Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
There is a lot more wisdom on these boards than any "specialist" I imagine you are going to get on the phone at Apple.
 
I went for the 6-core. Remember that Nehalems and Westmeres offer hyperthreading, so the 6-core will present 12 virtual cores to your software, and a 3.33GHz processor will have no problem keeping up with the virtual load. On my Dell (see sig) my video editing software uses all 8 virtual cores and is very fast. So I'm counting on my new 6-core to plough through the 12 parallel tasks with ease.

Question : Can somebody explain this hyperthreading thing? This sounds like a really big deal to me, the notion of 6 cores presenting as 12 cores to your software. Yet it has hardly any mention in the ongoing discussion.

Thanks!, Julian
 
Hyperthreading

Example:

Programs written for it will work as if they have actual physical hardware cores.

See my Corei7 iMac: it is a true quadcore but it has hyperthreading with 2 threads so programs written for it act like a octocore:

logic.jpg


Logic Audio shows as if I have 8 cores.
Istat menus shows 8 cores.
Boinc/Seti @ Home crunches 8 workunits at once.

etc.
 
So I'm currently using a Quad G5 and its pretty much dead - so I need a new machine (it keeps overheating and crashing after 30 minutes)...

Have you tried cleaning it lately? I have to vacuum out the heatsink of my Quad G5 every few months or it starts to grumble and moan like an arthritic old dog.
 
Avid MC?

I know Avid's been terrible at updating their software for Mac, so I'm curious when or if they'd ever take advantage of multiple processors.

I'm assuming right now the 6 core would be faster at running MC, but does anyone have thoughts for future Avid updates?
 
6, 8, or 100s?

I've wondered about what happens when highly parallel tasks (video) are coded to use cores in the graphics cards with OpenCL/CUDA. For parallel tasks, when your software supports it, it would make more sense to get a slower CPU, but a really nice GPU like an Nvidia Tesla or AMD FireStream. Then it matters much less if the main CPU is 6 or 8 core, if you have a GPU with hundreds of cores.
 
6 cores

Went with the 6-core with 5870 video card. RAM and HD upgrades will be 3rd party, probably OWC. Got two of these for our video production business.

The way I see it, these will do the job faster now and will serve us very well for running Final Cut Studio and Adobe CS5 for the foreseeable future. It's a bummer having only 4 RAM slots, but we'll probably do 12GB now (3 x 4GB), and if we need more RAM later, we can add another 4GB stick ... or swap in 8GB units eventually – hoping the 8GBs work (has that been confirmed?), and likely that the prices for 8GBs will fall by the time we need them.

While the 8-core has a tempting potential upgrade path to 12-core, the need for that many cores seems pretty far down the road for us. If we knew that we'd need that many cores sooner, for us would make sense to just take the plunge now. But 6 fast cores now will pay off immediately. And by the time that our new 6-core machines start feeling slow to us (how many years down the road?), odds are that brand new machines may prove to be a better bet because of new technologies than trying to upgrade our (by then somewhat outdated?) 2010s.

Bottom line, we're going for a sure thing now. Ordered today. Delivery estimated Aug 25 to 31.
 
I've wondered about what happens when highly parallel tasks (video) are coded to use cores in the graphics cards with OpenCL/CUDA.

For larger problems ..... nothing happens. One of the things folks are "griping" about here is that have "limited" amount of RAM memory on a Mac Pro. .... "There is only 12GB of RAM.. .can't possibly get any work done.". Well on these GPUGPU cards 2GB is typically the current limit and quite often it is less than that. As you as you have to round trip the problem chunks to/from RAM to the card and back you'll start to loose the speed advantages relatively quickly.

Likewise if the computation hits the I/O ( disk , NAS ,etc. ) with regularity for data... the GPGPUs won't deal with it. They are oriented toward more purely RAM resident computations.


For embarrassing parallel programs and data that are all ready partitioned and loadable into RAM that are distributed over a large number of nodes... yeah it will get traction. ( one of the top 10 supercomputers is leveraging some GPGPU cards to boost its computational rate pretty high. There will be more in the next couple of years on the list. )

For problems that are Frame buffer sized ( some images want to present) + some small addiitional data structures ( some textures/effects to blend into the scene ) it will work alot better. Jobs were in previous tech generations the CPU was computing stuff so that a less capable GPU could display the finished results.... that stuff the trend will be more likely to shift over.




