Teddy Gage
Forum Replies Created
-
@Eric – Awesome, would love to see the 3930 vs. 4930 results, let me know if you get a chance to run it.
-TG -
Hi folks, thanks everyone for contributing to this project, it was a real eye-opener. However I am no longer supporting this benchmark, I have a new project that does total CPU / RAM / GPU testing in AE CC here:
https://www.teddygage.com/AfterBenchCC/
RESULTS:
Add yours! email me at fstopdigital@gmail.com or on twitter @teddyrocksteady
@ Paul Roper –
I would NEVER rely on a hackintosh for real production work. They are too unreliable even if you know what you are doing. And even if built by a 3rd party (which will never happen, because they are breaking the law) you need to know how to work with a PC way before you’d consider messing around with faking OSX bootloaders and the like. They require a very specific combination of hardware and many features remain unsupported. Not to mention the hardware will be limited – you are probably losing a good percentage of CPU performance because it is not perfectly tuned for the OS.
You may be better off saving up for a new mac pro, although I’ve heard mixed reviews.
Honestly, my advice would be, if you are looking to stay Mac, is to look on the used market (ebay, classifieds) for a topped-out, but used, last-gen Mac Pro with the specs of your old work computer. You may be surprised what you can get for way less money.
Additionally, I have heard firsthand that the new mac pro GPU implementation is really crappy. The only app to seriously take advantage of their power is FCPX, and who cares about that. Not to mention zero CUDA or directX support. So I’d recommend looking at a PC with 2x GTX Titan or 2x GTX 780 in SLI if graphics is truly your priority. Yet you claim you need the speed solely for AE, and the GPU dependent effects are few and far between, unless you use AE raytracing all the time. People seem to believe GPU performance is all that matters when that couldn’t be further from the truth.
So I would save money and focus on getting a capable graphics card and spending money on CPU – you would want an overclocked intel i7 3930K or 4930K hex-core CPU – which in many tests on the benchmarks above performs much faster than dual-quad xeons. In a pc environment it is currently the fastest consumer processor on the market for after effects.
If you are out for pure speed in AE, you should switch to windows. I’m sure people will disagree, but there is no mac out there that touches the price / performance ratio of a custom built pc workstation. I’m not biased: I own and use both systems, but I would never do 3D work on a mac, and only do AE stuff on OSX in a pinch.
-
In the post you just responded to – a google docs spreadsheet for now
or the shortened URL
-
Thanks Walter, here are the results so far
I’m looking especially for new Mac Pro and 4930K processor, please either post results here, email to fstopdigital@gmail.com or on twitter @teddyrocksteady
-
Yes, Gtx 5xx series have better CUDA performance than 6xx and even 7xx series cards. This is because nvidia gimped the cuda performance of these cards in favor of open gl performance. They are fantastic cards for gaming but not graphics. Currently the gtx 580 is the best performing cuda card for the money, and as the cpu is not involved, its doesnt matter how many cores you have. HOWEVER this only applies to the raytracing engine. Once you have normal 3d engine engaged, six core machines will dominate, and even I only occasionally use raytrace engine for graphics. The gtx 680 is quite a good card for open gl 3d using cinema 4d for example, so consider that as well.
-
We have a winner. Where should I send your free totebag?
😀
-
I still think two 3930K machines are preferable to a single dual-cpu xeon setup. No matter how many cores you throw into one machine, it will still be tied up during render time. I have two master workstations in a dual-dvi KVM switch setup, and there’s always one rendering and one working. and if I need both rendering I have a networked 8TB NAS. One machine has a GTX titan, the other has a dual GTX 580 SLI setup for heavy CUDA and transcoding. That machine also has two SSDs in RAID 0 maxing out sata 6 bandwidth (500+ MBps IO) for disk-intensive fluid simulation etc. Not to mention it’s much more affordable, and a setup that’s easier to troubleshoot and fix (along with system backups of each machine). I get around 13.5 CB score on the 3930K @ 4.8 gHZ
-
You didn’t clear the cache between renders, so it was just spitting out the previously cached frames. hence the 1 sec render. You’ll have to clear the disk cache and RAM to get a new test.
however MP is irrelevant for this test, as it is turned off when artisan (the raytraced renderer) is used. Those results are in line with your gpu
-
The Titan is fantastic for Resolve and Premiere pro. Extra video memory and they can use all CUDA cores. Not sure about ability to use both cards on 590 or 690. Plus less heat and power generated. A Quadro is really only necessary for 3D apps that need viewport accuracy and performance, or if you need a 10bit monitor support / SDI output.
I highly recommend the Asus boards that Ian mentioned. As well as a better cooler. I’d recommend the corsair h80 or h100. That chip was meant to be overclocked. It is very easy, even for beginners.
-
AE being, of course, another exception.
I did some research of my own, and I don’t usually say this, but I think you are right. Although it really depends on a) what your definition of “faster” is and b) what you’re using the card for. I wasn’t aware of the massive difference in viewport performance for maya etc. Now whether that built into the drivers or is a factor of unlocked hardware is hard to tell. I would favor an answer in software. But I would disagree about the Titan being overpriced. It is really an amazing card. Especially once you’ve seen what it can do on multi-monitor 2k openGL performance, while running cool, not requiring SLI, having 6GB RAM and using fewer watts. Add to this the ability to enable double precision floating point calculation. I honestly believe GPU rendering is the future, and this card can compete and outperform a $3000 “pro” equivalent on that front, so I think it is a great value. Just maybe not to your average gamer who just wants to run the latest call of duty.