Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1287
Article Thumbnail

PugetBench for After Effects

Written on December 10, 2019 by Matt Bach
Share:

At Puget Systems, one of our primary goals is to provide fast and reliable workstations that are perfectly tailored to each customer's unique workflow. As a part of this, we publish an ongoing series of hardware articles for After Effects - as well as numerous other software packages - in order to discover what hardware configurations are optimal for these specific applications.

However, while this ensures that we are selling the right hardware, it does not give our customers a great idea of how much faster a new workstation might be compared to their existing system. In addition, with the amount of hardware that is currently available, it is impossible for us to test everything on our own, so we want to provide tools for other hardware reviewers to perform real world benchmarking in professional applications like After Effects.

To address both of these issues, we are making our After Effects benchmark available for download so that anyone can can perform the same testing we do in our hardware articles. Even better, this benchmark is compatible with both Windows and Mac-based systems which allows for cross-platform performance comparisons.

Download PugetBench for After Effects

PugetBench for After Effects free edition

Free Edition

Download V0.93 BETA

Suitable for personal use.
Results upload to the PugetBench online database.

PugetBench for After Effects commercialedition

Commercial Edition

More Information

For commercial use.
Includes optional uploading of results, email support, automation, and logging.

Be sure to check out the full list of benchmarks we have available!

PugetBench for After Effects

System Requirements

  • Adobe After Effects 2020 or Adobe After Effects CC 2019
  • 16GB of memory (at least 12GB of which is available for Ae to use)
  • Compatible with both Windows 10 and MacOS
  • After Effects/OS language must be set to English

Note that this benchmark is still in BETA. Plug-ins and customized preferences in particular may prevent the benchmark from running properly. If you run into any issues, we recommend resetting your preferences either by following Adobe's guide or by using our Adobe CC Preference Backup/Restore/Clear utility (Windows only).

How to run the benchmark

  1. Download and unzip the benchmark from the link above.
  2. Download a ZXP (Adobe plugin) installer and use it to install the plugin. We recommend using aescripts ZXP installer to install the .zxp file
  3. Launch After Effects
  4. Configure the necessary After Effects settings:
    • ​​​​​​​Enable "Allow Scripts to Write Files and Access Network" in "Edit -> Preferences -> Scripts & Expressions"
  5. Start the plugin from "Window->Extensions->PugetBench for After Effects"
  6. Configure any settings you want to change (especially if you have a commercial license)
  7. Run the benchmark!

When the benchmark is complete, it will give you an "Overall Score" and upload the results to to our online database (unless told not to with the commercial version). From the uploaded results, you can view the sub-scores and individual test results.

Troubleshooting

If you have any issues with the benchmark not properly running, firs verify that your system meets the System Requirements and that the version of After Effects you are using is supports. If there are no issues with your system or software, there are a number of troubleshooting steps you can try:

  • Make sure you are following each step listed in the How to run the benchmark section.
  • Make sure you have at least 12GB allocated for After Effects to use (set in the memory section of preferences).
  • Do not attempt to use the system while the benchmark is running. Defocusing the application window can break the benchmark in many situations.
  • If you are on MacOS and have renamed your drive, the benchmark will not be able to get your system specs. The only work-around at the moment is to rename your drive back to "Macintosh HD", or you can continue to run the benchmark without it being able to pull specs.
  • Clear the application preferences and reset them to default using Adobe's guide or our Adobe CC Preference Backup/Restore/Clear utility (Windows only).
  • Ensure that your OS and After Effects is set to English and that the path to the benchmark script does not have any special or non-English characters.
  • The benchmark does not work over an unmapped network drive (I.E. //192.168.1.2/Benchmark/). Mapped drives should work, but we highly recommend having the benchmark on a local drive as storage performance can affect the results.

Lower than expected performance can be caused by a number of factors, but the most common are:

  • Outdated BIOS/drivers/software
  • Running background processes (Antivirus, Windows Updates, etc.)
  • Windows power profile (High Performance is recommended for the best results)
  • Adobe application settings are not correct (GPU acceleration not enabled, etc.)

If you continue to have issues, we recommend posting in the comment section below - ideally with a screenshot of the problem and a copy of any log files that are generated by the benchmark.

How does the scoring work?

