Jump to content

Archived

This topic is now archived and is closed to further replies.

BigAl3D

New Mac Pro Specs Out

Recommended Posts

For those of you considering the new MP, what configuration do you think you'll choose? I'm sure there will be wide price range from the base $2,999 to a fully maxed out version near $10K.

 

  • # Cores (4-12)
  • GPU (D300-D700)
  • SSD (256-1024)
  • Memory (12-64)

Share this post


Link to post
Share on other sites

Setting aside the fact that many existing Mac Pro customers may not be happy with the new design (while I'm not 100% happy with it I like what I see enough that I'll be getting one), and given that the other Mac hardware discussion thread was closed... here's my understanding of how hardware under OS X benefits C4D usage. Someone tell me if I'm wrong.

 

 

1) Other than rendering, very little in the application is threaded for multiple CPUs, so getting 12 cores or 8 cores over 6 will not help you in non-rendering functions.

 

2) Since many processes use a single core, the best thing you can do to speed up C4D workflow on a Mac, is get the fastest CPU (in this case the 6-core 3.5GHz model). [That's assuming you find the 4-core model's stock RAM and GPU to be unacceptable based on the price... even though that model has 3.7GHz cores.]

 

2a) Is the typical setup in C4D that the CPU processes the imagery, and then passes it to the GPU for rendering in the viewport? Will improved OpenGL support in Mavericks, combined with the Fire Pro based GPUs in the new Mac Pro further improve the speed of operations, beyond the speed of the CPU? Or does C4D have to be updated explicitly to take advantage of the OpenGL features now present in Mavericks (tessellation shaders, et al)?

 

3) Nothing in C4D uses CUDA or OpenCL, nor cross-fire or dual-GPU processing... nor will it, so again we're looking at a single GPU's raw processing power and VRAM as the only factors.

 

FWIW it's a little disappointing to see that Adobe can optimize so many features in AE, Pr, and Ps for CUDA AND OpenCL, yet this is not on MAXON's radar. I'm not a programmer so I have to take MAXON's word that it's "not worth it". However I'm not sure that passes the smell test since we're talking about computational operations that should be very similar (i.e. image processing algorithms). 

 

Getting back to the specs and buying decisions, WRT to item 3 what I'm unclear on for C4D workflows is, where does the law of diminishing returns take over as far as having a GPU with 2GB of VRAM vs. 3 vs. 6? And do the extra stream cores in the two higher end cards, or their extra bandwidth, make any difference for major parts of the C4D workflow? Put another way: will it help users in the "time = money" department, on a day to day basis?

 

My guess would be, at the speeds we see for the AMD FirePro cards (which the new Mac Pro cards are derivatives of -- they're custom, not retail layouts), that more VRAM is more important than the additional processing units. But since the GPU is not used for rendering AFAIK, I wonder if the 3GB model is more than enough for C4D, and more than that is overkill. Also the higher end GPU will run a lot hotter than the other two based on what I've read.

Share this post


Link to post
Share on other sites

also dont forget about the constant problem many of the members of this forum had with c4d and amd cards.

 

Assuming you mean PC users as there are only a handful of AMD cards running on modern Macs (with Apple drivers)? I suspect the drivers for these cards will be highly tuned compared to older cards / that Apple and AMD will have put the time in on these, given how much Apple is promoting the dual GPU setup. They obviously feel this is the direction the industry (if not MAXON specifically) is going and are trying to set the trend. For them to ship shoddy drivers like they've done in the past would be a great way to torpedo the product before it ever gets traction.

 

In any case I hope someone from MAXON will chime in here and let us know where the money is best spent for this particular machine type.

Share this post


Link to post
Share on other sites

Setting aside the fact that many existing Mac Pro customers may not be happy with the new design (while I'm not 100% happy with it I like what I see enough that I'll be getting one), and given that the other Mac hardware discussion thread was closed... here's my understanding of how hardware under OS X benefits C4D usage. Someone tell me if I'm wrong.

