regarding aluminum it's not that expensive of a material. They aren't using a terribly high grade here so it doesn't contribute as much to the cost as you anticipate. In fact much of the cost of aluminum is often from processing it. The macbook airs and pros use a cnc process which should in theory be more labor intensive than the mac pro case. Beyond that they've provided machines using this form factor at $1500 and $2000 during the G5 era into the first generation of the mac pro. Right now the top macbook pro processor available to the 15" in a macbook pro is $568 retail. That machine is still cheaper than a mac pro with the inclusion of the thunderbolt chip. The 27" imac has the cost of a 27" ips panel (not the most expensive one of all but none of the 27" ones are dirt cheap) and uses a processor around $300.
That mac pro you mentioned is artificially inflated. Doing anything to turn it into a better volume machine including a redesigned form factor and up to date setup would encourage volume and potentially a much better deal on the machine. Remember what Apple charged for the original macbook air? It was grossly overpriced relative to its capability, and ended up with dismal sales. Also what do you really gain from single socket Xeons anyway? Going with normal desktop i7s that are updated more frequently would be better for a single socket machine.
Not with PCI slots.
It's already been suggested that there are newer more compact methods of implementation for this stuff. It's just a matter of what Apple wants to use. Really a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work (I'm serious, if anyone questions it I'll bring up links to companies, pictures of capture carts, etc). Some use laptops for lighter jobs but they suck. A lighter form factor here would be amazing for these kinds of guys. Rackmountable configurations could be good if there are any Xserves still in use. I realize it wouldn't be the most efficient 2U kind of setup, but it could add more market potential to such a machine.
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
It's quite a difference from the mac mini with a 45W processor, but the quad core mini isn't the same thing. Like I said it's not even close to the imac processor which is around 90W. By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
I think we need to get away from the idea that Apple built a performance machine in that 2010 baseline mac pro. For lighter computing it feels fast enough, but anything does these days if you've used computers for more than a decade. You have some really good choices at 65-95W and a couple okay ones at 45W. If Ivy Bridge surprises me and intel does actually come through, we may see see some great choices in smaller thermal envelopes. Regarding power, the mini still has slower hardware than what is available on the macbook pros and imacs. The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
Configurations like this have been fit into mini towers for years. Some of them really do suck, but not all are that way. In terms of heat Apple has quite a bit of experience trimming things down. Let's talk about thunderbolt for a minute here. It has a pretty limited ability to supply power. Currently it still lacks the throughput of solutions like SAS cards. I'd really like to see that change to the point of where it could start to edge out legacy ports. Even then PCI is still there. There are still some ways it could be implemented in the future. As mentioned SSD storage cards at some point in the future could provide a better means of internal storage. Presently a move like going for 2.5" bays could cut a minor amount of space.
If Apple is planning to completely ditch optical drives, that would cut further room needed. That's one I've always been a bit hesitant on because I still use it for a few things even though I abhor optical media. Downloads just aren't truly instantaneous and a lot of offices still have crappy methods for intra office file sharing of large documents. So anyway I'd like to see them either roll the low end of the mac pro line into a more cost effective line that can at least perform in line with the imac, or bring up the performance on the low end of the mac pro to a truly suitable level for its price point.
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth.
The director of Thunderbolt engineering says 40Gbps per port (20 up, 20 down):
http://www.pcmag.com/article2/0,2817...id=mxopWpE3s3y
Demoes like this one show 13Gbps:
http://www.youtube.com/watch?v=gk69p...eature=related
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it.
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
The Mini functions just fine as a desktop.
For some users yes it functions as a desktop. For the majority of Apple users it doesn't. That should be pretty clear as this conversation wouldn't be happening otherwise. Further Mini sales would be exploding considering Apples recent market successes.
The quad i7 performs close to a current quad Mac Pro. If the Mini can't function as a desktop then logically neither can the entry Mac Pro nor anything in between. Desktop parts do give better price/performance but not all that much to cross the line from usable to unusable.
Every single response here you keep falling back onto the processors speed as a key argument. First the quad doesn't even come close to Mac Pro performance, more importantly you are stuck with the built in GPU. All of this is running on a laptop like memory bus.
Taken as a whole though the Mini comes up short capability wise. A computer is made up of parts, or provisioned for parts beyond the CPU.
Thunderbolt uses two channels at 10Gbps each, one for PCI data (storage etc), the other for display and the bandwidth isn't shared. Intel specifically demoed a RAID system connected with a high-resolution display on the end of the chain to highlight this.