The scoring system used in our benchmark is based on the performance relative to a reference system with the following specifications:

  • Intel Core i9 9900K 8 Core
  • NVIDIA GeForce RTX 2080 8GB
  • 64GB of RAM
  • Samsung 960 Pro 1TB
  • Windows 10 (1809)
  • Adobe After Effects CC 2019 (ver. 16.1.1)

  • Overall Score: 1000
  • Render Score: 100
  • Preview Score: 100
  • Tracking Score: 100

The Overall Score is simply the average of the three types of tests and multiplied by 10 since bigger score = more important.

Update Log

Version 0.93 BETA

  • Fixed issue with After Effects 14.1.4 that prevented the RAM Preview tests from completing due to only 99 frames out of 100 being rendered.

Version 0.92 BETA

  • Moved to using node-wmi to gather system specs on Windows systems. This should be significantly faster and more reliable than the previous command line method
  • Added GPU driver and motherboard BIOS to the system specs for Windows systems
  • Misc bug fixes

Version 0.91 BETA

  • Improved system spec gathering reliability
  • System specs on MacOS now gather properly even if the system drive's name has been changed from the default "Macintosh HD"

Version 0.9 BETA (Major Update)

  • Results are now uploaded to our online database. This is required for the free version, but opt-in if you have a commercial license.
  • Removed the result screen at the end of the benchmark run now that the full results can be viewed on our benchmark listing page.
  • Moved from a script to a full plug-in. This was required for the result uploading, but also allows us much more flexibility for configurable settings and increases consistency between our various Adobe benchmarks.
  • License validation moved from the CLI utility to the plugin itself.
  • Added the ability to do a "Custom" test which does the same tests as normal, only on whatever compositions are on the root-level of the currently opened project.
  • Added ability to set the location for any disk-based tests (rendering).
  • Added tooltips for the various settings that can now be configured.
  • Status logs and configuration settings moved to "~Documents⧵PugetBench⧵After Effects⧵TIMESTAMP" since we cannot log directly to the plugin folder.
  • Removed "Progress" text layer since this information is now reported in the plugin UI.
  • Disk Cache is now cleared by directly deleting the cache files. This removed the need to click the "Are you sure you want to clear the cache?" window that used to come up
  • General bug fixes and stability improvements.

Version 0.82 BETA

  • The Adobe After Effects team helped us figure out most of the improvements in this update. Thank you to the dev team!
  • Improved method for starting the RAM Preview tests. This fixes numerous issues with the previous method
  • No longer need to manually set spacebar playback settings for "loop once"
  • We can now detect when not all frames were played during RAM Preview (due to user intervention or lack of RAM) and can throw an error
  • Now able to start the render queue without it looking like AE froze. Due to this, we no longer need to use AERender for the render tests and can use the built-in render queue
  • Misc improvements for how we loop through the different tests
  • Due to the change to using the render queue, the scores/results in this version may be slightly different than previous versions

Version 0.8 BETA

  • Added support for After Effects 2020
  • Renamed benchmark to "PugetBench for After Effects"
  • Added motherboard and RAM speed (Windows only) to the results page
  • Removed .csv log file support in the free edition (log files will be a feature in the commercial use version)
  • More code to ensure consistency between benchmark runs
  • Miscellaneous bug fixes and improvements

Version 0.52 BETA

  • Overall Score rounding error fix. Previous versions were rounding to the nearest 10. I.E. "950" instead of "948"

Version 0.51 BETA

  • Added Windows buildnumber to the Benchmark Results page
  • Making sure to set the renderer to "Classic 3D" on most comps, or "Cinema 4D" on the C4D comps (just incase it gets changed somehow)
  • More robust reporting for render errors such as not enough frames rendering or the render failing completely.
  • Clearing out export folder at the start of each run just so that aerender doesn't have to overwrite files.
  • Scoring fix for when a render completes in exactly a minute (the log reports different if it is less than a minute, over a minute, or exactly on a minute)

Version 0.5 BETA (Major Update)