 

 

1) Other than rendering, very little in the application is threaded for multiple CPUs, so getting 12 cores or 8 cores over 6 will not help you in non-rendering functions.

 

2) Since many processes use a single core, the best thing you can do to speed up C4D workflow on a Mac, is get the fastest CPU (in this case the 6-core 3.5GHz model). [That's assuming you find the 4-core model's stock RAM and GPU to be unacceptable based on the price... even though that model has 3.7GHz cores.]

 

2a) Is the typical setup in C4D that the CPU processes the imagery, and then passes it to the GPU for rendering in the viewport? Will improved OpenGL support in Mavericks, combined with the Fire Pro based GPUs in the new Mac Pro further improve the speed of operations, beyond the speed of the CPU? Or does C4D have to be updated explicitly to take advantage of the OpenGL features now present in Mavericks (tessellation shaders, et al)?

 

3) Nothing in C4D uses CUDA or OpenCL, nor cross-fire or dual-GPU processing... nor will it, so again we're looking at a single GPU's raw processing power and VRAM as the only factors.

 

FWIW it's a little disappointing to see that Adobe can optimize so many features in AE, Pr, and Ps for CUDA AND OpenCL, yet this is not on MAXON's radar. I'm not a programmer so I have to take MAXON's word that it's "not worth it". However I'm not sure that passes the smell test since we're talking about computational operations that should be very similar (i.e. image processing algorithms). 

 

Getting back to the specs and buying decisions, WRT to item 3 what I'm unclear on for C4D workflows is, where does the law of diminishing returns take over as far as having a GPU with 2GB of VRAM vs. 3 vs. 6? And do the extra stream cores in the two higher end cards, or their extra bandwidth, make any difference for major parts of the C4D workflow? Put another way: will it help users in the "time = money" department, on a day to day basis?

 

My guess would be, at the speeds we see for the AMD FirePro cards (which the new Mac Pro cards are derivatives of -- they're custom, not retail layouts), that more VRAM is more important than the additional processing units. But since the GPU is not used for rendering AFAIK, I wonder if the 3GB model is more than enough for C4D, and more than that is overkill. Also the higher end GPU will run a lot hotter than the other two based on what I've read.

 

 

1. Thats correct. I am not aware of any other multi-threaded feature within C4D. Unless things have changed, I believe shadow rendering (soft shadow pre-rendering) is done with only one thread.

 

2. Correct again.

 

3. Correct. No CUDA or OpenCL features. Only OpenGL for viewport speed/playback

 

VRAM only comes in to play when dealing with shadows, viewport textures (viewing them at a high res), and large complex scenes. 1 to 2 GB's of VRAM should be plenty for most C4D users. However, there are some circumstances when more would be needed, for example, I use TFD (Turbulence FD) which is a smoke/fire simulation plugin. It uses the GPU to simulate, and uses CUDA. It is HIGHLY dependent upon VRAM. If you dont have enough, it will automatically change over to the CPU for simulating. So more VRAM is a big plus when using it.

Share this post


Link to post
Share on other sites

1. Thats correct. I am not aware of any other multi-threaded feature within C4D. Unless things have changed, I believe shadow rendering (soft shadow pre-rendering) is done with only one thread.

 

2. Correct again.

 

3. Correct. No CUDA or OpenCL features. Only OpenGL for viewport speed/playback

 

VRAM only comes in to play when dealing with shadows, viewport textures (viewing them at a high res), and large complex scenes. 1 to 2 GB's of VRAM should be plenty for most C4D users. However, there are some circumstances when more would be needed, for example, I use TFD (Turbulence FD) which is a smoke/fire simulation plugin. It uses the GPU to simulate, and uses CUDA. It is HIGHLY dependent upon VRAM. If you dont have enough, it will automatically change over to the CPU for simulating. So more VRAM is a big plus when using it.

