Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1116
Article Thumbnail

Premiere Pro CC 2018 Workstation GPU Performance

Written on February 16, 2018 by Matt Bach
Share:

Introduction

While most Premiere Pro workstations end up using either an NVIDIA's GeForce or NVIDIA Titan GPU, there are times when using a more expensive workstation GPU is necessary. Workstation cards don't usually give you any higher performance, but the most common reason for using them is due to the fact that they can output 10-bit color. To be clear, if you simply need 10-bit support on a secondary monitor for full resolution preview or color grading, using something like a Blackmagic Decklink card (paired with a GeForce/Titan GPU) is often the recommended route to take. But if you only have a single display, or for whatever reason need 10-bit support on your primary display, using a workstation graphics card is the only way to do so since most consumer cards do not support displaying 10-bit color.

For those that need a workstation card for Premiere Pro, we wanted to determine how various cards from NVIDIA's Quadro and AMD's Radeon Pro lines perform. Information on our testing process can be found in the Test Hardware & Methodology section or you can skip right to the Conclusion. If you are more interested in the performance of NVIDIA's GeForce and Titan line, we also have a similar article available that covers various consumer oriented cards: Premiere Pro CC 2018 GPU Performance: NVIDIA Titan V 12GB.

Test Hardware & Methodology

To see how modern workstation graphics cards compare in Premiere Pro, we opted to combine a number of different GPUs with the Intel Core i9 7940X CPU which is currently the fastest all-around CPU for Premiere Pro. Our full test platform consists of the following hardware:

Skylake-X (X299) Test Platform
Motherboard: Gigabyte X299 AORUS Gaming 7
(rev 1.0)
CPU:

Intel Core i9 7940X 3.1GHz
(4.3/4.4GHz Turbo) 14 Core

RAM: 8x DDR4-2666 16GB
(128GB Total)
Hard Drive: Samsung 960 Pro 1TB M.2 PCI-E x4 NVMe SSD
OS: Windows 10 Pro 64-bit
Software: Premiere Pro CC 2018

The workstation GPUs we will be testing includes a number of cards from the NVIDIA's Quadro and AMD Radeon Pro line as well as the NVIDIA Titan V to act as a comparison point:

The footage used in our testing is shown below with links to where you can download it yourself:

Codec Resolution FPS Camera Clip Name Source
H.264 3840x2160 59.94 fps Panasonic GH5 60p Snow Handheld The Angry Video Guy
Panasonic GH5 Sample Footage
ProRes 422 HQ 3840x2160 24 fps Ursa Mini 4K City Train Station Blackmagic Design
Production Camera 4K Update
ProRes 4444 3840x2160 59.94 fps Canon C200 Untitled00024199 4K Shooters
Canon C200 Raw Footage Workflow
CinemaDNG 4608x2592 24 fps Ursa Mini 4K Interior Office Blackmagic Design
[Direct Download]
ARRIRAW 6560x3100 23.976 fps ALEXA 65 A003C025
(Open Gate spherical)
ARRI
ALEXA Sample Footage
RED 3840x2160
(11:1)
23.976 fps EPIC DRAGON A016_C001_02073O_001 RED
Sample R3D Files
RED 4096x2304
(7:1)
29.97 fps RED ONE MYSTERIUM A004_C186_011278_001 RED
Sample R3D Files
RED 6144x3160
(12:1)
23.976 fps EPIC DRAGON A007_C115_07181B_001 RED
Sample R3D Files
RED 6144x3077
(7:1)
23.976 fps WEAPON 6K S005_L001_0220LI_001 RED
Sample R3D Files
RED 8192x4096
(12:1)
23.976 fps WEAPON 8K S35 S002_C074_02065Z_001 RED
Sample R3D Files
RED 8192x4320
(9:1)
25 fps WEAPON 8K S35 B001_C096_0902AP_001 RED
Sample R3D Files
RED 8192x4320
(7:1)
23.976 fps EPIC-W 8K S35 S002_C074_02065Z_001 RED
Sample R3D Files
DNxHR HQ
8-bit
3940x2160 29.97 fps Transcoded from RED A004_C186_011278_001
DNxHR HQ
8-bit
6144x3160 23.976 fps Transcoded from RED A007_C115_07181B_001
DNxHR HQ
8-bit
8192x4320 25 fps Transcoded from RED B001_C096_0902AP_001

While this is by no means every codec available, we do feel that this covers a much wider range than our previous testing. In the future we may cut down on the number of RED clips and replace then with something like XAVC-S or AVCHD but we really wanted to see how the different compression levels impact performance.

Our testing was done with three different timelines to simulate different types of workloads.