For parallel tasks, when your software supports it, it would make more sense to get a slower CPU, but a really nice GPU like an Nvidia Tesla or AMD FireStream

Perhaps not necessarily slower CPU but it may blunt the 8 , 12, 16, etc. core CPU from becoming the standard mainstream desktop/laptop configuration. Likewise not necesarily the Telsla or FireStream bleeding edge in performance on the GPU side either.

That is arleady in motion with the upcoming AMD's Fusion and Intel's Sandy Bridge that should be visibly numerous systems by early 2011. So would get to the greater than 6 core in the system processor package, but they all wouldn't be of the same type.
 
is it even possible to buy four 4GB at that price 650$?

At 1066 speed.... yes. ;)

http://eshop.macsales.com/shop/memory/Mac-Pro-Memory#1333-memory

also is it possible to have like 2x4GB and 1x2G and 1x1GB memory in the slots, or do you always need to use the same GB memory modules?

if use the OWC option mentioned above, then no. The answer is it depends upon
who get them from and what the specs are. You'd also be killing of a bit of performance since with an odd set of pairings the interleaving will be turned off by the memory controller(s). Data will get fetched somewhat slower from memory. You don't have to match them up. System runs slightly faster if you do though.


It would probably be happier if had 2 x 4GB and 2 x 1GB. Each one of those pairs could be matched up and interleaved.
 
Don't bother ordering any RAM until places like OWC confirm what's usable. Even they're not sure about the vague restrictions on these new Intel chips. For instance, you might not even be able to use 8GB DIMMs in the 6/8/12 cores.

You want to stay away from
filling all 4 slots. Barefeats prooved, adding a 4th stick OM
thev6 core slows it down. That's why
jmyou see stock 3gb not 4th really slows it down. Not why but it's true so the most you can use is 24 gb 8x3. Apple really messed stuff upmand looking at the 2.93 from last year is almost as fast as the 12. So I am getting Lord willing, 6 8 core last year or 12,as thememory brings it's close in price.

Very confusing.
 
Not why but it's true so the most you can use is 24 gb 8x3. Apple really messed stuff upmand looking at the 2.93 from last year is almost as fast as the 12.

I disagree with the first part of your post. Barefeats showed that the 4th stick slows down synthetic memory benchmarks significantly, but they also demonstrated that real-world apps get plenty of memory bandwidth without the tripe-channel benefit. In fact if your apps need memory, the 4th stick is fantastic to add.

The second part, I don't have a clue what you are trying to say Apple messed up. The triple-channel RAM architecture is Intel's. While it would be nice of the single socket MP had 6 RAM slots, it's really not a big deal in practice.
 
Question : Can somebody explain this hyperthreading thing?

basically hyperthreading allows the processor to que up twice as much data pr working core, this is only effective when the core is stalled due to lack of data on one thread, it can switch over to the second thread while data is catching up on the first thread. so in all it is basically making the most of your cores.

edit: aaa sorry about the double post, didn't realize i was in this thread :p
 
by god no.
get the 8-core, the basic configuration also has 6GB ram vs 3GB in the 6core version.
so basically you get two additional cores and twice the ram for $200 less
None and both are the best option actually :) It depends on your needs. The 6-core has 6 cores at a higher frequency. The 8-core has 2 more cores but all cores are at a lower frequency. Some appliances benefit from the higher frequency, others from the larger amount of cores.

However, there is something else to be considered: memory. The single cpu Mac Pro only uses cpu's that support a maximum amount of 24 GB of memory (see Intel specs). That would come down to 2 GB per core in case of the 6-core cpu. The dual cpu Mac Pro's on the other hand are able to drive up an awful lot of memory (200+ per cpu, see Intel specs). The limitation on this one comes from the amount of dimms and the amount of mem on a dimm. Right now the max. is 8GB per dimm and you get 8 dimm slots so that would make 64 GB in total. That's 8 GB per core. In other words: if you need a lot of memory the single cpu Mac Pro's aren't an option, the dual cpu ones are.

In the end it just isn't about the cpu, it's about cpu, memory, gpu and even i/o (hdd/ssd). If you want a really fast Mac Pro you need to finetune it so the cpu, memory, gpu and i/o are well balanced.
 
None and both are the best option actually :) It depends on your needs. The 6-core has 6 cores at a higher frequency. The 8-core has 2 more cores but all cores are at a lower frequency. Some appliances benefit from the higher frequency, others from the larger amount of cores.

