Intel will come around eventually though they are really going to have to up their game if they want to complete with discrete graphics cards from nVidia and AMD. With all the money they are spending in research, I have my confidence in them that they will make great strides and we'll probably see something amazing by Broadwell or Skylake or Skymont. Haswell will be a nice jump too but not a big enough one.
Intel certainly is one company I'm positive on as Apple themselves is kind of making me annoyed with all the bs on the lawsuits, patents, and iPhones. I want more Mac news.
The people who build their own machines don't represent a large enough market. iMacs have fast GPUs, IPS displays, support up to 32GB RAM like any i5/i7 CPU. Thunderbolt makes up for the lack of PCI slots no matter how much people want it not to. Thunderbolt isn't just a fast connection, it is a multi-protocol PCI port, they just put it on the outside. If Apple built consumer towers again, it would just be wasted effort.
In what way? You're still limited by the drive speeds and it's not a good idea to put loads of drives in RAID 0. Thunderbolt speeds are more than adequate for anything storage related.
eSATA is very limited - it's just a storage protocol and doesn't officially support power. It will die out in favour of USB 3.
They didn't care initially but the unibody redesign made it obvious they do. The latest Mini is a work of art and it's perfectly suitable for the vast majority of desktop use. With 16GB RAM and an SSD, the only area it is lacking in is graphics but even then, the dGPU model performs ok.
It's easy to imagine Apple offering a quad-core desktop with an option of a high-end desktop GPU and PCI slots just like every other manufacturer for $500-1000 but Apple builds iconic computers and this inexpensive desktop couldn't be iconic because there's no way to simplify its design to epitomize its purpose without still falling short of the requirements of the target audience.
When you look at an iconic machine and wish that it was built the way everybody else builds them, you're missing the point about why they are built that way in the first place. Make Marilyn Monroe a brunette, take off the beauty spot, put her in a drab outfit and you have a nobody. That's what you are asking Apple to do with the iMac.
At this stage in time, there is no missing model in Apple's lineup, they just have one model too many. One that doesn't fit the pattern of Apple's design and it will be made to fit. When this happens, it will make it clear once and for all.
The story about eSATA is that it was much faster than anything else Apple was offering and Apple was stagnant, if not outright resistant to improved technology that was mature. Yes, today it is slower than Thunderbolt. Thunderbolt is not what you seem to think it is, however. Thunderbolt is a subset of external PCIe. It has two lanes, one each way. That is a lot less than is available from a proper internal PCIe slot. (e.g. 4 or 8 lanes if your board only has one 16 lane slot for the graphics card.) Additionally, there are two Thunderbolt chips that must process the data and the internal distance as well as the TB cable and any internal distance involved in the external device. Does that mean that TB is not useful? Of course not. It is probably the best available external interface at the moment.
All I will say about Apple design is that if it were more functional it would have a beauty all its own. Ditto as to thermal management. The signature element of Ive's designs is that they run hotter than necessary because of his design work. No, I am not a fan of his, even if Queenie did tap him on the shoulders.
Somewhere there was a remark that there are not enough Hack users to worry about building a machine for. Strictly speaking that may be true, but for every one of them how many more people are there who would buy the machine from Apple even though they would not go to the trouble to put together a Hack? I think it would be a large number. There is still a place for a desk side computer, especially for image processing.
As far as there being no missing model in Apple's lineup, you must work for Apple and/or be drinking the cool-aid because there are any number of missing models. One that was discussed just this evening at a meeting...a 13" MBP quad core with anti-glare screen. It's pretty obvious. There are a significant number of people who believe the 15" MBP to be too large and too heavy who, nonetheless want the other feathers and capabilities.
My point regarding RAID 5 was that it's definitely more finicky when it comes to consumer grade solutions. Raid 0, 1, and 10 may still be feasible even without enterprise firmware. Just grab caviar black drives + roc card or eSATA host card and software RAID. Make sure eSATA card and port multiplier chipset are made by the same manufacturer so you don't run into conflicts. My point was that when you started getting into parity stripes or dedicated parity drives (basically anything that splits off parity whether it's byte or block level), hardware and setup requirements become more stringent if you want an absolutely stable rig. If you're dealing with enormous archives of data, you don't really have to keep all of it on such a solution. Anyway this isn't really my area of expertise. I just know what to avoid in most circumstances, and built in Raid 5 functionality is not really something you'd see in a consumer level desktop.
I didn't so much mean this. My issue with the discrete graphics is that they still do not perform well, and their vram allocation misses the minimum mark for many things. It doesn't need a high end solution, just a better solution. Improvements to integrated graphics may eventually turn this aspect into a non issue, but it's a design flaw today. It annoys me somewhat that the mini is still languishing on this update.
Aside from an actual RAID array controlled from the computer, there are a variety of NAS RAID arrays that are useful even if they are only run through a Gigabit Ethernet connection which is doable from a Mini. TB NAS is good as well. I would not call a RAID 5 array an outright "consumer" feature, but there are a good many enthusiasts who want it for speed and safety...your data will survive the loss of one drive, but that is no substitute for a proper backup plan. RAID 5 is popular, comparatively speaking, with the image editing crowd and, no doubt, others. Sure, (a minimum of) three WD Black RE 2 TB drives for a RAID 5 array is not the cheapest solution in the world, but it offers a lot of capability. Having it internally not only has performance advantages over an external setup, but avoids the additional expenses associated with many external systems.
Though your remark about vram for discreet graphics cards was not directed to me, I quite agree with your opinion. Sadly, Apple seems to make a habit of languishing, as you put it, on comparatively straight forward updates of components.
The story about eSATA is that it was much faster than anything else Apple was offering and Apple was stagnant, if not outright resistant to improved technology that was mature.
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
Thunderbolt is not what you seem to think it is, however. Thunderbolt is a subset of external PCIe. It has two lanes, one each way. That is a lot less than is available from a proper internal PCIe slot. (e.g. 4 or 8 lanes if your board only has one 16 lane slot for the graphics card.)
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB latency is 9ns so not a problem.
Somewhere there was a remark that there are not enough Hack users to worry about building a machine for. Strictly speaking that may be true, but for every one of them how many more people are there who would buy the machine from Apple even though they would not go to the trouble to put together a Hack? I think it would be a large number. There is still a place for a desk side computer, especially for image processing.
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
As far as there being no missing model in Apple's lineup, you must work for Apple and/or be drinking the cool-aid because there are any number of missing models. One that was discussed just this evening at a meeting...a 13" MBP quad core with anti-glare screen. It's pretty obvious. There are a significant number of people who believe the 15" MBP to be too large and too heavy who, nonetheless want the other feathers and capabilities.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
It was always faster. I adopted it in 2005 (I think 2005) or so. At that time it was relatively good for large amounts of cheap storage with tolerable speeds. Even the early cards outpaced firewire in desktop settings. For laptops it caters to more of a similar market to the thunderbolt display, meaning people who primarily use notebooks. I never expected thunderbolt to take off quickly. It's kind of a weird implementation of something that the line needed long ago. I call it weird due to the need to piggyback display data even though this can tie up 2 channels. It's not the most efficient way to do things.
Quote:
Originally Posted by Marvin
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB
Okay this is more Marvin kool-aid. You post some great stuff at times, then you make up numerical claims that exceed intel's best case scenario for their own product roadmaps. Thunderbolt's actual speeds have nothing to do with PCIe 2 or 3. It's using 4 PCIe 2.0 lanes. Unless they update the chips, speed stays the same. Initial optical cables aren't fully optical anyway, and their purpose is just to get around length restrictions. Intel previously claimed it could hit 100Gb/s within a decade, but it comes down to how much bandwidth people are willing to pay for. They never claimed what you are claiming.
