The Mac Pro is Dead - Page 6

So that makes one. What else?

Apple's own pro apps via Compressor then Qmaster - After Effects - C4D - there's heaps more ... no point having all that heavy hitting capacity and not using it to full advantage.

Most of us set up the renders/transcoding to run at the end of the day - come back next am, all done. Sweet !

I say that with the qualification that some software isn't, never will be and doesn't benefit from threading. However the vast majority of apps available on Apple platforms use some amount of threading, either by design or via included libraries.

The advent of Snow Leopard (SL) and Grand Central Dispatch (GCD) resulted in improved threading for many apps without a major coding effort for the developers. Of course you need to understand that GCD is just one way to exploit extra execution engines in your hardware. Many of Apples included apps got a significant boost in performance simply the result of SL. Other apps useing various approaches to threading exist. Video processing, Ray Tracing, engineering software and other apps can be heavily threaded. Development tools can often start up multiple instances of a compiler to drastically reduce compile times.

These are heavier uses of threading but most Mac software uses some threading to keep the user interface responsive while things happen in background. For example printing a document can be handled in a thread while the main app continues to function. Like wise with spell cheating and other things that happen concurrently as you create a document.


I'm not sure what you mean by full advantage? How an app benefits from threading depends upon a couple of things. For one; each solution to a problem results in a program that may or may not heavily leverage threading. Two; programmer skill is significant here, just because programmer X implements a threaded solution, it doesn't mean programmer Y can't do a better job. Three; even highly threadable problems run into hardware limitations where adding more cores slows down the overall gain significantly. This due to communications and data transfer overhead. Four; one programmer might add one thread to an app and consider it fully threaded, another might add ten to the same app.

In the end what I'm saying is that there is no way to truly define an answer to your question.

I believe your choice of words is less than optimal here, as the word optimized implies optimal usage of a capability. I believe this is misguided for one thing you might not be able to ever find the optimal threading implementation for an app. Some problems are almost brain dead easy to thread but even then what does it mean to have an optimal solution.

The list of software that uses no threading at all is likely tiny compared to apps that leverage two or more threads. Think about your common tools that most user operate these days. Mail programs, word processors, web browsers and many others on everybodies desktop have some threading built in of one sort or another.

Something like some versions of Safari use multiple techniques such as running Flash in a separate process. This to reduce crashing but it also means that flash could be running on a different CPU. WebKit is currently being reworked to support more threading capabilities. Will this result in future versions of Safari using 24 cores efficiently? Probably not but those cores aren't exclusive to Safari anyways.

A few days ago, in another unrelated thread, I mentioned that it was time for the Mini to go quad core. I believe this is the case because it is the best way to support common use cases. For example it is not uncommon for people to have FaceTime, Skype or other video calling program running while doing other things. Some of these programs can take a significant amount of your CPU performance thus extra cores brings you back to parity. Also enhancements to apps means that they can better leverage those cores. Really if you look back to when dual cores where announced people dismiss them as unneeded, yet we saw very rapid adoption once the benefit was seen by users. SL gave Apple the long term infrastructure to make 2+ cores viable. So yeah a minimal of quad cores in a Mini makes sense these days. We simply have another level of expectations.

http://www.sonystyle.com/webapp/wcs/...52921644777498



Some specs:

• 24" touchscreen LED
• BluRay burner
• 2 TB HD
• i7 quad-core 3.3 gHZ
• WEB button: browse internet without a full Windows boot-up
• HD web camera and microphone with Exmor image sensor & face tracking
• Wireless keyboard and mouse
• Remote control
• WiFi 802.11b/g/n Bluetooth 2.1 + EDR Technology
• 4-pin i.LINK® port
• Composite Audio/Video Input
• Headphone Output
• HDMI Input
• HDMI Output
• USB 2.0 ports (3)
• USB 3.0 ports (2)
• NVIDIA® GeForce® GT 540M with 1GB VRAM
• IN-HOME warranty SERVICE
• Color: Black or White
• Wall mountable

There was a time when Apple would sue, win, and subsequently bankrupt a company for doing any design even remotely close to theirs.

Where the heck have those days gone?!

Maybe they'd be embarrassed to make a fuss over a machine whose specs put their "big iron" desktop MacPro to shame? Nah, nothing embarrasses them, probably better to think they seriously do NOT want to draw any extra attention to a computer like this Sony.

And for the record, I've been using Macs since May 1984... that year I invested in the machine, the fat mac 512k upgrade and a 10MB internal HDD to the tune of over 4 grand (would have been over 5 but I got an edu discount on the machine itself).