Basic
(Live Playback)

  • 4 Clips in series
  • No effects
  • No transitions

Lumetri Color Only
(Live Playback)

  • 4 Clips in series
  • Lumetri Color on all clips
  • No transitions

Heavy Effects
(Live Playback & Export)

  • 10 Clips stacked
  • Lumetri Color on all clips
  • Cross dissolve between all clips
  • Audio track
  • Includes:
    • Keyframed position, size, & crop
    • Gaussian Blur
    • Multicam sequence
    • 2x2 grid of clips
    • Text overlay

To be clear, these test sequences are not anything fancy and frankly don't end up with a very nice final product since they are simply copies of the same clip over and over. However, if you wish to replicate our testing we have all the project files and export presets available for download. You should be able to replicate any of our tests by downloading the clips from the original source, making 10 copies of each clip, and re-linking the media in the appropriate project file. Export time was simply recorded from when the "Export" button was clicked until the export completed. Live playback FPS (frames per second) was measured based the number of dropped frames relative to the total number of frames in the sequence. For example, if we dropped 100 frames in the 4K ProRes 422 HQ "Basic" timeline (which has 579 frames), that means we rendered a total of 479 frames over 24 seconds for an overall result of ~20 FPS. For ease of comparison, we also have the raw number of dropped frames listed in the Live Playback section.

Live Playback

Live playback performance is a challenge to accurately test since performance depends not only on the codec and resolution of your media, but also the playback resolution (full, half, quarter, etc.) and the effects you have applied. Since this is the first full round of testing with our new benchmark process, we decided to go all out and benchmark all our test media with not only three different timelines, but also at full, half, and quarter playback resolution. For more information on each test media and timeline, check out the Test Hardware & Methodology section.

 

Premiere Pro CC 2018 Live Playback Benchmark Quadro Radeon Pro

Individual Avg. FPS Results

[+] Show Raw Dropped Frame Results

Due to the huge amount of data gathered in our testing (over 800 different data points in total!), we are not going to go through everything point by point. If we tested a codec you typically work with, we highly recommend looking at the "Individual Avg. FPS Results" charts or the raw Dropped Frame results by clicking on the "Show Raw Dropped Frames Results" link. However, for a general overview of how each GPU performed we decided to average the results relative to the Quadro P5000 video card.

Likely, the first thing you may notice is that the Radeon Pro WX 9100 did not do particularly well in this test. The overall average chart is a bit shewed since the largest area it struggled was with RED footage, but even in general the performance was fairly lackluster once effects like cross dissolves and Lumetri Color was applied to the timeline. One thing we do want to point out is that our testing is done with the cache and system RAM completely cleared in order to replicate what the performance would be like if you were working on a fresh timeline. What was interesting was that while the Radeon Pro WX 9100 gave overall poor performance in this "first run" situation, if you played the same clip multiple times the performance was much higher. For example, in the RED 8K 9:1 testing at full res with no effects, we only received about 9 FPS in our test. However, if we played the clips a second or third time, it jumped right up to around 25 FPS. The NVIDIA cards do exhibit similar behavior, but it was never as dramatic as what we saw with the Radeon Pro WX 9100.

Something else that we fist discussed in our Premiere Pro CC 2018 Titan V article (which included various Titan and GeForce cards) is that the lower the playback resolution, the less the GPU model impacts performance. Of course, at a certain point the GPU doesn't matter since you can't get any better than 0 dropped frames, but even when we were dropping a decent number of frames the difference between the Quadro P4000 and the Quadro GP100 at half or quarter resolution was often less than 1-2 FPS. This suggests that at lower playback resolutions, the GPU takes a back seat to the CPU in terms of importance. This is a very significant finding since with 6K and 8K footage you will likely be dropping to half or quarter playback resolution if you add a decent number of effects in order to maintain near real-time playback. If this is something you will be doing regularly, then you are likely better off with a slightly lower end GPU and spending the cost savings on a more powerful CPU if possible.

At full res playback, however, there is definitely a very distinct difference in performance between each GPU. With 4K footage it may not be terribly noticeable unless you use a lot of accelerated effects, but with 6K and especially 8K footage the difference can be pretty dramatic. At those resolutions we would highly recommend the Quadro P5000 as a starting point and upgrading to the P6000 or even the GP100 depending on the kind of performance you need.

Exporting

Exporting is one of the biggest single time sinks for a Premiere Pro user and is often the go-to metric for measuring performance. For this test, we will be examining 4K, 6K, and 8K projects using all the different codecs listed in the Test Hardware & Methodology section. Added up, this works out to just over 40 different individual tests for each GPU resulting in just under 250 total data points.

Premiere Pro CC 2018 Export Benchmark Quadro Radeon Pro

Individual Export Time Results

Due to the amount of data we collected during our testing, we once again decided to compile the results into a single overall average relative to the Quadro P5000 video card. If you are primarily concerned about just one of the codecs or resolutions we tested, we highly recommend checking out the individual results - just click on any of the thumbnails to view the full-sized chart.

With that said, from an overall perspective we saw a small but noticeable performance improvement as we used more and more powerful (and expensive) GPUs. However, keep in mind that this is just an overall average. In some cases - like when using ARRIRAW footage or when exporting 8K projects to 8K H.265 - all the GPUs performed roughly the same. Other times, there is a decent (but not vast) difference in performance between the lower and higher-end cards.