  • Switching from "year.mon.day" version format to a more typical versioning scheme so that we can more easily show when things like the scoring has changed enough to not be interchangeable with previous versions.
  • Dropping color depth from 32 bpc to 16 bpc. At 32 bpc, you needed more than 20GB of RAM in order to RAM preview the UHD compositions and anything less simply stopped the test early resulting in artificially high scores for that test. Dropping to 16 bpc will allow people with 16GB of RAM to successfully run the benchmark.
  • Added a RAM checker that throws an error if you have less than 12GB of RAM available for AE to use. Anything less results in the same issue mentioned in the previous bullet point with artificially high scores.
    • This checker is not very robust at the moment since it has to read some configuration files from /Adobe/dynamiclinkmanager/*version*. I can't find a way to directly link the version of AE with the version of the DLM, so for now it is hard coded and allows you to keep running the benchmark if it breaks.
  • Added a "Progress" text layer on each test comp that updates automatically so you can visually see how far along the benchmark is.
  • Export (Final Render) test now uses aerender (background renderer) rather than the render queue built in to After Effects. This is nearly the same performance as using the render queue, but much easier to automate.
    • Added a "status" comp that shows you how far along it is since there is little to no visual feedback that After Effects isn't just broken
    • This fixes most of the issues we've seen with this test failing on some systems
    • This removes the need to do any security setting changes in MacOS (yay!)
    • This also opens up the possibility of adding a render node test in the future that spawns multiple instances of aerender.
  • Switching from Overall/Standard/C4D/Tracking scores to Overall/Render/Preview/Tracking to align better with what people are interested in. If someone only cares about C4D performance, they can look at the raw results for those compositions.
  • There is now a "Benchmark Results" comp that comes up at the end of the benchmark that displays a bunch of useful information including:
    • Benchmark version
    • Overall & section scores
    • FPS results for each individual test
    • System information like CPU, RAM, OS, GPU, GPU acceleration mode, and After Effects version
  • In addition to the log file with the results, the benchmark now also makes a PNG of the "Benchmark Results" comp
  • Adjusting benchmark tests:
    • Adding three new compositions:
      • "Expressions" - almost exclusively uses expressions
      • "Motion Blur" - Simply comp with motion blur being the limiting factor
      • "GPU Stress" - our attempt to show the maximum difference you would see with a higher-end GPU
    • Removing old compositions that were either redundant or are tested better with the new comps:
      • "Glitch Project"
      • "Turbulant Noise Keyframe"
      • "Rotation Keyframe"
      • "C4D Text Draft Project"
      • "C4D Track Draft Project"
    • Replaced video media in the "Phone Composite Project", "Tracking", and "C4D Track Final Project" with lower bitrate H.264 media. It doesn't significantly impact the performance, but results in a much smaller benchmark download size.
  • Miscellaneous bug fixes and improvements

Version 19.5.16 BETA

  • Fixed issue where the benchmark could not verify the loop playback setting on minor revisions of AE.

Version 19.4.4 BETA

  • Reordered the tests to Render->Half Res Preview->Full Res Preview in order to try to resolve issues with some systems improperly using the disc cache files for the render test which resulted in too high of scores.
  • Added additional checks during the Render tests so that if the benchmark is not able to start the render automatically, it will prompt the user to click the "Render" button manually.

Version 18.12.18 BETA

  • Minor bug fix

Version 18.12.17 BETA

  • Added more robust detection for when the benchmark starts/stops RAM Preview and Render testing. This is to fix a bug on some systems where if the test doesn't start within a few seconds, the benchmark thinks it already finished.

Version 18.12.10 BETA

  • Added Ae version check to make sure Ae 2019 (ver 16.x) is being used.
  • Initial benchmark errors (Disk Cache not disabled, Ae not using English, Ae version mismatch, etc) now allow for you to continue the benchmark anyways - although the benchmark will probably break so that isn't recommended.

Version 18.12.0 BETA

  • First release.

Looking for an After Effects Workstation?

Puget Systems offers a range of poweful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!
Tags: After Effects, Benchmark, Public, Download
ComputahNerd

Hi! thanks a lot for this! But, I tried to set preview to loop once, but the script still wont start and says I need to set preview to "loop once". Despite having restarted AE and even rebooted the machine and previews actually only playing once when testing. Any ideas?

Posted on 2018-12-06 06:47:13

Huh, interesting. Are you on Mac or PC? And you sure you are setting it for spacebar and not any of the other preview hot keys? That setting is a little bit weird in the setting files, so I'm worried there might be something weird with it. I'll take a look at it later to see if there is something I missed and worst case I'll swap it over to a warning rather than a hard error that stops the benchmark from being run.