You can run a Pegasus RAID drive and a display from the one Thunderbolt port and stick another display on the HDMI port.
That HDMI port does make the Mini very useful for certain applications. However the point about the TB port is still valid.
There's nothing they can do to stop the Mini getting great performance from Ivy Bridge.
BS they have crippled the Mini with every other release they can do it again with the Ivy Bridge (IB) machine. In fact I pretty much expect that they will debut the IB Mini in such a way that the iMac is a far better machine performance wise. It would be shocking to find a Mini effectively running as fast as it's contemporary iMac, with the same number of cores.
Right but a box that has a desktop motherboard and PCI slots along with a PSU big enough to handle the load. It needs to have enough room for large internal storage too. This is pretty close to the volume of the Mac Pro. So what do they make the enclosure from? Not plastic so it will be another aluminium monster that is heavy and expensive.
Have you even looked at a modern computer lately? There is no reason for a desktop mid range machine to be as large as an iMac these days. There is no need for the large bays, no need even for the 3.5" bays.
Remember we are going into 2012 all of the old requirements of the past decades are pretty much legacy items. Motherboards are comparatively small these days, there is no need for bays to support optical drives, solid state storage for the most part comes in laptop drive format or some sort of printed circuit card. There is no reason to suggest a massive box.
How do they make this machine cheaper? The Xeon they use in the entry model is around $300, which is how much the i7 would be.
All I know or need to know is that similar hardware is available and at reasonable prices. No I don't expect Apple to hit those low prices, but it clearly indicates that my price range is viable. Apple can come close though by doing what they always do, that is distill the machine down to a minimum required to deliver the required functionality.
Not with PCI slots.
What is it with PCI slots that is a problem. Engineers can put three of them into a 1U server box these days. it isn't a big deal.
The cheapest 6-core Mac Pro is $3700 and I don't see any 6-core i7s that can possibly go in a mid-range machine. They are 130W and nearly $600 and I don't remember them making a Sandy Bridge version. So we are stuck with quad-core for now and the quad-core Mini certainly does come close to the quad-core Mac Pro performance. The Ivy Bridge dual-cores will come close to the current-gen Mac Pro (though obviously the new Xeon in November will change this).
I really don't know why you are so intent on selling dual core IB chips. It is a step backwards especially if your software is heavily threaded or you are in the habit of running lots of threads. You have convinced yourself that the performance is there but we have nothing to support that. Further history has shown that multiple cores are more important than single core performance.
No reason you can't get a USB 3 adaptor for this and way easier to install.
I've stopped laughing now. The response above obviously indicates that you don't get it.
That would be like an iMac design but you can already see how cramped it is in there with no PCI slots and one HDD. If you want all the parts you mention, it has to be a mid-range tower. If you want a pizza box, you have to do without.
You are so damn convinced that this machine can only go into a mid range tower that you can't see the other possibilities. They could build the machine into a round ball for all I care. The reason for that statement, about the pizza box, is to get people to think outside the box so to speak. XMac is not the type of machine that is heavily constrained by yesterday's technology. Think of it as a machine deserving the engineering time of an Apple laptop focusing on desktop needs.
I will try to think of a test for this. As I've stated my opinion on the mac pro was that it was a downgrade of the 2008 machine from the processor aspect. The 2010 one had a decent gpu option though. Now when you do run longer tasks on it, that mac pro really isn't very fast, and it's a horrendous value. The gpu if you take the cto is about the only nice thing in it. The hard drive bays aren't even well designed in terms of cooling unless they fixed that on the 2010 one. The 2009 it was still really uneven.
These days a Mac Pro is really of value only when you can employ all of the possible cores or exploit it's bandwidth capabilities. Unfortunately for many users not all software can do that these days, thus an iMac can look really good if an all in one is viable.
If you have the link I'd love to see it. I've looked this stuff up and they seemed to be on shared total bandwidth but using separate protocols. I couldn't find anything from intel indicating that they were separated in bandwidth. It seems pretty high considering the peak bandwidth of displayport 1.2 is roughly 18Gb/s total. I'd like to know what kind of specs they were getting for throughput + what the display resolution was, and if this setup supports ddc based calibration as many people who would purchase such a display need that. On mine it goes through a shielded usb cable but not many operate that way.
