Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Wirelessly posted (iPhone: Mozilla/5.0 (iPhone; U; CPU iPhone OS 2_2_1 like Mac OS X; en-us) AppleWebKit/525.18.1 (KHTML, like Gecko) Version/3.1.1 Mobile/5H11 Safari/525.20)

These benchmarks look great. It makes me want one really bad.
 
Im thinking of getting the 2.26. Does anyone know if it would be possible to upgrade the processors to say the 2.66 or 2.93 in a couple of years?
 
Not every application benefits from multiple cores, and there are also natural limits to parallelization for applications that can be parallelized. Even miracle technologies like "Grand Central" won't change that. And Snow Leopard alone won't make existing apps faster - they must be written for Grand Central, and they must be able to take advantage from parallelization (which already rules out most desktop or office applications).

Games are traditionally single-threaded (Half-Life 2 being just one famous example), because they require heavy synchronization and threading only adds to complexity, but not to performance. Games want ONE fast CPU core and a fast graphics card to be happy.

In Final Cut Pro, the encoding of videos will be faster on a multi-core machine, but you also need lots of RAM to speed up the process (after all, each core requires RAM to work). How well Final Cut Pro scales beyond four cores is another question.

Handbrake is another app that takes advantage of multiple cores, but like Final Cut Pro this is a video encoder, so it's not very surprising.

Aperture also uses multiple cores for RAW processing and it is an application that can easily scale to all available CPU cores - it just needs to spawn a worker thread for each CPU and there's not that much to synchronize.

Most other applications are not very likely to benefit from multiple CPU cores, so you will almost always be better off with less, but faster CPU cores. And, again, Grand Central won't change that either. It will just make it easier to write applications that by definition can benefit from parallelization - but it won't be of any help for apps where there is nothing to parallelize.

You can also expect that most of Snow Leopard's promised performance boosts are not coming from heavy usage of Grand Central, but from very old fashioned code optimization, code rewrites and the dumping of legacy code.

But once again, Apple's marketing "just works", and you've all already bought into the secret weapon "Grand Central", believing that it will make a super computer even out of your toaster. It's more likely that all available additional computing power will be burned on some eye candy instead of finishing the task at hand faster. After all, the company wants to sell you faster machines at higher prices.
 
Have you priced it out with dual 2.93 GHz Xeons? EXPENSIVE YES!

I am not impressed that the Nehalem low end octacore model is not able to even beat an octacore 3.2 from 2008. That is not good news in the short run. In the short run, a quad core with higher clock speed is going to be a better investment for MOST current apps - until SL and the new "Pro" apps written to take advantage of this new technology.

Yeah, if I were Steve Job's brother I'd be urging him to drop the 4 core low end Mac Pros all together and replace them with a Corei7 systems which essentially have identical specifications, and price them competitively with currently available boxes using that chip - which I'm told start somewhere around $1,000 and even below.

I haven't looked but if there's a Corei7 machine with identical properties or identical in all but something as arcane as ECC facilities, that sells for around $1,000 then IMO that makes the $2,500 four core Mac Pro look like a pretty huge marketing mistake.

This is only the case for the 4-core single processor Mac Pros which are based on the 3500 series xeon processors. The 8-core dual processor Mac Pros don't have an identical less than half priced counterpart that I know of. :)
 
Games used to want one fast CPU core and one fast GPU. But nowadays new games start taking advantage from multiple cores and multiple GPU's. World of Warcraft uses 2 cores and doubles frame rates through the second core. GTA4 uses 4 cores and most games use 2-4 GPU's like COD 5.
 
Im thinking of getting the 2.26. Does anyone know if it would be possible to upgrade the processors to say the 2.66 or 2.93 in a couple of years?

As far as i knew yes, you can. It actually looks like Apple made it easier to do so than in previous Mac Pro models. So, thank you Apple!!!
 
Games used to want one fast CPU core and one fast GPU. But nowadays new games start taking advantage from multiple cores and multiple GPU's. World of Warcraft uses 2 cores and doubles frame rates through the second core. GTA4 uses 4 cores and most games use 2-4 GPU's like COD 5.

Yeah, Quake4 also has a Multi-Processor switch. I dunno if I can tell any difference or not... but the switch is there. ;)
 
Being an owner of an 8-core 2008 Mac Pro, I can't tell you how happy I am to see that the new machines do not blow the old ones out of the water.
 
Although the fact that it only accepts 8 Gig of RAM is a real disadvantage to any pro / freelance Music / Audio guys as RAM is very important for sofware instruments and sample playback etc. Also it has no scalability / upgrade path. Seems to me like the OCTO 2.26 should have been the base machine and priced where the QUAD currently sits.

That's why I said the NEW Quad-Core is essentially an overpriced xMac and NOT REALLY A MAC PRO...

That's my belief. It's NOT affordable, and it sticks with the "Mac Pro" branding for marketing purposes, but I believe our desired xMac was introduced last week at a hefty price tag.

It is hard for all of you real Pros who have to go backwards or spend so much more for the same performance. I really disagree with Apple's forward thinking here. I think Apple should have started the Mac Pro at Dual 2.66 GHz Octacore Nehalem to compete with its last generation.

HOWEVER, I think March 24 may show us all that Apple has some actual software to back up the numbers behind the new Octacore 2.26... at least I hope Apple can back it up somehow... someway...
 
So any quad benchmarks yet?

Are they faster than the old 8 cores (as Apple seems to want us to believe)?

Are they faster than the PC i7 boxes? Some people keep insisting that they perform better, I'd like to know if this is really the case or if those people are just getting sucked into the RDF.
 
Ah awesome, this has really helped me to make a decision.

