Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1210
Article Thumbnail

After Effects CC 2018: NVIDIA GeForce vs AMD Radeon Vega

Written on August 2, 2018 by Matt Bach
Share:

Introduction

While GPU acceleration has become fairly common in Adobe applications, in most situations it is much more important to have a powerful CPU, plenty of RAM, and fast enough storage. Despite this, a popular request we get is to compare AMD's Radeon Vega video cards to NVIDIA's GeForce cards. In previous articles we have compared these cards in Premiere Pro, Photoshop and Media Encoder, and today we will be rounding out our Adobe testing with After Effects. It is worth noting that while we will be focusing on After Effects performance in this article, choosing a specific GPU to use is a much more complicated topic. Many other factors including current pricing, reliability, power draw, noise level, and available cooler designs are all things that need to be considered.

If you would like to skip over our test setup and benchmark result/analysis sections, feel free to jump right to the Conclusion section.

Test Hardware & Methodology

For this testing, we will be using the following hardware and software:

This CPU, RAM, and storage combination we are using is among the fastest you can currently get for most tasks in After Effects which should give each GPU the chance to perform to the best of its ability. To compare AMD and NVIDIA, we chose a wide range of cards from both the Radeon and GeForce lines. We do want to point out that at the time we did this testing, it was difficult to source a quality AMD Radeon Vega card that was not factory overclocked. Rather than delaying our testing we decided to go ahead and use the overclocked cards even though it will slightly skew the results in favor of those cards.

In order to accurately benchmark the different systems, we used a range of After Effects projects that are mostly publicly available for download. The projects we used (along with their source) are:

Project Name Comp Name Tested Frames Notes
Countdown
by Rocketstock

(1920x1080)
Example Composition 0-40 (40 total frames)  
Electric FX
by Video Copilot

(1920x1080)
CloseCity
PlainSmoke
212-238 (26 total frames)
0-97 (97 total frames)
 
Animated Polygon
by Video Copilot

(1280x720)
Green Polygon 0-78 (78 total frames)  
GPU Stress
(3840x2160)
  32610-32710 (100 total frames)

4K H.264 video with:

  • Find Edges
  • Glow
  • Brightness/Contrast
  • Transform
  • Sharpen
  • Directional Blur
  • Lumetri Color
Cineware Party
by Cineversity

(1920x1080)
Party-Lite-004Full 0-169 (169 total frames)

"Video Wall" and "*.mov" layers
removed. Tested with C4D Renderer:

  • OpenGL
  • Standard (Draft)
  • Standard (Final)

RAM Preview - Raw Benchmark Data

[Click Here] to skip ahead to analysis section

RAM Preview - Benchmark Analysis

Since AE version 2015, we have seen a very sharp split in the type of CPU that works best for After Effects. While in the past a CPU with lots of cores would be great for everything, most effects and tasks in AE are now better with a CPU with fewer cores but a higher operating frequency. The exception to this is if you utilize the Cinema 4D CPU renderer where a high number of CPU cores can still make a massive difference in performance. Due to this, we have separated out our testing results between "standard" projects and those utilizing the Cinema 4D CPU renderer.


Before we get into the results themselves, we want to explain the scoring system we used to represent the average performance we saw with each GPU. In essence, a score of "20" would mean that on average that card was able to play our projects at 20% of the project's defined FPS. A perfect score would be "100" which would mean that the system was able to play it back in real time, although with the difficult projects we use for testing this should never actually occur.

Starting with the projects that use the C4D Renderer, we saw almost no difference in performance between any of the GPUs we tested. Since rendering with C4D is so heavy on the CPU, this is really to be expected as the CPU will almost always be a performance bottleneck. For the standard projects, however, we saw noticeably higher performance from the NVIDIA GeForce Cards. While pricing varies widely based on numerous factors like current sales or the popularity of bitcoin mining, in general you can think of the following cards as roughly costing the same amount:

  • AMD Radeon RX 580 8GB ~ NVIDIA GeForce GTX 1060 6GB
  • AMD Radeon Vega 56 8GB ~ NVIDIA GeForce GTX 1070 Ti 8GB
  • AMD Radeon Vega 64 8GB ~ NVIDIA GeForce GTX 1080 8GB

Using this as our primary point of comparison, we saw roughly 15% higher performance with the NVIDIA GeForce cards over their AMD Radeon equivalents when playing our standard projects.