You get two channels supporting 10GB/S each way. What I'm not clear on is the transmission of the Display Port signal and data on the same channel. It does look like bandwidth is shared in some Intel graphics I've seen while others imply Display Port being switched onto a channel.
More reading is in order but it does look like multiplexing of data is going on to some extent.
I don't have a lot of trust in intel (or Apple these days for that matter), so I'm waiting to see here. It's not a matter of being dismissive, but rather I think they're hyping it. I also remember how long Apple retained the Core2s in one of their laptops to avoid intel chipsets. I don't think that'll happen again, but it would be easy for them to go the mac pro route on this. Raise the price again (they've adjusted mini price points before) and downgrade their relative choices of processor grades for the machine. Mac pro used to start with mid range dual socket versions. Now they have the cheapest budget model. This is Apple. They've done this kind of thing before because they figure that people will grumble and buy it anyway.
I fully believe Ivy Bridge will be better. The problem I have is that Apple will cripple the processor in the Mini just like they have with every other Mini release. Well that and I have no confidence in dual core technology, if the Mini goes IB it really needs a quad core as a minimal implementation.
I'd like to comment on this. The mac pro is one of the larger towers of its type. This design was originally to accommodate the G5 processors with their massive heatsinks and the accompanying logic board design. In 2011 it looks like a dinosaur. Regarding aluminum it's not that expensive of a material. They aren't using a terribly high grade here so it doesn't contribute as much to the cost as you anticipate. In fact much of the cost of aluminum is often from processing it. The macbook airs and pros use a cnc process which should in theory be more labor intensive than the mac pro case. Beyond that they've provided machines using this form factor at $1500 and $2000 during the G5 era into the first generation of the mac pro. Right now the top macbook pro processor available to the 15" in a macbook pro is $568 retail. That machine is still cheaper than a mac pro with the inclusion of the thunderbolt chip. The 27" imac has the cost of a 27" ips panel (not the most expensive one of all but none of the 27" ones are dirt cheap) and uses a processor around $300.
A Dinosaur it is!!!! It is way to big to implement a mid range computer in. It has it's place but it isn't a place many people go to these days.
That mac pro you mentioned is artificially inflated. Doing anything to turn it into a better volume machine including a redesigned form factor and up to date setup would encourage volume and potentially a much better deal on the machine. Remember what Apple charged for the original macbook air? It was grossly overpriced relative to its capability, and ended up with dismal sales. Also what do you really gain from single socket Xeons anyway? Going with normal desktop i7s that are updated more frequently would be better for a single socket machine.
It's already been suggested that there are newer more compact methods of implementation for this stuff. It's just a matter of what Apple wants to use.
Whatever they use needs to be an industry accepted standard.
Really a highly portable machine with 1-2 high bandwidth slots could sell exceptionally well to photography and cinematography markets. These are industries that often lug mac pros along for location work (I'm serious, if anyone questions it I'll bring up links to companies, pictures of capture carts, etc).
I'm glad I'm not in that business!
However we do use PCs on carts. Sometimes it is just easier to have everything on a cart that passes as a mobile workstation. It is a place for cables, connectors, cameras, I/O boxes and the like.
Some use laptops for lighter jobs but they suck. A lighter form factor here would be amazing for these kinds of guys. Rackmountable configurations could be good if there are any Xserves still in use. I realize it wouldn't be the most efficient 2U kind of setup, but it could add more market potential to such a machine.
One day, if it ever takes off, TB supporting Macs may allow for that compact system. It is just that that day is a ways off from what I can see. In the end though I don't think your carts will go away. The computer might be smaller but the cart solves other issues. At work we often use the same sorts of carts to support laptops.
It's quite a difference from the mac mini with a 45W processor, but the quad core mini isn't the same thing. Like I said it's not even close to the imac processor which is around 90W. By your logic on Ivy Bridge if power consumption drops enough, something similar to what is in the imac today might start to look appropriate for the mini. I'm not so optimistic on it personally.
I'm convinced that power levels will drop enough to significantly upgrade the Mini. It really isn't a point to be argued as that is what the industry has done for years. The problem is Apple as they seem to purposely under power the Mini to the point that it doesn't even keep up with their own laptops.
Even if intel comes up a bit short with IB and the rumored and not so rumored massive power reductions we will still see an improvement. The other problem Intel has is with the GPU, they may have to start allocating a lot more transistors to that half of the chip. This will likely be driven by ant success AMD has with Bulldozer based Fusion chips.