Posted on 2018-12-06 06:55:57
ComputahNerd

Hey, yeah I'm on a PC. I have tried to restart AE many times and also tried to change all the other playback settings (range and play from) - It refuse to start. See screenshot: http://prntscr.com/lrcx82

Posted on 2018-12-06 06:59:02

Yea, looks like a bug then. I can't check it right now (and unfortunately I might not be able to tomorrow either), but when I can I'll check it and release an update. My guess is just that the script isn't finding that one setting since I actually have to do a file text search to get it and there is probably some weirdness going on.

Thanks for pointing out the bug!

Posted on 2018-12-06 07:08:06
ComputahNerd

Thank you - looking forward to test :) I was just curious as to how my threadripper 1950x performed - which I expect is quite low - as I noticed the slower responsiveness of the entire program after getting this CPU. Though based on your charts, even the 2950x scores pretty bad compared to Intel CPUs - proving that AMD threadripper series is a pretty bad investment for any AE users.

Posted on 2018-12-06 07:26:15

I couldn't find anything wrong, the only other thing I can think of is that if you aren't use Ae 2019 (ver 16.x) it wouldn't be able to detect that you changed that setting since the setting file wouldn't exist.

I just added a Ae version check to the script, but also made it so those errors are more of a warning. You can have it run the benchmark anyway, but it just means that it might break at some point or get stuck. So re-download it and give it a try.

Posted on 2018-12-10 21:40:28
ComputahNerd

thanks will test it. I use the very latest version of CC 2019 AE.

Posted on 2018-12-10 21:48:36
ComputahNerd

So even though it gave me the same warning now and also another which was called "unknown" it worked fine. It wen't through all the benchmarkings. It said also that it couldnt perform the changes, though I had set it up myself. (perhaps I should have started AE in administrator mode?)

Anyway, results for my 1950x was rather bad, as anticipated, but it was good to see that the experience I have had with the CPU made sense according to the test.
Overall:634 - standard:66 - C4D: 58 - Tracking: 63

Almost as bad as Mac Pro, So awful results both in standard and C4D scenes, which you guys have stated before. TR4 is by no means a good overall CPU for creators, except only those who need many cores for rendering for a cheap buck.

Posted on 2018-12-11 12:18:18
Hugo Scali

Hi, I wanted to ask permission to use the results of their benchmarks on my Spanish-language YouTube channel since I plan to make a video about the workstation. Since it is not true that always more nucleos is the best choice and the choice of components should be based on the determined task that they are going to execute. Obviously I will put in the description that the data was extracted and analyzed from pugetsystem and I will leave the link to the site.

This is my channel: https://www.youtube.com/cha...

Posted on 2018-12-09 08:15:59

Sure! We would appreciate having some reference to "Puget System's After Effects CC Benchmark" (or something along those lines) on the graphic itself in addition to a link in the description, but absolutely you can use results from our benchmark.

Posted on 2018-12-10 18:13:50
Nazar Malyy

Hi Matt. I have the same issue. Preview playback is set to "Play once" but anyway I getting message to set again to 'Play Once". (Even OS restart didn't help.)

=====================
AE version 16.1.0 (Build 204)
MacOS High Sierra 10.13.6
MacBookPro 2017 3.1GHz 16 GB RAM

Posted on 2019-04-04 13:05:14

16.1 moved the preferences file a bit. I'm working on an update right now that should fix it and keep it from happening in the future. Just have to run the benchmark on a couple of systems here to make sure I didn't break anything else, then I will upload it. Should be up before 3pm today (assuming it is working right).

Posted on 2019-04-04 18:45:00
Nazar Malyy

Thanks Matt! All working fine and I already know that I have "trash" :D ...probably I'll buy a new PC.
My lovely Apple became expensive "garbage" in nowadays.... it's sad :(

Posted on 2019-04-05 09:24:11
mateo5

Thanks for this!! it's really helpful :) Something that would make this more awesome IMHO is to include in the downloads the benchLog csv of scoring machine, in order to be able to compare in detail each one of the tests with the base config that sets the benchmark score. I for one, would love to see how many fps the base machine got in the polygon test RAM preview in order to compare to mine. In the polygon test I got: RAM Preview (Half Res.),Polygon Project,11.1,FPS.

