New Mac Pros

pegot

pegot

0
New Mac Pros

Since the new Mac Pro coming out later this fall will only be running AMD cards without (as currently known) any options to replace or upgrade the default GPU's (they may be soldered in) I suspect my next computer will be a PC.

But I don't think I could live without Cheetah and so I might consider building a Hackintosh.

Does any one in this community currently run Cheetah on a Hackintosh?

Would also love to hear other people's thoughts on the new Mac Pro's lack of CUDA support and what that might mean for your future 3d work and system setup.
 
Apple is clearly pro OpenCL and not CUDA, which is unfortunate for those of us who want to work with Octane Renderer in the short term, but I suspect better for everyone in the long-term. (And from what I've been reading, OpenCL has caught up to CUDA in terms of functionality and tooling, so the advantage of CUDA is going to disappear — especially since if you have a Haswell CPU/GPU combo and a high-mid-range NVidia card you've actually got more OpenCL horsepower in the Intel part than the NVidia part.

I wish Apple would simply allow higher-end users to buy a $500 dongle for PCs and let folks build their own (and Apple could even bless certain parts, right?).

Anyway, I'm not a 3d professional and I don't rely on my Mac for major rendering tasks so it doesn't bother me and the new Mac Pro is awesome, but if you're relying on a CUDA solution right now Apple isn't offering much.

I've been waiting for Apple to announce transparent cloud computing via Grand Central Dispatch. Theoretically Apple could offer scalable cloud-rendering transparently at OS level thanks to GCD + OpenCL but if they're planning it they haven't announced it. (This would basically give any Mac a cloud-based render farm…)

One thing to note: there was a suggestion in the keynote of expansion via thunderbolt (there's already third-party boxes of PCI slots that hook up to Thunderbolt Macs — so you can hook your Macbook Air up to a bunch of NVidia cards in an external box).
 
Otoy / Refractive has already developed a cloud based rendering system for Octane but I don't know when it will be released to the public yet. I read too that Open CL has vastly improved, but others say it is still not comparable to some CUDA functions - I don't know one way or another. But I do have Octane, 3d Coat, and just recently purchased Video Co-pilot's Element 3d and all those work on CUDA technology.

Why do you feel pushing us to OpenCL is better in the long run? Big Brother (I mean Apple) always claims they know what's best for the consumer but though I have invested into the Apple ecosystem I find it is harder for me to accomplish the things I want than if i were just using a PC. GPU expansion being one of them.

I too am attracted to a lot of what the new Mac Pro promises, and I wonder to what leaps Martin might take Cheetah to in being able to use the two AMD cards it will ship with. But being locked into these two cards for the life of a product, as powerful as they might be, is simply not acceptable for a machine that calls itself a "pro" workstation. If I would be spending so much money for a new workstation, I absolutely require the freedom to put in any damed cards I see fit for my work flow.

The Thunderbolt expansion boxes may be an option but from everything I have read thus far there is no proven solution to use them with GPU's yet. And performance, as opposed to internal cards, may also be an issue. Yet another sticking point with that solution is bound to be the price. After spending several grand on a new Mac Pro, one is now expected to shell out another grand for the necessary expansion boxes (I think most Thunderbolt expansion boxes would start around $800 - then there are the additional boxes and or cable adapters one needs for all your drives - is there even an adapter yet for Thunderbolt to eSata?).

BTY Podperson - I love your new animated avatar. Much less intimidating than the tentacle :)
 
Last edited:
OpenCL lets you write code once and it will run on (theoretically) anything with driver support, and in practice x86, integrated GPU, NVidia GPU, AMD GPU (I imagine ARM Cores as well should that ever become a significant case). CUDA only gives you NVidia. I really don't see how that can't be better than being locked into NVidia in the long run (and I imagine that Octane is probably being ported to OpenCL as we speak).

I too wonder how upgradeable the new Mac Pros will be. Given the relatively slow progress of CPUs lately, I imagine that a non-upgradeable new Mac Pro will be out of date before my 2012 Mac Pro is ;-) I guess we'll wait for the reviews (and pricing). If this is a $1500 product that's one thing. If it's a $5000 product that's another thing entirely.

Anandtech's review of the new Intel Haswell chipsets show Luxrender getting more performance out of Intel's midrange integrated GPU than a midrange discrete NVidia GPU (I can only assume NVidia hasn't spent much effort on their OpenCL drivers).
 