I think we need to get away from the idea that Apple built a performance machine in that 2010 baseline mac pro. For lighter computing it feels fast enough, but anything does these days if you've used computers for more than a decade. You have some really good choices at 65-95W and a couple okay ones at 45W. If Ivy Bridge surprises me and intel does actually come through, we may see see some great choices in smaller thermal envelopes. Regarding power, the mini still has slower hardware than what is available on the macbook pros and imacs. The imac is by no means a high end machine. The starting mac pro is also not a high end machine. Apple can't declare that they are simply by pricing them high.
It is a performance machine just not in the way you or I want.
Configurations like this have been fit into mini towers for years. Some of them really do suck, but not all are that way. In terms of heat Apple has quite a bit of experience trimming things down. Let's talk about thunderbolt for a minute here. It has a pretty limited ability to supply power. Currently it still lacks the throughput of solutions like SAS cards. I'd really like to see that change to the point of where it could start to edge out legacy ports. Even then PCI is still there. There are still some ways it could be implemented in the future. As mentioned SSD storage cards at some point in the future could provide a better means of internal storage. Presently a move like going for 2.5" bays could cut a minor amount of space.
It is all about cooling. More importantly a tighter enclosure can put any cooling capability it has to better use. Far less of your air flow gets used to move dead air. moving away from legacy devices and concepts further enhances cooling.
If cooling was a critical issue for Apple they could build upon the concepts used for Compact PCI and other card based systems. There is no reason to maintain old PC approaches to the age old issue of thermal design.
If Apple is planning to completely ditch optical drives, that would cut further room needed. That's one I've always been a bit hesitant on because I still use it for a few things even though I abhor optical media. Downloads just aren't truly instantaneous and a lot of offices still have crappy methods for intra office file sharing of large documents. So anyway I'd like to see them either roll the low end of the mac pro line into a more cost effective line that can at least perform in line with the imac, or bring up the performance on the low end of the mac pro to a truly suitable level for its price point.
No matter what they make, an XMac or new Mac Pro they need to build for the future not the past. Building for the future permits innovation.
The director of Thunderbolt engineering says 40Gbps per port (20 up, 20 down):
http://www.pcmag.com/article2/0,2817...id=mxopWpE3s3y
Demoes like this one show 13Gbps:
http://www.youtube.com/watch?v=gk69p...eature=related
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
I'd like you to first consider that the guy in that video has the most awesome job title ever created.
Everything I read previously suggested it was shared bandwidth. That would make it considerably more awesome. I do understand that the display is downstream only. We have a couple things that have been emerging over the past couple years. Display panels have been increasing slowly in resolution per inch and 10 bit out is starting to see software and gpu support. Regarding SSDs the newest ones are extremely fast. That's probably encouraged the emergence of them in the PCIe form factor as just a couple can quickly saturate a SATA channel. The SATA standard overall has been around for quite some time. I wouldn't be surprised to see it replaced by the next thing if prices fall on solid state technology. The SATA/HDD form factor was really never ideal for mobile technology although it's been fine in practice. The 8x 1080p streams comment is a case of where you are too trusting of manufacturers
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
It hasn't always seemed this way, and at times the release cycles seem a bit weird. Westmere focused more on ultra high end chips and now we haven't seen a replacement in the Xeon line for well over a year (counting from the time of intel's release, not Apple's). Just going by benchmarks never helped me. There are just too many factors that require more in depth testing.
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
It'll be interesting to see. There are some definite plateaus to this stuff when it comes to justifying upgrades. When tasks are able to run over a lunch break instead of overnight (say they would have taken 3 hours instead of 1) or overnight rather than over the weekend, or even when something decreases the amount of dedicated server hardware required, it can make a difference. For things involving responsiveness it's just a matter of whether you feel like you're having to stop and wait for the system, or required to taper down settings to avoid this.
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
They haven't so far. They come closer in some tests than others, such as on lighter tasks, but they're still behind both the imacs and laptops Regardless I'd figure out some kind of test to run on the best one available. I've read up on every test I can find. I've run some very basic testing on each in the Apple store. I expect that with more ram the difference would only increase.
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
The 12 core is a totally different beast from the others. For the price points Apple sets, the starting machine in that line should have used something like a w3670. Sure it's more expensive than what it is in it currently, but it would have maintained a clear demarcation in performance when you move up to a mac pro and it's well within reason for a machine of this price range. Right now it's like you take a moderate dip in performance then have to pay more just to get back to where you were, and nothing about the machine warrants it. Big heavy aluminum cases don't account for the vast majority of that price. nor do the internals.