Laugh all you want, but the HDMI IO is useful. You can use the display as a TV when you have a digital TV box. Why not use that screen? Reversely, you can use the BluRay/DVD drive to play back movies on a big-screen TV.

It does not look as good as an iMac and (more importantly) it does not have OSX, but if the price is right I can see this doing well.


On a different note, why is this in the MacPro thread?

edited for brevity

I am an ignoramus (albeit an enthusiastic one!) so I'll take your word for it.

However, there were some DAW benchmark tests last year that demonstrated that most DAW's do not take full advantage of hyper-threading and multi-core processing under OSX. It was better under SnowLeopard (compared to Leopard) but nowhere near as efficient as W7.

For perspective, the guy doing the testing builds custom DAW PC's for a living, so he is not entirely devoid of bias, but I believe his basic method is sound and don't think he could fake it completely without being found out.

http://www.dawbench.com/

hum...

Up to Intel® Core i7-2720QM quad-core processor (2.20GHz, up to 3.30GHz with Turbo Boost)

Not the same thing at all.

HDMI is digital lossless audio/video: yes it is very usefull: you can connect second LCD monitor, or run HDMI cable to your stereo receiver to view BluRay or streaming movies on big TV, listen streaming Pandora or your MP3 collection though some quality floor standing speakers. I wish iMac would come out with HDMI.


I agree, you can't beat iMac industrial design.



Sorry, my bad: I've had iMac thread opened in the other tab but posted in this thread instead.

Agree

Yes, i7-2720QM is Sandy Bridge.

HDMI In/out: Two separate HDMI ports.
HDMI In: connect external audio/video source to your Computer (cable box, camcorder, etc.) for capture/playback.
HDMI Out: use your computer as audio/video source and hook it up to a second monitor or large screen TV and/or high-end stereo system.

Thunderbolt. Seven monitors.


Apple TV. Better, as you're not restricted by wire.


AirPlay. Better, as you're not restricted by wire.


And for those reasons, they absolutely never will.


Thunderbolt is superior for the aforementioned reasons.

Just seven? Theoretically may be seven. There is no 1x7 Thunderbolt splitter available. You can get 1x16 HDMI splitter now.



"Better" if you don't mind paying $100 for AppleTV and agree to be locked into its limitations. And how do you get your video from AppleTV to TV? Via HDMI! AppleTV is just a $100 middle man with lots of limitations and no web browser.



AirPlay cost $100. Plus I would have to throw away my receiver to buy AirPlay compatible Denon receiver. The cheapest is $700 street, but the good audiophile quality AirPlay compatible Denon will cost twice more. HDMI is on the other hand industry standard works with most even 5 year old receivers/TVs.

Also AirPlay works over your Wi-Fi network and, as anybody who uses Wi-Fi knows, it's not always rock-solid. In testing some AirPlay speakers, we experienced the occasional dropout and even a few complete disconnections, but if you want go from computer to receiver wirelessly, there are much cheaper solutions compatible with any receivers/speakers (so you more equipment choices).
Direct HDMI wire will always be cheapest, most reliable and setup-free solution.

Yes, you can browse web on your TV using VAIO-HDMI connection and included wireless keyboard. You can also do it using Sony stand alone BluRay player (via Wi-Fi or Ethernet), or Logitech Google TV device.

You did not know it???

While I have no use for this machine, I don't think it dead yet. It will continue for a few more years, but very soon hardware will no longer matter. We are quickly approaching a world were the hardware is way beyond the needs of the software. Then the MacPro will die.

I'm making fun of the idea of doing it at all. I've known it was possible since before the Internet existed. There's just absolutely no reason to do it.

OK, I've been doing some reading and this article also suggests that most apps do not multi-thread efficiently:

http://macperformanceguide.com/Revie...ere-Cores.html


I'm gonna do some more reading, but so far I have not seen any evidence that 4+ cores yield significant real-world performance gains.

That page doesn't talk much about specific software. This has some:
http://macperformanceguide.com/Optimizing-Grades.html

One problem with the list is there aren't any video programs in that, most of those are still image apps of different kinds. Most still image software uses probably do fine with a quad, it's the video and maybe 3D users that might want more than that.

There are lots of things I don't know. That being the case I don't run around calling myself an ignoramus. Have a little self respect.

There is a lot more to the world than DAW. Some apps exhibited an immediate gain under SL that was very significant. Others needed revision.

The web site you linked to is complete BS. It is obvious if you listen to the tone of the writing that the guy has an agenda. For one he only tests with one software product and then puts a Windows system up against an outdated version of OS/X for half the test. Where he runs a Mac with 10.6 the numbers are very similar to the Windows machine.

In any event you can't build an opinion around one web site.

Nothing about the above web site impresses me at all. It is all marketing crap to draw customers to their products.