 

That's all useful information, Vilandra. Thank you!

 

WRT to VRAM I am considering a Turbulence FD purchase (X-Particles 2 also in the mix)... would the "mid range GPU" (D500) be acceptable for this use with 3GB of VRAM? Minimal switching in that case or can the TFD simulations get huge in a hurry, so 6GB much better than 3GB (as opposed to suffering from the law of diminishing returns where it doesn't help that much)? Any idea if X-Particles or other plugins are GPU-accelerted? I don't see anything on their system specs / FAQ page to suggest it is but who knows... maybe one of you guys has heard the developer speak on this issue.

Share this post


Link to post
Share on other sites

That's all useful information, Vilandra. Thank you!

 

WRT to VRAM I am considering a Turbulence FD purchase (X-Particles 2 also in the mix)... would the "mid range GPU" (D500) be acceptable for this use with 3GB of VRAM? Minimal switching in that case or can the TFD simulations get huge in a hurry, so 6GB much better than 3GB (as opposed to suffering from the law of diminishing returns where it doesn't help that much)? Any idea if X-Particles or other plugins are GPU-accelerted? I don't see anything on their system specs / FAQ page to suggest it is but who knows... maybe one of you guys has heard the developer speak on this issue.

 

The more VRAM the better when dealing with GPU accelerated stuff like TFD. TFD has a meter in the sim window that shows how much VRAM is being used (see attached image). If the meter goes over the maximum amount of VRAM your GPU has, it automatically selects the CPU.

 

Depending on how big your TFD sim is, you can easily exceed the GPU VRAM limit. But the most important part is making sure the voxel size is low enough as to not exceed your VRAM amount. TFD gives you control of the voxel size (or resolution) of the sim quality.

 

As far as X-Particles, I am not aware of any GPU feature for it, however, it is multi-threaded. Ive used it before and the viewport speed with X-Particles vs the standard Thinking Particles is a HUGE night and day difference.

post-5982-0-38537700-1383694523_thumb.pn

Share this post


Link to post
Share on other sites

The more VRAM the better when dealing with GPU accelerated stuff like TFD. TFD has a meter in the sim window that shows how much VRAM is being used (see attached image). If the meter goes over the maximum amount of VRAM your GPU has, it automatically selects the CPU.

 

Depending on how big your TFD sim is, you can easily exceed the GPU VRAM limit. But the most important part is making sure the voxel size is low enough as to not exceed your VRAM amount. TFD gives you control of the voxel size (or resolution) of the sim quality.

 

As far as X-Particles, I am not aware of any GPU feature for it, however, it is multi-threaded. Ive used it before and the viewport speed with X-Particles vs the standard Thinking Particles is a HUGE night and day difference.

 

You are, as they say in the business, "the man". :) 

 

Sounds like 3GB should be enough for some tasks (just based on what I see in that screenshot, but perhaps not for sims with huge numbers of voxels). So it will come down to a choice between having that extra 3GB of VRAM and the extra noise and heat that comes with it... or occasionally having to "offload" a sim to CPU (which likely will be 6 cores at 3.5 GHz, so no slouch by any means). Probably will save that money, add more RAM (for AE multiprocessing among other things) and go with the D500. 

 

Once I've gone the route I chose I'm sure I'll be back in here reporting my findings. Actually excited to see what the new C4D rendering options can do with just the 6 processors (12 threads) too, and the speed that comes from the lack of a spinning magnetic platter anywhere in the system.

Share this post


Link to post
Share on other sites

Apple recently started to engineer their products to break, so I can't really trust buying anything Apple.  Some examples are making the iphone screen break if the code is not put in right, nickel connecters in the phone so it will instantly rust when moisture is introduced, them gluing the fragile front glass to the lcd screen so it's very expensive to replace, and the design of the new iphone 5 seems to make the dropping force needed to break the screen even smaller.  Not to mention the warranty starting when the the store you bought it from bought it which can mean you have maybe 1 year less warranty then what you thought.  Don't buy apple.