The 6970M (even the 2GB version) is still significantly slower than the 5870 in all possible uses from gaming to rendering (After Effects, maya, etc). Anyway the 6970M is basically a compromise for the enclosure style of the imac, but it's better than what many of the older imacs received.
May I ask what suggested the current cpu power of the mini as being close to the i7 2600 (can't recall if they use the k version) in the cto imac? Everything I've found suggests outside of lighter computing that doesn't actually saturate the bus, the unit in the imac is pretty far ahead.
The director of Thunderbolt engineering says 40Gbps per port (20 up, 20 down):
http://www.pcmag.com/article2/0,2817...id=mxopWpE3s3y
Demoes like this one show 13Gbps:
http://www.youtube.com/watch?v=gk69p...eature=related
but it's 6-7Gbps up and down - thing is, display data will only go downstream. A real-world test would have to be pushing multiple uncompressed streams out while also writing to a RAID drive. I haven't seen a stress-test yet so it may not live up to its theoretical maximums but you'd need an SSD RAID drive and about 8x 1080p streams running at once to check.
I really need to find the Intel documents because my understanding is that the data is multiplexed onto the channels. This would make sense due to the requirement that the monitor is the last item on the chain.
They pretty much always demonstrate 100% speed increase every 2 years so this year shouldn't be any different. The only thing that may be exaggerated is how much power they save. I'm not expecting half but even 25% saving is significant enough.
Baloney. At best Intel has been indicating around 20 to 25% speed ups over the last couple of years. Speed ups not always realized in real life. I don't ever recall Intel saying their chips are 50% faster from one model year to the next. Not overall at least though some obscure functions might be. Some years you are lucky to get 10% speed increase in general performance.
Previously, there hasn't been a mobile technology as fast as Thunderbolt and laptops didn't have 4-core, 8-thread CPUs. I suspect the more that people become aware of the setups available, the less you'll see the Mac Pros being hauled around. The tower doesn't need to change, people do.
TB is all well and good but right now the world is stuck with PCI cards. On top of that TB adoption in the PC world is spotty at best so many of those cards will stay PCI for a very long time. Beyond that if the laptop needs a disk array for storage you might as well haul around a Mac Pro.
They will still use the mobile versions of the chips but they will perform the same as the iMac chips.
History does not support your position. I really don't see how you can keep saying things like this.
There isn't a Sandy Bridge i7 desktop chip higher than what's in the iMac AFAIK and the 6970M GPU is very powerful. I'd personally call the iMacs high-end consumer-level machines.
No, they won't match the 6, 8 or 12-core Mac Pro but in terms of price, they can't really do much better.
I don't think anybody here dismisses the idea that the iMac is a good value. For many it is an excellent platform. A limited platform it is though as you again focus on CPU performance as the only metric by which to judge a machines value.
I really need to find the Intel documents because my understanding is that the data is multiplexed onto the channels. This would make sense due to the requirement that the monitor is the last item on the chain.
I looked for data on this. I couldn't find anything conclusive.
Baloney. At best Intel has been indicating around 20 to 25% speed ups over the last couple of years. Speed ups not always realized in real life. I don't ever recall Intel saying their chips are 50% faster from one model year to the next. Not overall at least though some obscure functions might be. Some years you are lucky to get 10% speed increase in general performance.
If you're comparing comparable chips from one year to the next, I haven't seen any truly amazing speed gains from one year to the next in a while. As mentioned the mac pro if you concentrate on roughly the same price point has been in stasis for a long time. I'm not going to compare what you got for $2800 (less from some of the online retailer) in 2008 to what the 2011 12 core monster at $5000. Aside from the gpu, performance at the sub 3k range hasn't budged in these machines since 2008. Slow sales shouldn't be a surprise at that point.
TB is all well and good but right now the world is stuck with PCI cards. On top of that TB adoption in the PC world is spotty at best so many of those cards will stay PCI for a very long time. Beyond that if the laptop needs a disk array for storage you might as well haul around a Mac Pro.
I've seen some things being built into new form factors. As long as it's a well developed solution it could work. I'd still like to see the performance levels of a stress tested thunderbolt connection.
History does not support your position. I really don't see how you can keep saying things like this.