However, there is something else to be considered: memory. The single cpu Mac Pro only uses cpu's that support a maximum amount of 24 GB of memory (see Intel specs). That would come down to 2 GB per core in case of the 6-core cpu. The dual cpu Mac Pro's on the other hand are able to drive up an awful lot of memory (200+ per cpu, see Intel specs). The limitation on this one comes from the amount of dimms and the amount of mem on a dimm. Right now the max. is 8GB per dimm and you get 8 dimm slots so that would make 64 GB in total. That's 8 GB per core. In other words: if you need a lot of memory the single cpu Mac Pro's aren't an option, the dual cpu ones are.

In the end it just isn't about the cpu, it's about cpu, memory, gpu and even i/o (hdd/ssd). If you want a really fast Mac Pro you need to finetune it so the cpu, memory, gpu and i/o are well balanced.
8GB RDIMM's work in the Quads and Hex core systems as well (single processor versions of 2009/10 systems).

But the additional DIMM slots can potentially reduce the cost for memory capacity, as smaller DIMM's can be used to atain it.
 
I think we need to focus on just the 2010 systems (rdimms are used in the 2006 Mac Pro, current ones use the standard ddr3 dimms). Like I said, you can put up to 24 GB in the single cpu machines and 64 GB in the dual cpu machines (due to the amount of slots and the fact 8GB is the largest amount on a dimm atm). Both machines support 1, 2, 4 en 8GB dimms. The single cpu has only 4 slots, the dual cpu has 8 slots (4 for each cpu).
 
Minus 8 yes. If not, you know more than Intel does...

I've comprised a small comparison of all the available cpu's for the Mac Pro (single and dual cpu). As you can see it's either 24GB or 228GB (per cpu).
 
Minus 8 yes. If not, you know more than Intel does...

I've comprised a small comparison of all the available cpu's for the Mac Pro (single and dual cpu). As you can see it's either 24GB or 228GB (per cpu).

I don't know, but OWC says a single cpu mac can use 32gb ram

http://blog.macsales.com/6620-owc-confirms-2010-mac-pro-single-processor-models-can-use-32gb-ram

This is apparently actually tested, and not based on specs. If intel tested a single cpu in a board with only three memory slots, then 24GB would be the max (since when they did the test 8GB simms were the biggest available)

I'm kind of shocked that an intel.com site appears to show that you cannot do what has been done.

I would have liked to see apple install more memory slots. A "single" cpu should have 6, and a dual cpu should have 12 (or 24)

edit: but I can only assume they had a reason to only put 4 memory slots per cpu on the 2010 macpro
 
Last edited:
I think we need to focus on just the 2010 systems (rdimms are used in the 2006 Mac Pro, current ones use the standard ddr3 dimms). Like I said, you can put up to 24 GB in the single cpu machines and 64 GB in the dual cpu machines (due to the amount of slots and the fact 8GB is the largest amount on a dimm atm). Both machines support 1, 2, 4 en 8GB dimms. The single cpu has only 4 slots, the dual cpu has 8 slots (4 for each cpu).
It's been proven that 8GB DDR3 RDIMM's work in the 2009/10 machines.

So Yes, the 24GB limit advertised by Intel is in fact wrong (seems they haven't updated it, as that number was generated prior to the release of 8GB RDIMM; or 8GB UDIMM, which also works, but is currently only available using Samsung chips - no one else has that high a density Unbuffered parts yet). :eek:
 
I know Avid's been terrible at updating their software for Mac, so I'm curious when or if they'd ever take advantage of multiple processors.

I'm assuming right now the 6 core would be faster at running MC, but does anyone have thoughts for future Avid updates?

Avid is multicore aware. It performs well while editing, distributing the workload along all cores but rendering out of avid does use's about 50% or less of all cores it gets confused while rendering it seems:(
 
I don't know, but OWC says a single cpu mac can use 32gb ram

http://blog.macsales.com/6620-owc-confirms-2010-mac-pro-single-processor-models-can-use-32gb-ram

This is apparently actually tested, and not based on specs. If intel tested a single cpu in a board with only three memory slots, then 24GB would be the max (since when they did the test 8GB simms were the biggest available)
It only shows that the OS sees the installed dimm modules but it doesn't show if it is able to use it. The same thing happened to the early core 2 duo systems. The chipset had a memory allocation bug so Apple limited it to 3GB instead of 4. You could put 4GB in the machine and it showed up as 4GB in system profiler and about this Mac however only 3 GB could be used. The last reply from OWC is a link to a list with the testing results in various configurations. But those are just what they are: numbers. That's about all you can find at OWC about it. Mac Performance Guide also questions it. The only source saying you can use 32GB memory is OWC and they not really have proven it to work.

edit: but I can only assume they had a reason to only put 4 memory slots per cpu on the 2010 macpro
They officially support up to 16 GB for those models so having 4 slots would make sense (4*4GB).