Share this post


Link to post
Share on other sites

Apple recently started to engineer their products to break, so I can't really trust buying anything Apple.  Some examples are making the iphone screen break if the code is not put in right, nickel connecters in the phone so it will instantly rust when moisture is introduced, them gluing the fragile front glass to the lcd screen so it's very expensive to replace, and the design of the new iphone 5 seems to make the dropping force needed to break the screen even smaller.  Not to mention the warranty starting when the the store you bought it from bought it which can mean you have maybe 1 year less warranty then what you thought.  Don't buy apple.

 

So because you've had some bad experiences with their consumer phone product, Apple is now "engineering" all their products to break, and "don't buy Apple"? Your axe to grind is pretty clear here. My iPhone 4s has been dropped multiple times, gotten liquid on it, and endured a lot of clanking around in my car compartments, on my desk, etc. Not so much as a scratch and works flawlessly after roughly 2 years of use. My friends also have iPhones. Zero complaints. iPods used for workouts and being bounced around all the time... zero complaints.

 

My iPad 3 the same. My Mac Pro lasted almost 5 years before i had a problem with it, and that endured a fair amount of clanks from chair legs and other abuse as well (namely being dropped by the delivery man from a height of about 5 feet). The Mac Pro before that lasted about 4 years also and could've lasted longer. MacBook Pro lasted through 2 years of frequent presentations and use, sold to family member, still going strong after another two years. Zero screen defects, etc. Wife's MacBook going strong after 5 years. Sisters the same.

 

So I guess I'll keep right on buying Apple like I have for the last 20 years, with zero significant problems to speak of in the last 10-12, with any of their products. Only the older, pre-OS X products ever gave me cause to take it into the shop. Their products are far more durable and reliable than any I've bought or seen from: Dell, HP, or Acer. All of those use cheaper components, connectors, and have worse customer service.

 

 Also your warranty theory is flat out false. The warranty is exactly what it says it is on the EULA document. Never heard anyone being turned away (for any product) that was returned with a defect during warranty period, which is one year from the customer's date of purchase. Maybe some shady retailers not affiliated with Apple tries to make that claim but Apple itself does not. In fact almost every instance I've ever seen, people took the item into the Apple Store for a fix, and they usually walked out with a brand new replacement.

 

Your assertion is without substantial evidence and patently ridiculous, especially since not one person has yet even had the opportunity to use the new Mac Pro, let alone surmise it's "designed to break". Surprising to see that kind of BS in a technical forum like this one. I'd expect to see nonsense like that in Macrumors.com forums, not C4D Cafe.

Share this post


Link to post
Share on other sites

Good for you Zmotive, but most everything I said was a technical fact from an engineers perspective not a rumor.  Most of it is with newer apple products, so I'm not surprised you have not had a problem with your 5+ year old devices.  It's not my experience as I have never and will never buy an apple product.  On top of that the DRM on their stuff alone is enough reason not to buy.  I also hear people getting automatic updates to their OS which either makes their computer or device work slower or not at all.  I hear too many bad things about apple to ever buy any of their stuff.

 

The warranty thing does only happen if you buy from a fishy reseller.  Apple would not want the class action against them.

Share this post


Link to post
Share on other sites

^Oh please cut the .  You didn't have to bring any of this fud in here.  This topic has nothing to do with what ever it is you are selling.  What ever rumor or what ever bullshit story you read on the "interwebz" keep it to yourself because thats all it is, bullshit.

 

I will now laugh out loud at your comment..

:lol:

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.

YOUTUBE CHANNEL:

ABOUT US:

C4D Cafe is the largest CINEMA 4D community. We provide facilities for discussion, showcasing and learning our favorite software :) Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, get your own private messenger, post status updates, manage your profile and much more. If you need to find solution to your problem or otherwise ask for help, Cafe is the right place.
×
×
  • Create New...