Westmere was only a moderate gain over the original Nehalem in similar configurations. It was also a die shrink year. Lowered power consumption could allow for a beefier chip within the same thermal spec in the mini, but that doesn't seem to be suggested here. I've been quoting cpu specs because they're something easy to identify and can represent one of the significant cost factors in construction and a significant factor in performance assuming they aren't bottlenecked by something else.
I don't think anybody here dismisses the idea that the iMac is a good value. For many it is an excellent platform. A limited platform it is though as you again focus on CPU performance as the only metric by which to judge a machines value.
They have a good cpu, an ok gpu, four slots for ram, and now thunderbolt which should be more useful by next year (too little stuff made for it currently). I dislike having to buy the display as one unit. It can make sense but it's a very expensive forced purchase. I do think that for a lot of people it's still a better choice than the mac pro, and it works out pretty well for someone who would have otherwise purchased an external Apple display. I've never been a fan of Apple displays. They have a nice aesthetic but they've always lacked quality control. I'm still waiting to hear what is so great about glossy. If you hate the sparkle seen on some ips displays it's an LG thing. They use an obnoxiously strong anti reflective coating although it tends to help against color bias from weak reflections. On a consumer display a weaker AR coating can be applied which won't have the sparkle issue but also will not show reflections unless they're taking a lot of direct lighting from a lower angle.
It would be shocking to find a Mini effectively running as fast as it's contemporary iMac, with the same number of cores.
I meant the next Ivy Bridge quad Mini would match the current high-end iMac, I worded that a bit ambiguously earlier.
Here are the Cinebench benchmarks for the Mini, iMac and Mac Pro:
http://www.barefeats.com/mini11_01.html
http://www.barefeats.com/imac11b.html
http://barefeats.com/wst10.html
They show 30-60% faster for the top iMac vs the top Mini, which Ivy Bridge will make up next year. The entry Mac Pro is only 5-20% faster. My point is that if people can use these machines as desktops, the Mini is perfectly capable of the same tasks.
No I don't expect Apple to hit those low prices, but it clearly indicates that my price range is viable.
Other manufacturers ship desktop towers in much larger volumes.
What is it with PCI slots that is a problem. Engineers can put three of them into a 1U server box these days. it isn't a big deal.
You wouldn't put GPUs in a server box though and people would try to in a desktop and it won't handle the heat in a 1U space or the power usage.
I really don't know why you are so intent on selling dual core IB chips. It is a step backwards especially if your software is heavily threaded or you are in the habit of running lots of threads.
They work like 4-core chips. The dual-core 2.7GHz Mini performs around the same as the quad 2.5GHz iMac.
I've stopped laughing now. The response above obviously indicates that you don't get it.
I don't see why you want to have PCI slots in a machine for ethernet and other comms ports when you can plug a tiny adaptor inline and get the same functionality. You can have a USB modem for fax, TV tuner, audio capture device, ethernet using USB but you'd rather have 4 giant PCI slots making the whole machine 30% bigger. In that respect, your'e right, I don't get it. I don't get why you want something in a PCI form-factor instead of something the size of your thumb.
The 6970M (even the 2GB version) is still significantly slower than the 5870
The 5870 is under 50% faster. You both seem to consider something 50% faster to be significantly faster. I would say that 100% faster is significant. If you are playing a video game, 50% is the difference between running at 20FPS and 30FPS. All you have to do on the lower one is reduce anti-aliasing or resolution.
You also both keep mentioning the capabilities of the highest-end parts but we're talking about a mid-range machine. Yes, they can take the BTO quad i7 from the 2nd iMac and put it in a small enclosure along with 4 RAM slots and 2-4 HDD bays but they can't get a double-wide PCI slot in it supporting a 200W 5870 GPU too.
This stuff as you said will plateau like all computer hardware. Just look at games consoles. Now that they've hit a certain performance level, nobody cares about the upgrades any more, it's all about software.
I agree that there are more powerful designs now but Apple doesn't build new products knowing they have no future. They build products that start out low and have growth.
For years, I wanted the same kind of mid-range machine described in this thread and I've seen the underpowered hardware come and go but for me, the Ivy Bridge Mini brings it to an end. If the middle one breaks 9k in Geekbench and bundles a Radeon 7000 series GPU and has USB 3, I just don't see the reasons for building a bigger machine.
If people want good price/performance, they need to get an over-clocked PC dedicated to rendering and/or gaming. If they want a stylish, stable, powerful, everyday machine, get a Mini.
If people want good price/performance, they need to get an over-clocked PC dedicated to rendering and/or gaming.
Word.