Quote:
Originally Posted by Marvin
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
"From the standpoint of profit margins, the worst performing part of IT has been PCs and it's getting even worse"
It's not a good idea to run in a direction you see everyone else running away from, especially the biggest players.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
That article mentions they should focus on high margin areas. Look at the pricing on some of their Z line. If they're still moving workstations, at least one part of their PC offering is doing well. Much like Dell they offer a lot of configurations, but with HP if you wish to order anything non standard, the markup is quite high. They're either moving mid priced workstations in stock configurations or custom ones with high markups on the configuration aspect.
Intel previously claimed it could hit 100Gb/s within a decade, but it comes down to how much bandwidth people are willing to pay for. They never claimed what you are claiming.
That should have been 50Gbps for 2015 although it might be dual 50Gbps channels in both directions but still 50Gbps per channel in one direction:
That article mentions they should focus on high margin areas.
Which is not the consumer tower market. If you sell a product with a high margin, it needs to have a selling point. A consumer tower is the same as everyone else is selling, how does Apple command a premium on it?
The other issue that also affects the Mini is what display do they bundle with it? Their $1000 27"? So they'd be selling a quad-core i7 box with up to 32GB RAM for $1000 along with a $1000 27" display and a $69 mouse and a $69 keyboard = $2138.
Hmmm, let's see how much that quad-core i7 iMac with the 27" display is that takes up to 32GB RAM and comes with the keyboard and mouse: $2200.
Now, obviously you could buy a 3rd party display with the tower and you get your PCI slots and your multiple drives but most consumers don't care about this stuff any more and eventually nobody will care about this stuff just like nobody will care about optical drives being taken out of everything.
The way Apple makes good money on the iMac is by selling you a display with your computer because they make the markup on the display. Take the display out of the equation and all they are doing is giving that margin to another display manufacturer.
Of course, that would suggest they get rid of the Mini but they can't do this as they can't get the iMac down to that price and it would cut out all the usage scenarios of the Mini. That's why the Mini will stay in the lineup. They will try to make the iMac more compelling but it won't matter, the Mini will keep increasing in power.
By the way, I realize that I jump around on comparisons. I haven't always been happy with the quality of Apple's mobile devices, so I tend to favor their stationary form factors.
The main point is, whatever people claim the bottlenecks are now, they will be resolved within the next 3 years.
Which is not the consumer tower market. If you sell a product with a high margin, it needs to have a selling point. A consumer tower is the same as everyone else is selling, how does Apple command a premium on it?
You should know enough not to believe intel's promises. It's still 3 years away anyway, and I still think the idea of external gpus is a dumb one. If you look at the 7970m, it's quite powerful, yet the tdp is way down. Pretty soon such a thing will not be worth it as many of those external pci enclosures can't hold the hottest or largest gpus anyway. The bottlenecks aren't really a claim. You can't hook a display + other stuff on a thunderbolt connection unless it's the Apple display without throttling as display data and regular data cannot share a channel. I don't know why you keep clinging to the idea that you can take a mac pro and reproduce it there. Even the mac pro is IO constrained at the single package level. In the past they've simply assigned more theoretical bandwidth than what really exists. It's like if you look at crossfire setups, you can't give each of them 16 lanes, as the PCI lanes simply aren't there in those chipsets. On the Xeon end they've been moved into the cpu package, so if you're in need of any kind of extreme bandwidth, your only option would be a dual package model.
Quote:
Originally Posted by Marvin
Hmmm, let's see how much that quad-core i7 iMac with the 27" display is that takes up to 32GB RAM and comes with the keyboard and mouse: $2200.
Now, obviously you could buy a 3rd party display with the tower and you get your PCI slots and your multiple drives but most consumers don't care about this stuff any more and eventually nobody will care about this stuff just like nobody will care about optical drives being taken out of everything.
The way Apple makes good money on the iMac is by selling you a display with your computer because they make the markup on the display. Take the display out of the equation and all they are doing is giving that margin to another display manufacturer.
Of course, that would suggest they get rid of the Mini but they can't do this as they can't get the iMac down to that price and it would cut out all the usage scenarios of the Mini. That's why the Mini will stay in the lineup. They will try to make the iMac more compelling but it won't matter, the Mini will keep increasing in power.
I'm well aware of their strategy to sell you a display too. I can't use the imac or thunderbolt displays. They're too bright, too reflective, and lack profile stability. I realize this won't be a big deal to everyone. Given everything I require on Linux, I'd switch immediately. There are just some things that don't run or don't run well. I do think they've made a few mistakes on the mini. Internal power supplies make the thermals suck. A modern desktop should be silent under load whenever possible, and they're working against that. I think if they're going to add discrete graphics, they shouldn't go totally low end mobile on them. Apple markets on simplicity. They should be able to meet minimum spec requirements on basically anything in that year without checking (minimum requirements are generally not high, even for powerful applications).
RBR seemed to be going the direction of suggesting a small tower that could also function as a file server, although I'm not sure how that would work out. The problem with the desktop is that it has become stagnant, and the biggest percentage gains in performance have been seen at the low end. The Airs have posted some impressive gains. Desktops do offer some superiority in long term reliability. Just look at the number of complaints on Apple chargers and replacement batteries at the Apple store. They never had great reviews on either product as no one likes to replace those things. It's just that charger cords prone to twisting where you have to discard the whole unit, and batteries known to bulge (I've had it happen) when old are both significant issues. On a side note I hope they fixed that battery thing on the rMBP.
I still think the idea of external gpus is a dumb one. If you look at the 7970m, it's quite powerful, yet the tdp is way down. Pretty soon such a thing will not be worth it as many of those external pci enclosures can't hold the hottest or largest gpus anyway.
So you agree you don't need PCI slots. If you have a 7970M internally and Thunderbolt handles the IO, you're all set.
The bottlenecks aren't really a claim. You can't hook a display + other stuff on a thunderbolt connection unless it's the Apple display without throttling as display data and regular data cannot share a channel.
That's just the usual vague fearmongering. You can't attach devices that use up all the bandwidth without seeing a performance drop, that's obvious and applicable to any setup. This has been tested out:
You can support two Thunderbolt displays and at least two peripherals on an iMac so 3 x 27" 2560 x 1440 + 2 Thunderbolt devices. In 2014, you will be able to add more without certain devices running at a lower speed.
I find it odd how people want an affordable tower and then suggest it's necessary because the iMac couldn't handle setups that require you to spend over $4k to saturate. Either you want an affordable setup or you don't. If you do, the iMac is sufficient and pretty decent value.
so you end up at $1580. If Apple sold that machine, they'd make $220 vs $550 on the iMac and yet as far as the consumer is concerned, they are spending comparable amounts (a savvy buyer wouldn't go for the cheapest brands).
It's like if you look at crossfire setups, you can't give each of them 16 lanes, as the PCI lanes simply aren't there in those chipsets. On the Xeon end they've been moved into the cpu package, so if you're in need of any kind of extreme bandwidth, your only option would be a dual package model.
Who has Crossfire or SLI in their Mac Pro? This is a setup for gaming. Plus you get a 300W limit so what cards are you going to use? A single high-end GPU uses that.
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB latency is 9ns so not a problem.