Yup, also while the $2500 and $3000 Mac Pro single processor quads are an incredibly bad deal right now there's speculation that the daughter cards (riser boards) that seat the CPU and memory may indeed be interchangeable with dual processor octad parts. If true and there's no reason to doubt it at this point, users can also upgrade from a single processor to a dual processor system - assuming they can find the parts for purchase. Some stealthy mac hardware hackers claim that such parts should be available no problem and point to new (as in not used) 2008 mac pro mother boards being available if one knows where to look.

Personally I don't think it's a financially viable option. More than likely the faster processors will not come down in price to a reasonable upgrade level for about 3 or 4 years. At least so it is true of the 2006/2007 compatible mac pro processors. At that time upgrading while certainly will improve performance and very significantly, might not look like such a good decision in the face of the then new 2011 or 2012 mac pro line-up - assuming there still is one and the planet is still here and all. :D
 
So any quad benchmarks yet?

Are they faster than the old 8 cores (as Apple seems to want us to believe)?

The new octo 2.26 performs around the same as old octo 2.8. It's impossible that the quad 2.66 can even come close to old 2.8 on multi threaded tasks. Their single threaded performance will be similar though.
 
So far I'm not impressed by these numbers. I have octo 2.8 which I use OC'd to 3185 Mhz. My geekbench result is 12560 and my Cinebench results are

Single 3678
Multi 21598

Which are much better than octo 2.26 and worse than octo 2.93. Considering octo 2.26 costs 500$ more than my octo 2.8 this is not impressive at all. Unless these new machines can be OC'd significantly somehow in the future, or SL brings much more speed to these things, I don't see why anyone would purchase octo 2.26 or octo 2.93 on that matter, since an OC'd 3.2 will outperform 2.93 for sure.
 
You can pick-up a second hand or apple refurb 3.2GHz for less than the 2.26GHz model and get equal or better performance :)

Where? Apple doesn't have any listed on their website, and the pricing on the previous models doesn't seem to be much of a deal. Any links to good deals would be appreciated.

I dont see why it would reduce performance?:apple:

Because instead of the ram running in three channels, it only runs in two. More ram, but it's slower.

That's why I said the NEW Quad-Core is essentially an overpriced xMac and NOT REALLY A MAC PRO...

That's my belief. It's NOT affordable, and it sticks with the "Mac Pro" branding for marketing purposes, but I believe our desired xMac was introduced last week at a hefty price tag.

It is hard for all of you real Pros who have to go backwards or spend so much more for the same performance. I really disagree with Apple's forward thinking here. I think Apple should have started the Mac Pro at Dual 2.66 GHz Octacore Nehalem to compete with its last generation.

HOWEVER, I think March 24 may show us all that Apple has some actual software to back up the numbers behind the new Octacore 2.26... at least I hope Apple can back it up somehow... someway...

QFT. And while I look forward to software updates (I pray that LOGIC is included, but I'm not hopeful for that bastard stepchild), I doubt there will be huge performance improvements.
 
So far I'm not impressed by these numbers. I have octo 2.8 which I use OC'd to 3185 Mhz. My geekbench result is 12560 and my Cinebench results are

Single 3678
Multi 21598

Which are much better than octo 2.26 and worse than octo 2.93. Considering octo 2.26 costs 500$ more than my octo 2.8 this is not impressive at all. Unless these new machines can be OC'd significantly somehow in the future, or SL brings much more speed to these things, I don't see why anyone would purchase octo 2.26 or octo 2.93 on that matter, since an OC'd 3.2 will outperform 2.93 for sure.

What are you using to overclock? If its that lame ZDNet utility how did you get around the fact that it speeds up the system timing - clock and all, making it actually impossible to get higher scores out of any of the current benchmarking utilities Xbench, CineBench, and Geekbench?

Do you have a link to your Geekbench results page?
 
What are you using to overclock? If its that lame ZDNet utility how did you get around the fact that it speeds up the system timing - clock and all, making it actually impossible to get higher scores out of any of the current benchmarking utilities Xbench, CineBench, and Geekbench?

Do you have a link to your Geekbench results page?

I use ZDNet utility. System clock is reset to regular speed at reboot always. So after a reboot you start getting the actual scores out of any benchmarking program.

Edit: I'm trying to locate my geekbench results in their website but I'm having problems loading the site. Maybe their servers are busy.
 
The new octo 2.26 performs around the same as old octo 2.8. It's impossible that the quad 2.66 can even come close to old 2.8 on multi threaded tasks. Their single threaded performance will be similar though.

Makes sense. I'd still like to see the quad results particularly compared with an i7 machine.
 
Handbrake is another app that takes advantage of multiple cores, but like Final Cut Pro this is a video encoder, so it's not very surprising.

Of all the apps I use that are multithreaded, I use this one the most. I'd love to see some encode times for 2 hour movie from VideoTS to mp4 using and H.264 codec.

Right now I do a 2-pass encode with a turbo first pass. Pass one is encoding in the 250fps range and the second pass is around 150fps. I want to say that for a 2 hour movie my encodes are in the 25-40 minute range depending on the settings.
 
It is nice that the machines are doing well but this confirms that the gulf between iMacs and Mac Pros is even wider. Thus many of us are wondering when Apple will fill that hole.

They did, they're charging $2499. Mostly to create a nice neat pricing structure. Its now set up so that dual core iMacs don't over lap with quad core Mac Pros which don't overlap with octo core Mac Pros. They're created their own little alternative reality of the computer market. Each step gets double the cores and double the memory sockets. Its great if you're a top dollar executive with OCD, not so great if you're a Mac user making less than six digits trying to make a living.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.