Last edited:
As a once (and still) proud owner of a 2006 vintage liquid-cooled quad G5 Mac (now retired, to be handed down to my grandchildren to take on Antiques Roadshow in 2050), I'm definitely in the market for an updated "Pro" machine. In recent years I've been working on MacBook Pros and Mac Minis, which are good for iOS development work, not bad at all for 3D modelling, and OK as far as rendering is concerned. The latter is less of an issue for me as I'm typically importing Cheetah models/animations into Unity rather than doing a lot of renders, but I still could do with a beefier platform to fully accommodate my workflow. So I've been waiting patiently for the announcement of the new Mac Pro since Tim Cook mentioned last year that Apple hadn't abandoned the platform completely and were working on "something great".

Whether you regard it as a real "Pro" machine or not, it's a stunning piece of engineering. It had to be a beast to live up to expectations - and in many ways it appears to be - but I have to say I still feel a tinge of disappointment about the out of the box (or should I say tube?) capabilities. For some reason I still don't understand, Apple has always provided rather lame default graphics card options which lagged behind the state of the art available in the PC world. OK sure, dual AMD FirePros with 6GB each certainly represent - at least right now - a formidable amount of horsepower. But as pegot said, upgradeability is key for a Pro machine. (There is some speculation on the Apple forums about whether these might sit on removeable daughter boards, which if so could offer a relatively straightforward upgrade path). Or perhaps I'm just not getting Apple's new thinking for "Pro" machines, ie an effectively sealed interior which you upgrade every two years - most of your storage etc being external. As Pod implies - that would be difficult to swallow at a $5K entry price point but maybe acceptable if a lot cheaper. We'll know more later in the year, though I'm not holding my breath...

Despite the progress of OpenCL, as a compute fan (I'm particularly interested in the numerical simulation of black hole collisions) I would have been pleased to see an nVidia-based CUDA piece as an configurable BTO graphics option, such as an integrated Tesla part. (And maybe that will still happen). However given the size and power requirements of Tesla, I don't know if the Kepler GK110 GPUs could be integrated into the new Mac Pro in the same way the FirePros have been. In any case, one Tesla card is never enough - just ask the Titan folks at Oak Ridge.... ;-). Perhaps Thunderbolt 2 will offer some distributed compute solutions, as Pod suggests.

Whatever, I too would love to hear Martin's thoughts on how Cheetah might take advantage of the new Mac ("Darth") Pro hardware, and in particular the evolution of the Cheetah renderer. I know that Pod has previously expressed strong views on the Cheetah renderer in this forum, and good though it is, some folks prefer to use other renderers such as Octane which can take advantage of GPU acceleration. If Cheetah goes OpenCL, I'm ordering one!
 
Hi,
to be honest I'm a little bit speachless concerning Apples announcement. Until the announcement of the new MacPro I would have taken every bet that OpenCL is an end of live technology on the Mac since Apples interest in supporting OpenCL (fixing bugs, answering emails concerning OpenCL, etc.) was close to zero.

In its current state OpenCL is not very reliable for complex computations. For example OpenCL theoretically runs on Intel CPUs, AMD GPUs and NVIDIA GPUs. But in practice I found cases where I got three different results depending on which CPU/GPU I ran the OpenCL kernel.:confused:

Apple also has the tradition not to build in the fastest GPUs in its hardware. The result is that probably in 95% of the installed Macs the CPU is much more powerful for OpenCL than the GPU. Even on the latest Retina MacBook Pro the CPU is dancing circles around the GPU when it comes to OpenCL performance. So why limiting the Ram to 1 GByte of GPU VRAM when the CPU has access to 16 GByte or more? Why using OpenCL if it only pays out for maybe 5% of the current users?

For simple task like calculating image filters the picture might be different. For complex calculations like raytracing Apple still has to do quite some homework on the software (OpenCL framework). Lets hope Apple finally wakes up and gets some serious interest in improving OpenCL. Then the developers will finally use it.

CUDA can be considered as pretty dead on the Mac. With no Nvidia option in the MacPro I would be surprised if too many developers would support it on the Mac anymore.

So the current situation on the Mac is that CUDA makes no sense anymore and OpenCL is pretty immature. :confused:

For the near future the CPU is therefore the best bet for raytracing on the Mac. And I have quite some improvements to it in the queue. In parallel I will work on my experimental OpenCL renderer so that I have something to release once the time has come.

Bye
Martin

P.S. On the PC the situation is completely different of course.
 
Last edited:
Sweet, I already ordered my GPU monster.:smile: Looks like this years christmas is in January.:redface:

Bye
Martin
 
Martin said:
Yes, I couldn't resist. The D700 upgrade was priced very fairly so I took the big GPU. I can't await to start playing with OpenCL on it.:smile:

Bye
Martin
Click to expand...