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
"From the standpoint of profit margins, the worst performing part of IT has been PCs and it's getting even worse"
It's not a good idea to run in a direction you see everyone else running away from, especially the biggest players.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
I am not sure what your problem is with eSATA. I have used it and it is far and away superior to USB 2 or FW 800. I find it inexcusable that Apple failed to adopt it. It may not have been "perfect" in a mobile platform, but it was/is still quite useful. OWC even had a mod for the iMac that simply put an eSATA port in it and connected to an unused SATA port on the logic board. It is not rocket science.
I will say this one last time. Thunderbolt is an improvement over the other currently available external data transfer protocols used by Apple. It is not, however,the equal of, or a replacement for, an internal PCIe. The transfer rates speak for themselves.
It is true that the market share of laptops/tablets/mobile devices has been growing for quite a few years and probably will continue to grow. That is not to say that there is not a place for something other than a laptop. You fail to comprehend the concept of offering a mix of products so that the company has a solution for more than one need. Indeed, you contradict yourself in regard to the 13" MBP. It really should offer a better CPU and screen. Many people prefer its size (and weight) over that of the 15" MBP, but want a better machine.
Apple is coming to a fork in the road where any number of long term users are abandoning the platform because of the lack of support for even the hardware they have been using. If Apple wants to concentrate on iPhones & iPads because they are the most profitable, that is the company's choice, but they should not be surprised if that makes the company an increasingly niche player, even if it is a large niche player.
so you end up at $1580. If Apple sold that machine, they'd make $220 vs $550 on the iMac and yet as far as the consumer is concerned, they are spending comparable amounts (a savvy buyer wouldn't go for the cheapest brands).
Who has Crossfire or SLI in their Mac Pro? This is a setup for gaming. Plus you get a 300W limit so what cards are you going to use? A single high-end GPU uses that.
And there's the explanation. You might as well give up the idea of switching if you aren't prepared to let go some of the legacy baggage.
It's not any louder than the one with the external PSU.
Hardly, PC manufacturers don't make quiet machines. The iMac is almost silent even under load.
My main problem with the imac is long term thermal issues. The display also doesn't age well because they cheap out on circuitry. It has all of the signs of a hack job if you view one that's more than a year or two old even if it's been used lightly. The number of aging imacs I've seen with purple splotches is ridiculous. I would go as far as to suggest that the only potentially reliable Macs they're still making are the mini, the pro, and the Air. I've had and seen issues with the rest at one point or another. I've owned a number of Macs, then between everyone I've known who owns them, it adds up to dozens of devices. The point being I encounter them constantly. Anyway my point on crossfire was that they oversubscribe the lanes there. It's not really 32 lanes. It wasn't a why don't macs have this kind of issue.
Quote:
Originally Posted by Marvin
So you agree you don't need PCI slots. If you have a 7970M internally and Thunderbolt handles the IO, you're all set.
That's just the usual vague fearmongering. You can't attach devices that use up all the bandwidth without seeing a performance drop, that's obvious and applicable to any setup. This has been tested out:
You can support two Thunderbolt displays and at least two peripherals on an iMac so 3 x 27" 2560 x 1440 + 2 Thunderbolt devices. In 2014, you will be able to add more without certain devices running at a lower speed.
I find it odd how people want an affordable tower and then suggest it's necessary because the iMac couldn't handle setups that require you to spend over $4k to saturate. Either you want an affordable setup or you don't. If you do, the
Huh? I mainly want a better mac pro. Regarding the xmac thing, I mostly wanted to be part of the conversation there. Thunderbolt creates more problems than it solves on a mac pro, which is why I don't care if it gets it. You need a displayport connection on the edge of the board or you end up with some weird rewiring. The thunderbolt displays are the only chainable displays, and at their pricing I'd rather have two of the 27" NECs as they're just better displays than Apple has ever made. Anyway other displays haven't tested well when it comes to chaining, and it does slow down the other devices. You can't run 2 displays and the TB raid at full performance specs here. That is oversubscribing the bus as it's saturated prior to that point. If you're wondering why others want different desktop options, Apple bundles things to get you into a high minimum sale with the imac, yet they don't last. Computing needs for most people are not growing very fast at the moment (note most people, especially those who buy imacs and macbook airs), so it annoys me that they are not built to handle reasonable wear. Others have complained about getting rid of the display after a couple years. I think it generally looks terrible long before the hardware is truly incapable, and I'm not fabricating this.
Quote:
Originally Posted by Marvin
Hardly, PC manufacturers don't make quiet machines. The iMac is almost silent even under load.
Have you ever looked at a modern PC? Most are silent, especially workstations.
I am not sure what your problem is with eSATA. I have used it and it is far and away superior to USB 2 or FW 800.
You can't run a Keyboard/Mouse, pen drive or video camera over eSATA. It's a storage protocol and one that doesn't supply power. It is a limited utility port. It doesn't matter if it happened to be faster than USB 2/FW800, USB 3 was always going to be faster and more useful. Now that we have USB 3, it is completely irrelevant. The issue is, we should have had USB 3.0 4 years ago.
If Apple had put eSATA onto machines 8 years ago, they'd be removing them today and people would be left with eSATA devices that didn't work. I think they made the right decision not to support it. It's good to avoid popularising technology you know in advance that you are going to deprecate.
Similar deal with Firewire - it was going to ramp up to 3.2Gbps or so but the whole standard is now regarded as legacy, same as Ethernet. Phil called FW800 and Ethernet legacy technologies when he introduced the retina MBP. That's why they are only supported with adaptors.
The technologies going forward are Thunderbolt and USB 3.
Thunderbolt is an improvement over the other currently available external data transfer protocols used by Apple. It is not, however,the equal of, or a replacement for, an internal PCIe. The transfer rates speak for themselves.
And when the ports are 50Gbps each, what then? Even if people feel TB isn't a good enough replacement now, it soon will be. That's all Apple is bothered about - backing the right horse. TB is an elegant way to connect high performance peripherals to a Mac, PCI slots aren't.
Indeed, you contradict yourself in regard to the 13" MBP. It really should offer a better CPU and screen. Many people prefer its size (and weight) over that of the 15" MBP, but want a better machine.
Right but it's not a new model so it's not the same as a mid-range tower vs an iMac. I think they will improve the 13" display and spec (more likely a better GPU than quad CPU) but eventually, it won't even be an additional option. It will become the only option. The Retina MBP isn't just a separate option, it will replace the old-style 15".
Apple used to sell PowerMac towers alongside the iMacs and the iMacs replaced them. They're not going backwards.
Apple is coming to a fork in the road where any number of long term users are abandoning the platform because of the lack of support for even the hardware they have been using. If Apple wants to concentrate on iPhones & iPads because they are the most profitable, that is the company's choice, but they should not be surprised if that makes the company an increasingly niche player, even if it is a large niche player.
The number of users who would abandon Apple are the niche players now. Apple has sold 365 million iOS devices so far. That's not a niche market.
My main problem with the imac is long term thermal issues. The display also doesn't age well because they cheap out on circuitry.
Some iMacs fail and some Mac Pros fail but a great number of iMacs are still running ok, even G5 iMacs from 8 years ago. I don't see any widspread evidence that Apple's small form factor machines have any long term issues due to their design.
You can't run a Keyboard/Mouse, pen drive or video camera over eSATA. It's a storage protocol and one that doesn't supply power. It is a limited utility port. It doesn't matter if it happened to be faster than USB 2/FW800, USB 3 was always going to be faster and more useful. Now that we have USB 3, it is completely irrelevant. The issue is, we should have had USB 3.0 4 years ago.
I do agree that eSATA was a waste and for the most part a poor investment. Apple has dragged its feet excessively with USB3 and has caught considerable flack for it. About the only thing here to debate is the 4 year term as it took a long time for suppliers to get the bugs out of the USB3 chipsets.