Conclusion

Premiere Pro CC 2018 Workstation GPU Performance

Looking at just Live Playback at full resolution with the "Heavy Effects" timeline and overall Exporting performance, we get a pretty good idea about how the various workstation cards stack up. Really, the only surprise was the poor performance of the AMD Radeon Pro WX 9100. Otherwise, there is a small - but noticeable - increase in performance as you use higher and higher end cards.

If your workflow requires the use of a workstation GPU, the card you should get really depends on your workflow. If you work with 4K footage and don't use many effects, the Quadro P4000 should be more than adequate. If you tend to use Lumetri Color, Cross Dissolve, or other GPU accelerated effects, however, we would highly recommend starting with the Quadro P5000.

Similarly, if you work with 6K or 8K footage then the Quadro P5000 is really the entry card you should aim for. The more effects you utilize, however, the more benefit you will see from upgrading to a Quadro P6000 or even a Quadro GP100. Keep in mind, however, that even though the Quadro GP100 gives terrific performance in Premiere Pro, it is actually about on par with the GeForce GTX 1080 Ti. So while it is a great card, unless you really need the 10-bit display support you can save somewhere around $6,000 by using the GTX 1080 Ti instead.

Tags: Premiere Pro, NVIDIA, Titan V, P4000, P5000, P6000, GP100, WX9100
manish

i have 1 major confusion regarding "geforce" & "Quadros".
that if we have titan xp paired with blackmagic pcie card that can output 10 bit to a monitor but still geforce cards can render 10 bit colors or not???

if geforce cards can output 10 bit color than why should we purchase quadro over cheap geforce cards..??

Posted on 2018-02-17 17:06:43

GeForce (and Titan) cards can not output 10-bit color - that is why you need something like the Blackmagic Declink cards. What those cards do is interface directly with supported software (Premiere, After Effects, DaVinci Resolve, etc.) to display highly accurate and 10-bit color to a dedicated display. This isn't a normal desktop display, however. In most software, it is simply get a full screen preview of what you are doing - just with more accurate color and 10-bit.

You would use a Quadro card if you need 10-bit not just for a dedicated preview/grading monitor, but your normal desktop monitors as well.

Posted on 2018-02-17 20:18:24
Wun Yung

I have a Vega Frontier, and I think it outputs 10bit color also.

Posted on 2018-02-18 01:46:39
manish

so what you thing i should buy a titan card paired with blackmagic card to watch and render 10 bit color?
are there titan cards render out 10 bit color is final output video or we still need quadro cards for render out 10 bit color ?

Posted on 2018-02-18 13:26:47
manish

also if you \r forum can provide some 10 bit monitors options for us that can be intresting .

Posted on 2018-02-17 17:07:57
Tugrul_512bit

Thats too many cuda cores. I wish quadros had nv-link like their tesla cousins.

Posted on 2018-02-18 07:05:51
Streetguru

Did you guys already get in the Radeon SSG for 8K footage testing? If not is that something you'd consider doing? Would just be interesting to see after all the claims of it being capable of doing 8K RAW scrubbing/playback. They're only $6999 on B&H, could possibly colab with Linus or MKBHD for exposure there.

https://www.bhphotovideo.co...

Posted on 2018-02-19 17:37:11

A bit frugal with the 128GB of RAM. Good lord.

Posted on 2018-02-19 23:18:50
Jacob Pawloski

I'm disappointed their wasn't a more consumer based graphics card like the 1080 Ti so that we can compare consumer grade to workstation grade. Its nice to know what an extra few grand for a workstation GPU will buy you...

Posted on 2018-02-24 21:42:56

We linked to our GeForce article several times if you are interested in cards like the 1080 ti: https://www.pugetsystems.co.... We don't directly compare GeForce and Quadro cards, but both articles include the Titan V that can be used as a common point of reference.

Posted on 2018-02-24 21:45:23

Have you done any testing of Premiere with *both* a Quadro AND a GeForce card (thinking 1080ti primary + P2000 secondary) and if that would benefit performance? Or just a waste? Very curious.

Posted on 2018-04-29 18:59:08

Premiere doesn't work very well with multiple GPUs so adding a P2000 secondary likely won't make of a difference for performance. In fact, mixing GeForce and Quadro can sometimes cause problems so unless you really need to we generally recommend against doing so.

Posted on 2018-04-30 17:23:04
Caio Rodrigues Pirinelli

hey P.S., GREAT work on those articles, wish you guys had a base here in brazil.

i have a z800, gtx 1060 with 2 monitor and i want to add a 3rd monitor for output full screen (don't need 10bit for now), so my question is, will my performance be affected by the numbers of monitos or it's best me to add a BM intencity 4k or alike for the 3rd monitor?

tks

Posted on 2018-05-03 03:25:48
Radek .

I think this article is sponsored by nvidia.
I have Ryzen 7 1700 OC 3.8Ghz + RX 580 8GB and clip "60p Snow Handheld" from Gh5 4k 60fps maintains 60 frames during playback with LUMETRI. There is no frame drop do 54FPS. and rx 580 card is 2x time weaker than wx 9100. And in puget system test is frame drop to 54FPS. it's weird for me

Posted on 2018-07-14 10:50:05