Assuming you are in the audio workstation market would you base your hardware selections upon one web site and opinion? It is kinda disturbing to me that you would even bring this web site up as evidence of anything. Besides that how often is absolute performance of a value greater than the other advantages of Mac OS/X. I know some people have a tendency to tie themselves and their careers to one software product, but do you really think that is wise?

If you aren't in the audio workstation market then why bring up the subject anyways.

carry on

That's about as polarized a first post as I've ever seen.

You're also completely wrong. The previous sentence is not an opinion.

Oh, right. Obligatory

What's your problem, man? Take a chill pill. We're just talking.

It is one thing to have a low opinion of somebody else, it is a vastly different thing when you express a low opinion of yourself. That is what I was concerned with in the first comment.

I don't have a low opinion of myself, but thanks for your concern. Fact is, I am an ignoramus, or an informed layman, if you will. All I know is what I've read on the net and in magazines. No personal expertise. As such I think the qualification is correct. But not knowing does not imply I am stupid.

I've done a lot of googling on the subject, and although there are no real empirical tests, what I have found so far (some of which I have linked to in this thread) indicates that most software does not really take advantage of multi-core (or hyperthreading, ftm) processing. Not just DAW's, before that post I linked to some reviews of photo software by MacPerformanceGuide.

I am well aware of DAWbench's agenda, and I pointed that out when I posted the link. BTW, his last tests are 6 months old and he DID use the latest version of SL for that. There were clear improvements with regards to multi-processing. His methods are not perfect, but I am not aware of any other similar tests so it'll have to do.

What I find more telling in this regard, is the (normally very vocal) Mac community's silence. They are always quick to point out OSX' superiority, but I have not found any tests that rebut DAWbench's claims. Except one, he did the very same test with Logic and it reigned supreme. All cross-platform software performed better under Windows. I'd be very happy to find evidence to the contrary, so if you have any please share. When I get a new MacPro I will definitely run the test again on both OSX and Win7 to see for myself.

I am content to be a Mac user, raw performance does not tell the whole story. But I also want good information and have no problem if the little anecdotal evidence that is out there tells me that OSX may not be superior in all aspects of computing.

I guess I must still have my head in the "Clouds" but maybe there is life after death... first the Lion is shaping up for release then FCP unleashed, and now New MacPro under development, http://www.9to5mac.com/63107/prototy...ble-stackable/

hope they take the opportunity to ditch some of the old legacy junk and make a generational leap statement for pro home server to rack and studio processor, could be interesting...

Someone over at MacRumors did a mock-up based on 9to5's speculation:


I'm all for smaller, as long as it keeps cool and quiet. And a lower base price would be good. Yeah right.

Done right the revised pro could serve many more users than the current one at a lower price. The problem with the current model is the focus on one class of customer.

5" wide and 19" tall seems like it would be fairly easy to topple over and the internals would be positioned strangely. Putting the optical drives and hard drives in sideways will use up a lot of space plus a vertical tray-loader is not a good design.

A single or dual slot-loading optical makes most sense, maybe none at all, like the XServe.

This design may also have implications for the PCI slots. The XServe can only get one x16 full-length and one x16 half-length in. This may mean the Mac Pro would have a single free expansion slot as the GPU would go in the main one.

I also wonder what they will do with the handles, which are not the most efficient use of space when it comes to server use but handy for moving a very heavy machine around.

Another interesting thing is that Intel doesn't have any chips to go in a new one yet but with a new Final Cut and Thunderbolt, there seems like a very compelling reason to offer an upgrade. A redesign seems like the best way to go with some new options (new Radeon GPUs), Thunderbolt and maybe some new pricing structure.

Maybe the idea is horizontal placement?

Many desktops mare already in this size range. As to the optical if Apple was smart they would implement a rotatable module, if they implement optical at all. Personally I still see the optical as a waste of space for the vast majority oaf users.

Even here Apple has options. They could go to notebook sized hard drives for example. This would save a lot of space. Another option would be to implement solid state storage. In a server the best way to implement that is via a PCI-E card. The point I'm trying to make is that Aplle needs to move forward and not dwell on past configurations or acceptable practice. The best possible deal would be a Fusion IO boot drive along with more conventional bays for bulk storage.

Centuries ago man invented the screw! Not to be an a$$ but handles can be dealt with in a number of ways.

Sure they do. At least I remember some SB based Xeons being released a week or two ago. Granted these aren't top end Xeons but they would be a substantial upgrade anyways.

One key thing to realize is that technology will allow for a significant repackaging of the Pro. There are a lot more options for storage these days than ever before. That is just one part of the platform, chipsets take less power these days and even things like power supplies are shrinking.