If Apple had put eSATA onto machines 8 years ago, they'd be removing them today and people would be left with eSATA devices that didn't work. I think they made the right decision not to support it. It's good to avoid popularising technology you know in advance that you are going to deprecate.
This I agree with 100%' especially with eSATA which frankly is a crappy standard.
Similar deal with Firewire - it was going to ramp up to 3.2Gbps or so but the whole standard is now regarded as legacy, same as Ethernet. Phil called FW800 and Ethernet legacy technologies when he introduced the retina MBP. That's why they are only supported with adaptors.
Ethernet is not a legacy port, it is the backbone of most business networks and more that a few home networks.
The technologies going forward are Thunderbolt and USB 3.
And when the ports are 50Gbps each, what then? Even if people feel TB isn't a good enough replacement now, it soon will be. That's all Apple is bothered about - backing the right horse. TB is an elegant way to connect high performance peripherals to a Mac, PCI slots aren't.
Right but it's not a new model so it's not the same as a mid-range tower vs an iMac. I think they will improve the 13" display and spec (more likely a better GPU than quad CPU) but eventually, it won't even be an additional option. It will become the only option. The Retina MBP isn't just a separate option, it will replace the old-style 15".
it will be very interesting to see how the Mac Book Pro line up pans out. I could see Apple giving people options for a very long time. Retina is easy to market to simpletons but for challenging users may opt for the extra features of the conventional machines.
Apple used to sell PowerMac towers alongside the iMacs and the iMacs replaced them.
That is a most unfortunate turn of events. For many of us the iMac is a terrible solution.
They're not going backwards.
The number of users who would abandon Apple are the niche players now. Apple has sold 365 million iOS devices so far. That's not a niche market.
Some iMacs fail and some Mac Pros fail but a great number of iMacs are still running ok, even G5 iMacs from 8 years ago. I don't see any widspread evidence that Apple's small form factor machines have any long term issues due to their design.
The only serious issue I'm aware of is the failure of the home button on iOS devices. Overall though I'd have to say that for portable devices they actually have fairly rugged products.
Yeah even though the iMac has more horsepower per say, I like the display and computer separate as I have now.
Put in the Ivy Bridge processors, bump up the RAM, Intel HD 4000, and SSD. One model having the GeForce 640M LE would be nice as I said before. If not then let it be. Have the base model either remain at $599 or drop lower (without an SSD anyway if they get to $499) and if they don't make a discrete model, then drop the server to $799 or $899.
You can't run a Keyboard/Mouse, pen drive or video camera over eSATA. It's a storage protocol and one that doesn't supply power. It is a limited utility port. It doesn't matter if it happened to be faster than USB 2/FW800, USB 3 was always going to be faster and more useful. Now that we have USB 3, it is completely irrelevant. The issue is, we should have had USB 3.0 4 years ago.
If Apple had put eSATA onto machines 8 years ago, they'd be removing them today and people would be left with eSATA devices that didn't work. I think they made the right decision not to support it. It's good to avoid popularising technology you know in advance that you are going to deprecate.
Similar deal with Firewire - it was going to ramp up to 3.2Gbps or so but the whole standard is now regarded as legacy, same as Ethernet. Phil called FW800 and Ethernet legacy technologies when he introduced the retina MBP. That's why they are only supported with adaptors.
The technologies going forward are Thunderbolt and USB 3.
You know well enough that Apple spins things however they like in marketing. You may also notice that my posts sound more grumpy after recent hardware failures (in this case another swollen battery). A year ago people were doubtful over whether Apple would even release usb3 drivers and suggested that thunderbolt was the only thing needed. FW800 should have been legacy long ago. eSATA should have never made it off the ground. The reason both survived was due to lack of better solutions. A second thunderbolt port on the new ones is a huge improvement in itself. I haven't bought any firewire device in years that didn't also support usb just in case the firewire option was dropped and I needed to use it a bit longer.
Quote:
Originally Posted by wizard69
I do agree that eSATA was a waste and for the most part a poor investment. Apple has dragged its feet excessively with USB3 and has caught considerable flack for it. About the only thing here to debate is the 4 year term as it took a long time for suppliers to get the bugs out of the USB3 chipsets.
This I agree with 100%' especially with eSATA which frankly is a crappy standard.
Years ago it was somewhat of a makeshift solution. It was stable and could sustain 200MB/s~ easily enough depending on chipset. Keep in mind during the G5 era, you got two internal bays and hard drive capacities were far more limited. I'm actually in the process of phasing out older eSATA stuff. It worked back then. Now it's simply dated, but usb3 should have come out last year. They went super conservative as usual. Unless something carries a coolness factor, Apple is very conservative on what they add. It may be an unwillingness to take on construction or support costs.
I'm really hoping that within the next two days we will know what is up with the Mini. The desire is that it gets beefed up a bit if we are to remain without an XMac. Like you I have problems with the current iMac so the desire to see the Mini get a little respect is strong.
By the way I could see a redesigned iMac I could go for reluctantly. Right now there is no way. I would still be a little pissed at having to buy an all in one but it may be the only way to midrange performance on the desktop. Apple would have to address serviceability and a few other negative factors with the current iMac though.
In any event this is a Mini thread so let's get the train on track. Who here is expecting a new Mini in the next 2 days? If they don't come with Mountain Lion then I have to really wonder what the hell is up with Apple. I can't see any technology they could be waiting on for this machine otherwise.
Yeah even though the iMac has more horsepower per say, I like the display and computer separate as I have now.
Put in the Ivy Bridge processors, bump up the RAM, Intel HD 4000, and SSD. One model having the GeForce 640M LE would be nice as I said before. If not then let it be. Have the base model either remain at $599 or drop lower (without an SSD anyway if they get to $499) and if they don't make a discrete model, then drop the server to $799 or $899.
You know well enough that Apple spins things however they like in marketing. You may also notice that my posts sound more grumpy after recent hardware failures (in this case another swollen battery). A year ago people were doubtful over whether Apple would even release usb3 drivers and suggested that thunderbolt was the only thing needed.
Some people believe in the tooth fairy too. Personally I always believed that Apple was held up by Intel when it came to USB 3 support. I was under the impression that many engineers at Apple where frustrated buy the lack of USB 3 just like the rest of us.
FW800 should have been legacy long ago. eSATA should have never made it off the ground. The reason both survived was due to lack of better solutions. A second thunderbolt port on the new ones is a huge improvement in itself. I haven't bought any firewire device in years that didn't also support usb just in case the firewire option was dropped and I needed to use it a bit longer.
Somethings like eSATA crop up due to a failure to fill a need with other products. FireWire, in the projected faster standards, could have filled that roll but it was already on the way out as a standard. So eSATA slips in to fill a whole with regard to external connectivity. ESATA being dirt cheap to implement was likely a factor too.
Years ago it was somewhat of a makeshift solution. It was stable and could sustain 200MB/s~ easily enough depending on chipset. Keep in mind during the G5 era, you got two internal bays and hard drive capacities were far more limited.
Yes but my storage needs grow every year. It has gotten to the point that I'm thinking about a RAID of some sort for fat storage.
I'm actually in the process of phasing out older eSATA stuff. It worked back then. Now it's simply dated, but usb3 should have come out last year. They went super conservative as usual.
In this case I think they simply where cheap. Apple didn't want to bother with engineering drivers for a chipset for a year or two of use and didn't want to solder extra hardware on their motherboards. Third party chips became very stable and power efficient at least a year and a half ago, so there are no other technical reasons not to implement USB 3.