Santa was very generous this year ... :cool:
So that day will surely be a long day ...
I love to see some nice pictures of that beast once you get it.
I mean other pictures than those from the marketing team

And when you will publish the gecko/marty time benchmark on the forum, the crowd will go nuts ... :smile:
 
Cheetah makes use of 8 or 12 cores, D700s

I'm planning to buy a 6 core nMP with d500s fairly soon - after the dust settles a bit and I can read some reviews.

My work is ~ 80% Photoshop and 20% Cheetah. From what I've read, PS won't take advantage of more than 6 cores or the d700s (don't quote me on this).

Martin, or anyone in-the-know, would Cheetah3D experience much performance gain by upgrading to 8 or 12 cores? How about upgrading to the d700s?

Thanks in advance!
--Shift Studio.
 
Hi,
Cheetah3Ds renderer and some other performance critical algorithms are multi-threaded and therefore profit from additional cores. Especially rendering profits almost linearly from more CPU cores.

I'm quite sure that the interest in OpenCL will increase once the new Mac Pros ship. And therefore also more applications will use OpenCL. So having good OpenCL performance it definitely important for the future.

Photoshop CS6 already supports OpenCL. I think it is already turned on by default in the performance preferences.

Bye
Martin
 
Martin said:
Hi,
Cheetah3Ds renderer and some other performance critical algorithms are multi-threaded and therefore profit from additional cores. Especially rendering profits almost linearly from more CPU cores.

I'm quite sure that the interest in OpenCL will increase once the new Mac Pros ship. And therefore also more applications will use OpenCL. So having good OpenCL performance it definitely important for the future.

Photoshop CS6 already supports OpenCL. I think it is already turned on by default in the performance preferences.

Bye
Martin
Click to expand...

Wow, so a 12-core machine could render ~ 2x faster than 6-core all else being equal.

How about d500s vs d700s? whats the approximate performance gain there?

So if Cheetah use was my main objective, the upgrades would be well worth it. Right?

Thanks,
--Shift Studio.
 
I wouldn't go for more than six cores (the six core machine has more RAM capacity). The extra cores are super expensive — e.g. about the same price as another Mac Pro (and actually slow down your single-core speed), and the fact that Martin just bought a D700-powered Mac Pro should tell you something. He's going to implement OpenCL support (I'm guessing it will be the headline C3D 7.0 feature — that's what I'd do in his place).

Ars Technica actually recommends you get the 6-core box with D700s with only as much RAM as you need as the best value-for-money (the D700 upgrade is actually a bargain according to them; the equivalent card for PCs costs 3x more!). Their test machine is cheerfully rendering 12 4k video streams at once.
 
Last edited:
Shift Studio said:
Wow, so a 12-core machine could render ~ 2x faster than 6-core all else being equal.
Click to expand...

No, not really since the 12 core CPU runs at just 2.7 GHz while the 6 core runs at 3.5 GHz.

The number of cores multiplied by CPU clock rate gives a better idea of the real performance.

6 * 3.5 GHz = 21,0 GHz
12 * 2.7 GHz = 32,4 GHz

And as you can see it's not even close to twice as fast.

Shift Studio said:
How about d500s vs d700s? whats the approximate performance gain there?
Click to expand...

At the moment Cheetah3D doesn't profit too much from the GPU but I want to change that. Since I don't even have my MacPro yet it's difficult to say when that will be the case. v7 would also have a rich feature without a OpenCL renderer.:wink:

I personally picked the d700 not just because it's faster but also because it has twice the memory. And you can never have enough video RAM once Cheetah3D offers a GPU renderer.

So the faster GPU is a good future investment at the moment.

Shift Studio said:
So if Cheetah use was my main objective, the upgrades would be well worth it. Right?
Click to expand...

I picked the 6 core since the price for the 12 core is indeed pretty steep. Especially for an approx. 50% speed up.


Bye
Martin
 
Last edited:
Thanks Pod and Martin!

I think it'll be six core for me too and probably the d700s - I don't want to get a couple of years down the road and regret buying d500s.

I'm pretty excited about a hot new machine!

--Shift Studio.
 

I'll probably have to get a quad, and the d700 cost 1k more in that configuration.:redface:
Is it worth it, or should I sell my 1970 custom Les Paul Custom and get the six core d700.
Anyone interested? :roll:

gibpai27a.jpg

​
 
Last edited:
Hi,
sounds like you love your guitar. So maybe its better to wait until Cheetah3D really can utilize the new GPUs.

Bye
Martin
 
You must log in or register to reply here.
Back
Top