@nanofrog: thought you meant the dimms used in the 2006 models but you're correct about registered memory. I took that for granted since unregistered ecc is nearly impossible to find. I never doubted the 8GB dimms to work but that in absolutely does not mean the cpu is able to support a higher max than stated by Intel. It isn't proven that 4*8GB works properly in the single cpu Mac Pro. OWC is the only one I could find who claims it does and does very little to actually prove it. I think we need more user experience in this area.
 
It only shows that the OS sees the installed dimm modules but it doesn't show if it is able to use it. The same thing happened to the early core 2 duo systems. The chipset had a memory allocation bug so Apple limited it to 3GB instead of 4. You could put 4GB in the machine and it showed up as 4GB in system profiler and about this Mac however only 3 GB could be used. The last reply from OWC is a link to a list with the testing results in various configurations. But those are just what they are: numbers. That's about all you can find at OWC about it. Mac Performance Guide also questions it. The only source saying you can use 32GB memory is OWC and they not really have proven it to work.


They officially support up to 16 GB for those models so having 4 slots would make sense (4*4GB).
I realize what you're getting at here, but there's a serious flaw in the logic by DigiLloyd. Yes, the 24GB spec was determined from 6x DIMM slots, and Yes it was also via 4GB UDIMM's. Thus 24GB.

But that was done before 8GB DIMM's of any kind were out for DDR3 (Samsung's UDIMM's or any vendor making RDIMM's).

Now since the memory controller is actually the same between the SP and DP parts, RDIMM's actually do work in the SP versions of the Xeon line (35xx, 36xx, 55xx, and 56xx parts - all LGA1366 sockets and use the same memory controller - the difference is the 2nd QPI channel in the DP versions, which require a different chipset). The chipset does not matter in terms of memory capacity, as the memory controller is now on the CPU's die, not the chipset (memory controllers were located in the Northbridge in previous families, such as the parts used in the 2006 - 2008 systems, but changed in 2009 with the Nehalem architecture = memory controller on the CPU die). Systems engineering at it's finest (same memory controller on the LGA1366 parts, as ECC functionality is just disabled for the i7 variants). ;) :D

This can be confirmed by looking to non Apple sources if it works or not (SP servers/workstations), and the information demonstrates that it does in terms of the memory controller within the CPU's (8GB UDIMM or RDIMM works on Nehalem architecture).

What I don't get, is why Intel hasn't updated that spec since RDIMM's have released to market. :confused:

In the case of the i7 variants on the LGA1366 parts, it's no longer correct either, but only if the Samsung parts are used for 8GB DIMM's (RDIMM is not supported; UDIMM can work, but the ECC functionality won't, so it operates as standard non-ECC DDR3).

@nanofrog: thought you meant the dimms used in the 2006 models but you're correct about registered memory. I took that for granted since unregistered ecc is nearly impossible to find. I never doubted the 8GB dimms to work but that in absolutely does not mean the cpu is able to support a higher max than stated by Intel. It isn't proven that 4*8GB works properly in the single cpu Mac Pro. OWC is the only one I could find who claims it does and does very little to actually prove it. I think we need more user experience in this area.
See above.

The information is based on what the memory controller can actually do, and there's no evidence that Apple's firmware implementation is borked in this case (one user here is running the Samsung 8GB UDIMM's IIRC without issue, as does the 8GB RDIMM's by a few users as well).

Now it's possible that the memory capacity is reported at 32GB via UDIMM, but only work as 16 in a MP, but there won't be a gain in performance between 4GB UDIMM and 8GB UDIMM. No such report yet, though as 8GB UDIMM's are rare, it's not likely we'll see this one bore out one way or the other in a MP.

The Register chip on RDIMM changes things a bit (gets around the impedance problem of so many memory chips - Samsung did it by increasing the memory density in their memory chips thus keeping it within spec impedance wise). What reports we do have here on MR, seem to indicate that there is a significant performance increase (i.e. applications can use the additional capacity).
 
I know Avid's been terrible at updating their software for Mac, so I'm curious when or if they'd ever take advantage of multiple processors.

I'm assuming right now the 6 core would be faster at running MC, but does anyone have thoughts for future Avid updates?

They don't properly support it under Windows... so before it runs okay in Mac hell will freeze over. I find the interface actually works better in Mac though, snappier.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.