Apple might have had an arguement in that they would have to support those third party chips for a very long time. The problem with that is that they screw with the good will of their users. The need for USB 3 support is legitimate and to ignore that support for almost two years after viable solutions exist just infuriates customers.
Unless something carries a coolness factor, Apple is very conservative on what they add. It may be an unwillingness to take on construction or support costs.
Yep they are cheap. That and they seemingly have ignored the Mac hardware lineup for a long time. I mean really the Mac Pro was turned into an industry joke with the last so called update. Sadly no USB 3 there either. If the Mini doesn't update soon it will become another joke. Seriously selling a Mini these days without USB 3 support is so anti consummer that even Apple will start to suffer.
In any event this is a Mini thread so let's get the train on track. Who here is expecting a new Mini in the next 2 days? If they don't come with Mountain Lion then I have to really wonder what the hell is up with Apple. I can't see any technology they could be waiting on for this machine otherwise.
If not within the next week than possibly sometime in August. Either way, I cannot see them not releasing a mini at all in 2012.
I'm really hoping that within the next two days we will know what is up with the Mini.
The iMac has actually gone longer without an update, which is unusual. It makes sense for them to spread the releases out though as the next event will likely be the iPhone launch on the anniversary of Steve's death (October 5th), given that the last one was Oct 4th.
They wouldn't update the Mini before the iMac either and we know Mountain Lion is coming in July. They may as well do all 3 this week.
It'll still fall short of some expectations, I think it will be:
Core i5, HD4000
Core i5, GT 640M
Core i7 quad, HD 4000
4x USB 3.0, HDMI, single TB.
I wonder if they will drop FW800 from the desktop machines. I don't think it makes sense to drop ethernet as they have plenty of room but dropping FW800 will save some money and push adoption of USB 3. A 5th USB 3 port would be nice if they did this though.
The iMac is their best selling desktop so while it hasn't been left without a major update as long as their best selling machine - the MBP (473 days), I expect an update soon.
Ethernet is not a legacy port, it is the backbone of most business networks and more that a few home networks.
We sure install a lot of Cat 5 and Cat 6 cable and data jacks in new businesses, stores, schools, hospitals, etc for ethernet to be going away like some people think. Sure we install more wireless nodes now but we haven't seen any business willing to trust wireless completely.
Comments
Intel certainly is one company I'm positive on as Apple themselves is kind of making me annoyed with all the bs on the lawsuits, patents, and iPhones. I want more Mac news.
Quote:
Originally Posted by Marvin
The people who build their own machines don't represent a large enough market. iMacs have fast GPUs, IPS displays, support up to 32GB RAM like any i5/i7 CPU. Thunderbolt makes up for the lack of PCI slots no matter how much people want it not to. Thunderbolt isn't just a fast connection, it is a multi-protocol PCI port, they just put it on the outside. If Apple built consumer towers again, it would just be wasted effort.
In what way? You're still limited by the drive speeds and it's not a good idea to put loads of drives in RAID 0. Thunderbolt speeds are more than adequate for anything storage related.
eSATA is very limited - it's just a storage protocol and doesn't officially support power. It will die out in favour of USB 3.
They didn't care initially but the unibody redesign made it obvious they do. The latest Mini is a work of art and it's perfectly suitable for the vast majority of desktop use. With 16GB RAM and an SSD, the only area it is lacking in is graphics but even then, the dGPU model performs ok.
It's easy to imagine Apple offering a quad-core desktop with an option of a high-end desktop GPU and PCI slots just like every other manufacturer for $500-1000 but Apple builds iconic computers and this inexpensive desktop couldn't be iconic because there's no way to simplify its design to epitomize its purpose without still falling short of the requirements of the target audience.
When you look at an iconic machine and wish that it was built the way everybody else builds them, you're missing the point about why they are built that way in the first place. Make Marilyn Monroe a brunette, take off the beauty spot, put her in a drab outfit and you have a nobody. That's what you are asking Apple to do with the iMac.
At this stage in time, there is no missing model in Apple's lineup, they just have one model too many. One that doesn't fit the pattern of Apple's design and it will be made to fit. When this happens, it will make it clear once and for all.
The story about eSATA is that it was much faster than anything else Apple was offering and Apple was stagnant, if not outright resistant to improved technology that was mature. Yes, today it is slower than Thunderbolt. Thunderbolt is not what you seem to think it is, however. Thunderbolt is a subset of external PCIe. It has two lanes, one each way. That is a lot less than is available from a proper internal PCIe slot. (e.g. 4 or 8 lanes if your board only has one 16 lane slot for the graphics card.) Additionally, there are two Thunderbolt chips that must process the data and the internal distance as well as the TB cable and any internal distance involved in the external device. Does that mean that TB is not useful? Of course not. It is probably the best available external interface at the moment.
All I will say about Apple design is that if it were more functional it would have a beauty all its own. Ditto as to thermal management. The signature element of Ive's designs is that they run hotter than necessary because of his design work. No, I am not a fan of his, even if Queenie did tap him on the shoulders.
Somewhere there was a remark that there are not enough Hack users to worry about building a machine for. Strictly speaking that may be true, but for every one of them how many more people are there who would buy the machine from Apple even though they would not go to the trouble to put together a Hack? I think it would be a large number. There is still a place for a desk side computer, especially for image processing.
As far as there being no missing model in Apple's lineup, you must work for Apple and/or be drinking the cool-aid because there are any number of missing models. One that was discussed just this evening at a meeting...a 13" MBP quad core with anti-glare screen. It's pretty obvious. There are a significant number of people who believe the 15" MBP to be too large and too heavy who, nonetheless want the other feathers and capabilities.
Quote:
Originally Posted by hmm
My point regarding RAID 5 was that it's definitely more finicky when it comes to consumer grade solutions. Raid 0, 1, and 10 may still be feasible even without enterprise firmware. Just grab caviar black drives + roc card or eSATA host card and software RAID. Make sure eSATA card and port multiplier chipset are made by the same manufacturer so you don't run into conflicts. My point was that when you started getting into parity stripes or dedicated parity drives (basically anything that splits off parity whether it's byte or block level), hardware and setup requirements become more stringent if you want an absolutely stable rig. If you're dealing with enormous archives of data, you don't really have to keep all of it on such a solution. Anyway this isn't really my area of expertise. I just know what to avoid in most circumstances, and built in Raid 5 functionality is not really something you'd see in a consumer level desktop.
I didn't so much mean this. My issue with the discrete graphics is that they still do not perform well, and their vram allocation misses the minimum mark for many things. It doesn't need a high end solution, just a better solution. Improvements to integrated graphics may eventually turn this aspect into a non issue, but it's a design flaw today. It annoys me somewhat that the mini is still languishing on this update.
Aside from an actual RAID array controlled from the computer, there are a variety of NAS RAID arrays that are useful even if they are only run through a Gigabit Ethernet connection which is doable from a Mini. TB NAS is good as well. I would not call a RAID 5 array an outright "consumer" feature, but there are a good many enthusiasts who want it for speed and safety...your data will survive the loss of one drive, but that is no substitute for a proper backup plan. RAID 5 is popular, comparatively speaking, with the image editing crowd and, no doubt, others. Sure, (a minimum of) three WD Black RE 2 TB drives for a RAID 5 array is not the cheapest solution in the world, but it offers a lot of capability. Having it internally not only has performance advantages over an external setup, but avoids the additional expenses associated with many external systems.
Though your remark about vram for discreet graphics cards was not directed to me, I quite agree with your opinion. Sadly, Apple seems to make a habit of languishing, as you put it, on comparatively straight forward updates of components.