Making it impossible for existing users to upgrade and artificially limiting drive size. Don't think so.

I personally think they should as they can offer more bays (e.g 8 bays) to allow enough storage for people migrating and to reduce heat internally as well as noise but they may not want to cut off the 3.5" options, especially for servers.


They can certainly bolt the handles on but that would be very unlike Apple. One design I had considered was something like an XServe but with a base plate to allow it to sit vertically:



This wouldn't be quite so bad to carry as you'd carry it like you would an iMac. You would take off the base plate to mount in a 1U server space. The ports wouldn't be laid out that way but it's an idea of what it could look like.


The released Westmere-EX are too expensive. The Westmere-EP coming in Q4 would have been most likely but the following chip would be ok in the entry model:

http://ark.intel.com/Product.aspx?id=52277

Nothing changed for the higher up models though.

Interesting idea, Marv'...

Maybe after 8 years...Apple is set to reinvent the tower...(maybe they'll reinvent their prices while their at it...)

Lemon Bon Bon.

Tell me you don't pull out 5 year old drives to put into a brand new computer when you upgrade. Besides there is nothing artificial about it. Many of todays 3.5 inch drives have laptop sized component inside as it is. As long as the new machine supports enough devices to present a nicely RAIDed volume to the user who cares what the physical size of the drives is?

Just look at Apple's Blade SSD's, these give just as good performance as devices in traditional physical form factors. Yet I still here people expressing misgivings about those devices. Why? I really don't know, but people need to realize that new technology enables the functionality thy crave. You can't dwell in the past and hope to navigate the future.

Eight bays would be over kill if you are trying to market a machine to a wide array of users. It would be great for a server only machine but it would use a lot of space that could potentially be used for other hardware.

As for the 3.5" option, many servers these days have left that size behind. One can simply stuff more stuff in the box with the smaller form factor. Not to mention you lower thermal loads. Plus today many of those so called 3.5" devices come with laptop sized guts anyways.

In the old days a lot of rack mounted equipment came with handles, so Apple could easily provide a front panel with handles and not look out of place one bit. However that really isn't my point, the point is if Apple designs a new Pro to these dimensions they have all the flexibility in the world to come up with a viable solution that addresses both the desktop users needs and the rack mount users needs.

8 if they used 2.5" drives to match the capacity of 3.5" drives but I guess even then it's a bit overkill. They do need more than 4 though because people setup drives in RAID 01 and they don't have a boot disk. They can possibly have 2 SSD blade slots and 4 standard hard drive slots.

I wish they'd put 64-128GB SSDs in all their machines. I don't see the point of their ultraportable exclusively having the fastest boot drive out of their entire lineup.


I think one important thing for server use is hard drives they can exchange from the front. With a Mac Pro on a rack, if a drive in a RAID system fails, you have to pull the whole machine out, flip down the side, take out the tray, replace the drive and then put it all back.

It doesn't mean they have to be individually visible at the front like the XServe but there can be a long tray that pulls out while they are still connected and running. Like an optical drive slot but a long tray with 4 drives on it.

I understand the needs of some for lots of RAID storage, that really isn't the problem here. The problem is Apple needs to design a new Mac Pro that addresses some of the old models problems. The big problem being that it doesn't sell enough to justify it's existence. I'd hat to see Apples only slot bearing desktop go the way of XServe.

So what I'm saying is that if you add to many bays and increase costs dramatically youmwill end up with another Pro very few will buy. I realize the need is there for massive storage but the Pro needs to be profitable for Apple. Right now all the signs are that it isn't profitable at all.

I know some will go of the wall over this statement but the writing has been on the wall for some time now. The Pros limited appeal and relatively high costs mean that there is a shrinking market for the current design. It would be a big mistake for Apple to come out with another Pro that does not appeal to a wider audience.

I tend to agree, I was extremely disappointed that the new MBP did not have any Blade SSD slots. This regression kinda makes youmwondernwhat the term "Pro" means to Apple. The
only thing in disagree with is the SSD size, 128 GB is way too small these days.

Some one at Apple must have been smoking some of those funny California cigs when they offered up the Mac Pro as a replacement for the XServe. I still think it is possible for Apple to address both the RackMount requirements and the desktop needs in one machine. Maybe what Apple needs here is a design where then drive module slides out the front

I was thinking a short tray myself withnthendrives stacked. The point is Apple has lots of options here. A stack of laptop sized drives might end up being 56 mm high that should give Apple plenty of room. They could add air filtration into the same assembly.

In any event the thought that started this thread seems to be real - the current Pro is dead! Apple needs a machine that attracks a wider array of users.

is good news?