Final Render - Raw Benchmark Data

[Click Here] to skip ahead to analysis section

Final Render - Benchmark Analysis

Once again, since the results should vary widely based on whether or not you utilize the Cinema 4D CPU renderer, we have separated our testing results between "standard" projects and those utilizing the C4D renderer:

In case you missed it explained in the previous section, the score shown in the chart above is a representation of the average performance of each GPU for this test. In essence, a score of "10" would mean that on average that card was able to export or render our projects at 10% of the project's set FPS.

Similar to the RAM Preview tests, once again the GPU really doesn't have an impact for the projects that use the C4D Renderer. For the standard projects, however, the AMD Radeon cards were slightly faster than their NVIDIA GeForce counterparts. It was only by 2%, however, so in the real world you likely wouldn't notice much of a difference.

Conclusion

NVIDIA GeForce vs AMD Radeon Vega After Effects CC 2018 Benchmark

The overall score in the chart above is a weighted average of our testing results based on what our customers tend to be the most concerned about. Overall, RAM Preview of standard projects is 40% of the score while the Ram Preview (C4D Renderer), Final Render (Standard), and Final Render (C4D Renderer) each contribute 20% towards the score.

With the results weighted in this manner, we get a pretty good estimation of what most users can expect in After Effects with each of these cards. Using the same rough pricing equivalents we used earlier (RX 580 ~ GTX 1060, Vega 56 ~ GTX 1070 Ti, and Vega 64 ~ GTX 1080) it is pretty clear that NVIDIA is going to give you higher performance in After Effects for your dollar. The exact performance you would see in your own projects will obviously vary based on exactly what you are doing, but in our testing the NVIDIA GeForce cards scored about 7% higher than their AMD Radeon equivalents.

Keep in mind that this is comparing factory overclocked AMD Radeon Vega cards against stock NVIDIA GeForce cards. While this probably didn't actually affect the results by a large amount, we would estimate that if we used stock AMD Radeon Vega cards the performance would likely be 1-2% lower than what we saw in our testing.

Tags: After Effects, Radeon, Vega, RX 580, GeForce, 1060, 1070, 1070 Ti, 1080, 1080Ti
Troy Guerrero

Radeon Pro Duo 32gb is much better than 1080 TI.

Posted on 2018-08-20 21:27:22

We have not tested that card in After Effects so it is possible that the Radeon Pro Due is faster than the GTX 1080 Ti, but I doubt it to be honest. The Pro Duo has a peak FP32 performance of 11.5 TFLOPS while the Vega RX 64 is 12.7 TFOPS. That is just theoretical, however, and since the Pro Duo is on the older Polaris architecture and the fact that multi-GPU has additional overhead that needs to be taken into account, the Vega RX 64 should actually be a bit faster relatively in real-world software like After Effects. Given that the GTX 1080 Ti was about 8% faster than the Radeon Vega RX 64, I would guess you are looking at somewhere around 15-25% higher performance with a 1080 Ti over the Radeon Pro Duo. Maybe a bit more even.

That is assuming that you are not maxing out on VRAM, however. If what you are doing in AE uses more than 11GB of VRAM, the Pro Duo will most likely end up being much faster than the 1080 Ti since running out of VRAM really affects performance.

Posted on 2018-08-20 21:40:11
Troy Guerrero

Please test Threadripper 2 in After Effects, thanks!

Posted on 2018-08-20 21:28:50

Actually in process of that right now, but we will likely wait to publish until the end of the month or early next month after the 2950X 16 core is out. The 2990WX 32-core is really interesting, but AE in general doesn't work well with high core counts so it very likely isn't going to be faster than the 16 core in most cases.

Posted on 2018-08-20 21:31:02
TA Nie

Matt, hurry! We're going to start buying replacement editing stations soon! 20$ bro!

Posted on 2018-08-29 17:05:22

I'm actually just waiting on the Threadripper 2950x 16 core, but unless it somehow is vastly better than what we saw with the 2990WX or the older 1950X, the new Threadripper CPUs are still well behind Intel for After Effects. So don't hold off on new systems, just get Intel CPUs like what we list in our AE workstations https://www.pugetsystems.co...

Posted on 2018-08-30 04:30:51