Cheers
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB latency is 9ns so not a problem.
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
http://www.pcworld.idg.com.au/article/431125/who_wants_pc_business_dell/
"From the standpoint of profit margins, the worst performing part of IT has been PCs and it's getting even worse"
It's not a good idea to run in a direction you see everyone else running away from, especially the biggest players.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
Quote:
Originally Posted by Marvin
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
It was always faster. I adopted it in 2005 (I think 2005) or so. At that time it was relatively good for large amounts of cheap storage with tolerable speeds. Even the early cards outpaced firewire in desktop settings. For laptops it caters to more of a similar market to the thunderbolt display, meaning people who primarily use notebooks. I never expected thunderbolt to take off quickly. It's kind of a weird implementation of something that the line needed long ago. I call it weird due to the need to piggyback display data even though this can tie up 2 channels. It's not the most efficient way to do things.
Quote:
Originally Posted by Marvin
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB
Okay this is more Marvin kool-aid. You post some great stuff at times, then you make up numerical claims that exceed intel's best case scenario for their own product roadmaps. Thunderbolt's actual speeds have nothing to do with PCIe 2 or 3. It's using 4 PCIe 2.0 lanes. Unless they update the chips, speed stays the same. Initial optical cables aren't fully optical anyway, and their purpose is just to get around length restrictions. Intel previously claimed it could hit 100Gb/s within a decade, but it comes down to how much bandwidth people are willing to pay for. They never claimed what you are claiming.
Quote:
Originally Posted by Marvin
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
http://www.pcworld.idg.com.au/article/431125/who_wants_pc_business_dell/
"From the standpoint of profit margins, the worst performing part of IT has been PCs and it's getting even worse"
It's not a good idea to run in a direction you see everyone else running away from, especially the biggest players.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
That article mentions they should focus on high margin areas. Look at the pricing on some of their Z line. If they're still moving workstations, at least one part of their PC offering is doing well. Much like Dell they offer a lot of configurations, but with HP if you wish to order anything non standard, the markup is quite high. They're either moving mid priced workstations in stock configurations or custom ones with high markups on the configuration aspect.
http://www.macrumors.com/2012/03/09/intel-working-to-boost-thunderbolt-speeds-with-move-to-pci-express-3-0/
Falcon Ridge in 2014 will see 20Gbps without optics.
That should have been 50Gbps for 2015 although it might be dual 50Gbps channels in both directions but still 50Gbps per channel in one direction:
http://www.tomshardware.com/news/Thunderbolt-silicon-photonics-optical-networking-50-gigabits-per-second-Jef-Demain,12668.html
The main point is, whatever people claim the bottlenecks are now, they will be resolved within the next 3 years.
Which is not the consumer tower market. If you sell a product with a high margin, it needs to have a selling point. A consumer tower is the same as everyone else is selling, how does Apple command a premium on it?
The other issue that also affects the Mini is what display do they bundle with it? Their $1000 27"? So they'd be selling a quad-core i7 box with up to 32GB RAM for $1000 along with a $1000 27" display and a $69 mouse and a $69 keyboard = $2138.
Hmmm, let's see how much that quad-core i7 iMac with the 27" display is that takes up to 32GB RAM and comes with the keyboard and mouse: $2200.
Now, obviously you could buy a 3rd party display with the tower and you get your PCI slots and your multiple drives but most consumers don't care about this stuff any more and eventually nobody will care about this stuff just like nobody will care about optical drives being taken out of everything.
The way Apple makes good money on the iMac is by selling you a display with your computer because they make the markup on the display. Take the display out of the equation and all they are doing is giving that margin to another display manufacturer.
Of course, that would suggest they get rid of the Mini but they can't do this as they can't get the iMac down to that price and it would cut out all the usage scenarios of the Mini. That's why the Mini will stay in the lineup. They will try to make the iMac more compelling but it won't matter, the Mini will keep increasing in power.
By the way, I realize that I jump around on comparisons. I haven't always been happy with the quality of Apple's mobile devices, so I tend to favor their stationary form factors.
Quote:
Originally Posted by Marvin
http://www.macrumors.com/2012/03/09/intel-working-to-boost-thunderbolt-speeds-with-move-to-pci-express-3-0/
Falcon Ridge in 2014 will see 20Gbps without optics.
That should have been 50Gbps for 2015 although it might be dual 50Gbps channels in both directions but still 50Gbps per channel in one direction:
http://www.tomshardware.com/news/Thunderbolt-silicon-photonics-optical-networking-50-gigabits-per-second-Jef-Demain,12668.html
The main point is, whatever people claim the bottlenecks are now, they will be resolved within the next 3 years.
Which is not the consumer tower market. If you sell a product with a high margin, it needs to have a selling point. A consumer tower is the same as everyone else is selling, how does Apple command a premium on it?
You should know enough not to believe intel's promises. It's still 3 years away anyway, and I still think the idea of external gpus is a dumb one. If you look at the 7970m, it's quite powerful, yet the tdp is way down. Pretty soon such a thing will not be worth it as many of those external pci enclosures can't hold the hottest or largest gpus anyway. The bottlenecks aren't really a claim. You can't hook a display + other stuff on a thunderbolt connection unless it's the Apple display without throttling as display data and regular data cannot share a channel. I don't know why you keep clinging to the idea that you can take a mac pro and reproduce it there. Even the mac pro is IO constrained at the single package level. In the past they've simply assigned more theoretical bandwidth than what really exists. It's like if you look at crossfire setups, you can't give each of them 16 lanes, as the PCI lanes simply aren't there in those chipsets. On the Xeon end they've been moved into the cpu package, so if you're in need of any kind of extreme bandwidth, your only option would be a dual package model.
Quote:
Originally Posted by Marvin
Hmmm, let's see how much that quad-core i7 iMac with the 27" display is that takes up to 32GB RAM and comes with the keyboard and mouse: $2200.
Now, obviously you could buy a 3rd party display with the tower and you get your PCI slots and your multiple drives but most consumers don't care about this stuff any more and eventually nobody will care about this stuff just like nobody will care about optical drives being taken out of everything.
The way Apple makes good money on the iMac is by selling you a display with your computer because they make the markup on the display. Take the display out of the equation and all they are doing is giving that margin to another display manufacturer.
Of course, that would suggest they get rid of the Mini but they can't do this as they can't get the iMac down to that price and it would cut out all the usage scenarios of the Mini. That's why the Mini will stay in the lineup. They will try to make the iMac more compelling but it won't matter, the Mini will keep increasing in power.
I'm well aware of their strategy to sell you a display too. I can't use the imac or thunderbolt displays. They're too bright, too reflective, and lack profile stability. I realize this won't be a big deal to everyone. Given everything I require on Linux, I'd switch immediately. There are just some things that don't run or don't run well. I do think they've made a few mistakes on the mini. Internal power supplies make the thermals suck. A modern desktop should be silent under load whenever possible, and they're working against that. I think if they're going to add discrete graphics, they shouldn't go totally low end mobile on them. Apple markets on simplicity. They should be able to meet minimum spec requirements on basically anything in that year without checking (minimum requirements are generally not high, even for powerful applications).
RBR seemed to be going the direction of suggesting a small tower that could also function as a file server, although I'm not sure how that would work out. The problem with the desktop is that it has become stagnant, and the biggest percentage gains in performance have been seen at the low end. The Airs have posted some impressive gains. Desktops do offer some superiority in long term reliability. Just look at the number of complaints on Apple chargers and replacement batteries at the Apple store. They never had great reviews on either product as no one likes to replace those things. It's just that charger cords prone to twisting where you have to discard the whole unit, and batteries known to bulge (I've had it happen) when old are both significant issues. On a side note I hope they fixed that battery thing on the rMBP.