Posted on 2018-12-08 12:02:58

The main reason we didn't get too much into the reference scores is because that score is not going to change over time. So as there are driver updates, Windows updates, Ae updates, etc. the actual performance of a system with a i9 9900K is going to change more and more from the reference score. What I would do is to compare your results to those in out hardware articles. https://www.pugetsystems.co... right now is the only article using this exact benchmark, but it includes results for CPUs like the i9 9900K.

Posted on 2018-12-10 18:11:11
Craig Marshall

Thanks for this tool! But I'm confused by the results on our systems. Our benchmarks are way above the "top" score. Most of our systems are showing an overall score of 6100 or more. Other metrics are equally out of range (Standard score: 798, C4D: 804, etc). Any idea what I'm missing here?

Posted on 2018-12-12 22:04:47

Huh, that is interesting. Could you grab the log file (labeled BenchLog_XXXXXX.csv in the benchmark script folder) and toss it to me in an email at labs@pugetsystems.com ? Those scores look like they are just multiplied by 10, but I can't think of how that could be possible. The other thing it could be is that either the disk cache isn't being cleared (if you have run the benchmark multiple times) or the viewport resolution isn't getting properly set to half/full. Seeing the log file will let me determine if it is a score calculation issue or something wrong with the actual run.

The other thing you could try is to reset your Ae Preferences. You can just make a copy of the preference folder, then go back to it after you run the benchmark if you don't want to lose any of your custom preferences: https://helpx.adobe.com/aft...

Edit: Actually, include the StatusLog_____.txt file as well if you can.

Posted on 2018-12-12 22:55:29

Hey Craig, I just sent you an email, but for anyone else who had this issue it should be fixed in the 18.12.17 version I just uploaded. I think the issue was that on some systems, it takes longer to start the render than I expected. That means that sometimes the benchmark would think that the render was done before it actually even started. I added some extra code to help make sure that doesn't happen so hopefully it should work for you now!

Posted on 2018-12-17 23:55:35
Alex S.

Hello, I have the same problem here and downloaded the script on friday, so it should be up to date. My overall score is 19900... So what can I do to be sure of the result ? And moreover, how can I interpret these results (just : ok it's a good machine, or is it more precise...?) ?
I also had to click "render" for every final render, is that normal ?
Thanks

Posted on 2019-03-11 13:29:05

No, that isn't normal. If you are on a Mac, did you follow the directions to add the two applescripts to Accessibility so they could run correctly? https://www.youtube.com/wat...

Most of the time, abnormally high scores are caused by either that or disk cache not being disabled. So try making sure that is disabled before running the benchmark, and make sure you let the After Effects window stay in the foreground. If you are doing other things on the system at the same time, it may not be able to run properly.

Posted on 2019-03-11 17:08:26
Alex S.

I didn’t add the scripts to accessibility but I launched them before to be sure they passed security issues.
And disk caches were empty. Actually, this is a new machine, that’s why I make some tests.
By the way, as for the results : the higher the better or the lower the better ?

Posted on 2019-03-11 17:22:56

You need to check them in Accessibility. Running them once adds them to that list, but doesn't actually enable them to allow other applications to run them.

Higher results are better, but it should be almost impossible to get an overall score above 1000. Something like 19900 means that it wasn't actually doing some of the tests, just immediately failing and counting something that should have taken 60 seconds (or whatever) as having been completed in 1 second.

Posted on 2019-03-11 17:26:23
Alex S.

I add the scripts to Accessibility, but it still blocks on render. The log says : "Add render to queue
12:59:51 - Starting Final Render
12:59:51 - Send Return
12:59:51 - Waiting for Final Render"
And then I have to click manually on render.
The thing is that I'm on 10.13, and don't have the "automation" tab in general settings

Posted on 2019-03-12 13:03:40
Alex S.

Maybe I should specify that clicking "return" in After Effects doesn't do anything. The action on the button can only be manual at this point. No shortcut available

Posted on 2019-03-12 13:11:58
Alex S.

ok, I redid it and tried to click on render as fast as possible to simulate an automatic click on the render button, and I got a score of 39000... Something went wrong again I guess :/.

Posted on 2019-03-19 16:06:38
WB

Hello, same problem here on a MacPro 5.1 running latest high sierra with a SSD.
overall score of 6831,75 (standard score 566,2 / C4D : 1563.8 / tracking : 36,5)
I has to run the scripts manually since they didn't start, even using Accessibilty like in your video.