So you agree you don't need PCI slots. If you have a 7970M internally and Thunderbolt handles the IO, you're all set.
That's just the usual vague fearmongering. You can't attach devices that use up all the bandwidth without seeing a performance drop, that's obvious and applicable to any setup. This has been tested out:
http://www.macworld.com/article/1163773/thunderbolt_how_devices_affect_each_other_on_a_daisy_chain.html
You can support two Thunderbolt displays and at least two peripherals on an iMac so 3 x 27" 2560 x 1440 + 2 Thunderbolt devices. In 2014, you will be able to add more without certain devices running at a lower speed.
I find it odd how people want an affordable tower and then suggest it's necessary because the iMac couldn't handle setups that require you to spend over $4k to saturate. Either you want an affordable setup or you don't. If you do, the iMac is sufficient and pretty decent value.
Even if you get a cheap-brand PC like:
http://www.newegg.com/Product/Product.aspx?Item=N82E16883229338
that's $880 with the spec of the next top of the line iMac, which will cost $2200 but a comparable display costs $700:
http://www.amazon.com/HP-ZR2740w-27-inch-Backlit-Monitor/dp/B005MR4P0W/
so you end up at $1580. If Apple sold that machine, they'd make $220 vs $550 on the iMac and yet as far as the consumer is concerned, they are spending comparable amounts (a savvy buyer wouldn't go for the cheapest brands).
Who has Crossfire or SLI in their Mac Pro? This is a setup for gaming. Plus you get a 300W limit so what cards are you going to use? A single high-end GPU uses that.
And there's the explanation. You might as well give up the idea of switching if you aren't prepared to let go some of the legacy baggage.
It's not any louder than the one with the external PSU.
Hardly, PC manufacturers don't make quiet machines. The iMac is almost silent even under load.
Quote:
Originally Posted by Marvin
It wasn't a good solution though. Without power, you can't use a mobile eSATA drive so you have to power it from USB anyway meaning you need to plug in 2 cables. It is double the speed of FW800 in desktop scenarios and it has taken 8 years to supercede it but it was always going to be a temporary option. It shouldn't have taken 4 years to get USB 3 on Macs.
That's what I think it is and I think it's enough to make up for the lack of PCI slots. Having slower PCI is only detrimental if you use the speed on offer from fast slots and very few PCI cards do. Thunderbolt will ramp up in speed in due time. PCIe 3 has the potential to double the speed of Thunderbolt taking it to x4 speed. The Mac Pro only offers x16 (for GPU), x8, x4, x4. There's a roadmap for optical connections to hit 100Gb/s by 2015, that's almost x16 PCIe 3. TB latency is 9ns so not a problem.
The largest desktop manufacturer in the world (HP) shipped 13 million computers last quarter. Of that, 30% are desktops. This means the highest potential market for Apple is 3.9 million units per quarter and that's assuming they ship as many units as HP, which they can never do. Right now, they sell about 1.6 million desktops. HP nearly abandoned the desktop market last year and even Dell (the second largest desktop manufacturer in the world) isn't too enamoured with it:
http://www.pcworld.idg.com.au/article/431125/who_wants_pc_business_dell/
"From the standpoint of profit margins, the worst performing part of IT has been PCs and it's getting even worse"
It's not a good idea to run in a direction you see everyone else running away from, especially the biggest players.
That's not really a new model, just a CPU/display option and a 13" Retina laptop is a possibility. Your mid-range desktop is a new model/form factor.
I am not sure what your problem is with eSATA. I have used it and it is far and away superior to USB 2 or FW 800. I find it inexcusable that Apple failed to adopt it. It may not have been "perfect" in a mobile platform, but it was/is still quite useful. OWC even had a mod for the iMac that simply put an eSATA port in it and connected to an unused SATA port on the logic board. It is not rocket science.
I will say this one last time. Thunderbolt is an improvement over the other currently available external data transfer protocols used by Apple. It is not, however,the equal of, or a replacement for, an internal PCIe. The transfer rates speak for themselves.
It is true that the market share of laptops/tablets/mobile devices has been growing for quite a few years and probably will continue to grow. That is not to say that there is not a place for something other than a laptop. You fail to comprehend the concept of offering a mix of products so that the company has a solution for more than one need. Indeed, you contradict yourself in regard to the 13" MBP. It really should offer a better CPU and screen. Many people prefer its size (and weight) over that of the 15" MBP, but want a better machine.
Apple is coming to a fork in the road where any number of long term users are abandoning the platform because of the lack of support for even the hardware they have been using. If Apple wants to concentrate on iPhones & iPads because they are the most profitable, that is the company's choice, but they should not be surprised if that makes the company an increasingly niche player, even if it is a large niche player.
Quote:
Originally Posted by Marvin
so you end up at $1580. If Apple sold that machine, they'd make $220 vs $550 on the iMac and yet as far as the consumer is concerned, they are spending comparable amounts (a savvy buyer wouldn't go for the cheapest brands).
Who has Crossfire or SLI in their Mac Pro? This is a setup for gaming. Plus you get a 300W limit so what cards are you going to use? A single high-end GPU uses that.
And there's the explanation. You might as well give up the idea of switching if you aren't prepared to let go some of the legacy baggage.
It's not any louder than the one with the external PSU.
Hardly, PC manufacturers don't make quiet machines. The iMac is almost silent even under load.
My main problem with the imac is long term thermal issues. The display also doesn't age well because they cheap out on circuitry. It has all of the signs of a hack job if you view one that's more than a year or two old even if it's been used lightly. The number of aging imacs I've seen with purple splotches is ridiculous. I would go as far as to suggest that the only potentially reliable Macs they're still making are the mini, the pro, and the Air. I've had and seen issues with the rest at one point or another. I've owned a number of Macs, then between everyone I've known who owns them, it adds up to dozens of devices. The point being I encounter them constantly. Anyway my point on crossfire was that they oversubscribe the lanes there. It's not really 32 lanes. It wasn't a why don't macs have this kind of issue.
Quote:
Originally Posted by Marvin
So you agree you don't need PCI slots. If you have a 7970M internally and Thunderbolt handles the IO, you're all set.
That's just the usual vague fearmongering. You can't attach devices that use up all the bandwidth without seeing a performance drop, that's obvious and applicable to any setup. This has been tested out:
http://www.macworld.com/article/1163773/thunderbolt_how_devices_affect_each_other_on_a_daisy_chain.html
You can support two Thunderbolt displays and at least two peripherals on an iMac so 3 x 27" 2560 x 1440 + 2 Thunderbolt devices. In 2014, you will be able to add more without certain devices running at a lower speed.
I find it odd how people want an affordable tower and then suggest it's necessary because the iMac couldn't handle setups that require you to spend over $4k to saturate. Either you want an affordable setup or you don't. If you do, the
Huh? I mainly want a better mac pro. Regarding the xmac thing, I mostly wanted to be part of the conversation there. Thunderbolt creates more problems than it solves on a mac pro, which is why I don't care if it gets it. You need a displayport connection on the edge of the board or you end up with some weird rewiring. The thunderbolt displays are the only chainable displays, and at their pricing I'd rather have two of the 27" NECs as they're just better displays than Apple has ever made. Anyway other displays haven't tested well when it comes to chaining, and it does slow down the other devices. You can't run 2 displays and the TB raid at full performance specs here. That is oversubscribing the bus as it's saturated prior to that point. If you're wondering why others want different desktop options, Apple bundles things to get you into a high minimum sale with the imac, yet they don't last. Computing needs for most people are not growing very fast at the moment (note most people, especially those who buy imacs and macbook airs), so it annoys me that they are not built to handle reasonable wear. Others have complained about getting rid of the display after a couple years. I think it generally looks terrible long before the hardware is truly incapable, and I'm not fabricating this.
Quote:
Originally Posted by Marvin
Hardly, PC manufacturers don't make quiet machines. The iMac is almost silent even under load.
Have you ever looked at a modern PC? Most are silent, especially workstations.
You can't run a Keyboard/Mouse, pen drive or video camera over eSATA. It's a storage protocol and one that doesn't supply power. It is a limited utility port. It doesn't matter if it happened to be faster than USB 2/FW800, USB 3 was always going to be faster and more useful. Now that we have USB 3, it is completely irrelevant. The issue is, we should have had USB 3.0 4 years ago.
If Apple had put eSATA onto machines 8 years ago, they'd be removing them today and people would be left with eSATA devices that didn't work. I think they made the right decision not to support it. It's good to avoid popularising technology you know in advance that you are going to deprecate.
Similar deal with Firewire - it was going to ramp up to 3.2Gbps or so but the whole standard is now regarded as legacy, same as Ethernet. Phil called FW800 and Ethernet legacy technologies when he introduced the retina MBP. That's why they are only supported with adaptors.
The technologies going forward are Thunderbolt and USB 3.
And when the ports are 50Gbps each, what then? Even if people feel TB isn't a good enough replacement now, it soon will be. That's all Apple is bothered about - backing the right horse. TB is an elegant way to connect high performance peripherals to a Mac, PCI slots aren't.
Right but it's not a new model so it's not the same as a mid-range tower vs an iMac. I think they will improve the 13" display and spec (more likely a better GPU than quad CPU) but eventually, it won't even be an additional option. It will become the only option. The Retina MBP isn't just a separate option, it will replace the old-style 15".
Apple used to sell PowerMac towers alongside the iMacs and the iMacs replaced them. They're not going backwards.
The number of users who would abandon Apple are the niche players now. Apple has sold 365 million iOS devices so far. That's not a niche market.
Some iMacs fail and some Mac Pros fail but a great number of iMacs are still running ok, even G5 iMacs from 8 years ago. I don't see any widspread evidence that Apple's small form factor machines have any long term issues due to their design.
The only serious issue I'm aware of is the failure of the home button on iOS devices. Overall though I'd have to say that for portable devices they actually have fairly rugged products.
Put in the Ivy Bridge processors, bump up the RAM, Intel HD 4000, and SSD. One model having the GeForce 640M LE would be nice as I said before. If not then let it be. Have the base model either remain at $599 or drop lower (without an SSD anyway if they get to $499) and if they don't make a discrete model, then drop the server to $799 or $899.
Quote:
Originally Posted by Marvin
You can't run a Keyboard/Mouse, pen drive or video camera over eSATA. It's a storage protocol and one that doesn't supply power. It is a limited utility port. It doesn't matter if it happened to be faster than USB 2/FW800, USB 3 was always going to be faster and more useful. Now that we have USB 3, it is completely irrelevant. The issue is, we should have had USB 3.0 4 years ago.
If Apple had put eSATA onto machines 8 years ago, they'd be removing them today and people would be left with eSATA devices that didn't work. I think they made the right decision not to support it. It's good to avoid popularising technology you know in advance that you are going to deprecate.
Similar deal with Firewire - it was going to ramp up to 3.2Gbps or so but the whole standard is now regarded as legacy, same as Ethernet. Phil called FW800 and Ethernet legacy technologies when he introduced the retina MBP. That's why they are only supported with adaptors.
The technologies going forward are Thunderbolt and USB 3.
You know well enough that Apple spins things however they like in marketing. You may also notice that my posts sound more grumpy after recent hardware failures (in this case another swollen battery). A year ago people were doubtful over whether Apple would even release usb3 drivers and suggested that thunderbolt was the only thing needed. FW800 should have been legacy long ago. eSATA should have never made it off the ground. The reason both survived was due to lack of better solutions. A second thunderbolt port on the new ones is a huge improvement in itself. I haven't bought any firewire device in years that didn't also support usb just in case the firewire option was dropped and I needed to use it a bit longer.
Quote:
Originally Posted by wizard69
I do agree that eSATA was a waste and for the most part a poor investment. Apple has dragged its feet excessively with USB3 and has caught considerable flack for it. About the only thing here to debate is the 4 year term as it took a long time for suppliers to get the bugs out of the USB3 chipsets.
This I agree with 100%' especially with eSATA which frankly is a crappy standard.
Years ago it was somewhat of a makeshift solution. It was stable and could sustain 200MB/s~ easily enough depending on chipset. Keep in mind during the G5 era, you got two internal bays and hard drive capacities were far more limited. I'm actually in the process of phasing out older eSATA stuff. It worked back then. Now it's simply dated, but usb3 should have come out last year. They went super conservative as usual. Unless something carries a coolness factor, Apple is very conservative on what they add. It may be an unwillingness to take on construction or support costs.
By the way I could see a redesigned iMac I could go for reluctantly. Right now there is no way. I would still be a little pissed at having to buy an all in one but it may be the only way to midrange performance on the desktop. Apple would have to address serviceability and a few other negative factors with the current iMac though.
In any event this is a Mini thread so let's get the train on track. Who here is expecting a new Mini in the next 2 days? If they don't come with Mountain Lion then I have to really wonder what the hell is up with Apple. I can't see any technology they could be waiting on for this machine otherwise.
Apple might have had an arguement in that they would have to support those third party chips for a very long time. The problem with that is that they screw with the good will of their users. The need for USB 3 support is legitimate and to ignore that support for almost two years after viable solutions exist just infuriates customers.
Yep they are cheap. That and they seemingly have ignored the Mac hardware lineup for a long time. I mean really the Mac Pro was turned into an industry joke with the last so called update. Sadly no USB 3 there either. If the Mini doesn't update soon it will become another joke. Seriously selling a Mini these days without USB 3 support is so anti consummer that even Apple will start to suffer.
If not within the next week than possibly sometime in August. Either way, I cannot see them not releasing a mini at all in 2012.
The iMac has actually gone longer without an update, which is unusual. It makes sense for them to spread the releases out though as the next event will likely be the iPhone launch on the anniversary of Steve's death (October 5th), given that the last one was Oct 4th.
They wouldn't update the Mini before the iMac either and we know Mountain Lion is coming in July. They may as well do all 3 this week.
It'll still fall short of some expectations, I think it will be:
Core i5, HD4000
Core i5, GT 640M
Core i7 quad, HD 4000
4x USB 3.0, HDMI, single TB.
I wonder if they will drop FW800 from the desktop machines. I don't think it makes sense to drop ethernet as they have plenty of room but dropping FW800 will save some money and push adoption of USB 3. A 5th USB 3 port would be nice if they did this though.
The iMac is their best selling desktop so while it hasn't been left without a major update as long as their best selling machine - the MBP (473 days), I expect an update soon.
Quote:
Originally Posted by wizard69
Ethernet is not a legacy port, it is the backbone of most business networks and more that a few home networks.
We sure install a lot of Cat 5 and Cat 6 cable and data jacks in new businesses, stores, schools, hospitals, etc for ethernet to be going away like some people think. Sure we install more wireless nodes now but we haven't seen any business willing to trust wireless completely.