I'll try agin.

Posted on 2019-04-04 14:50:50

I'm working on an update right now that should fix that, or at least make it smart enough to know when the renders aren't starting and ask you do do some manual input. It will still technically result in a tiny bit lower score than if it worked automatically (since it takes a real person a split second to click on buttons), but I don't think it will skew the results more than a fraction.

I just have to run the benchmark on a couple of systems here to make sure I didn't break anything else, then I will upload it. Should be up before 3pm today (assuming it is working right).

Posted on 2019-04-04 18:46:44
CPH

Hi,

First, thanks for the help so far … am trying to diagnose why my new Radeon RX580 makes no difference in AE rendering time (in 'Mercury GPU Acceleration - Metal' or '...OpenGL', versus '...Software Only'). Was about to run the Puget Sound Benchmark but the last step has 'PugetBenchmark.jsxbin' grayed-out (see attachment):
https://uploads.disquscdn.c...

This page was very helpful in trying to understand these issues –– https://www.pugetsystems.co... –– *I'm only interested in Standard Render speeds btw*, for what I do.

I can return the card if nec. But am wary of NVIDIA in Macs, due to potential hassles and learning curves, and also cost is a factor.

System:
Razer Core X - Thunderbolt 3 eGPU
Radeon RX 580 8GB card
MacBook Pro (13-inch, 2017, Thunderbolt 3 ports, 2.3 GHz Intel Core i5, 16GB RAM, Mojave 10.14.2)
Primary software: After Effects/Premiere/CC2019)
Output drive: Samsung Portable SSD T5

~

Have re-started AE btw. Also, the Activity Monitor/GPU History window shows the rendering is being done by the Radeon card.

Posted on 2018-12-15 15:51:36
CPH

Hi,

First, thanks for the help so far … am trying to diagnose why my new Radeon RX580 makes no difference in AE rendering time (in 'Mercury GPU Acceleration - Metal' or '...OpenGL', versus '...Software Only'). Was about to run the Puget Sound Benchmark but the last step has 'PugetBenchmark.jsxbin' grayed-out (see attachment).

This page was very helpful in trying to understand these issues –– https://www.pugetsystems.co... –– *I'm only interested in Standard Render speeds btw*, for what I do.

I can return the card if nec. But am wary of NVIDIA in Macs, due to potential hassles and learning curves, and also cost is a factor.

System:
Razer Core X - Thunderbolt 3 eGPU
Radeon RX 580 8GB card
MacBook Pro (13-inch, 2017, Thunderbolt 3 ports, 2.3 GHz Intel Core i5, 16GB RAM, Mojave 10.14.2)
Primary software: After Effects/Premiere/CC2019)
Output drive: Samsung Portable SSD T5

~

Have re-started AE btw. Also, the Activity Monitor/GPU History window shows the rendering is being done by the Radeon card.

~

https://uploads.disquscdn.c...

Posted on 2018-12-16 00:43:04

Are you sure you are trying to run a script and not trying to open a project in Ae? That is the only reason I can think of why that would be grayed out like that. "File > Scripts > Run Script File". If that isn't it, I don't know what else it could be. I would try resetting your preferences and if necessary, uninstall and reinstall After Effects. After that, you should contact Adobe support since not being able to run a script indicates that there is something wrong with either your system or After Effects.

Posted on 2018-12-17 17:40:16
CPH

Thanks very much for your reply Matt … I've found the following solution:

––after selecting "File > Scripts > Run Script File", the resulting window has a pulldown menu at the bottom, with "Enable: Adobe JavaScript file" pre-selected
––but when I select "Enable: Adobe JavaScript binary file" instead, "PugetBenchmark.jsxbin" becomes newly available (is no longer grayed-out). I can then run the test successfully.

Unfortunately, I'm traveling until Boxing Day, and so I'm away from my EGPU+. I'll run the rest of my tests after I'm back, and will be in touch again after that.

Have a good holiday.

Posted on 2018-12-19 02:16:54
CPH

https://uploads.disquscdn.c...

Posted on 2018-12-19 02:18:19
Savage

Hi trying to set up this benchmark and I get a popup immediately on running script that says Puget Systems After Effects Benchmark Complete Overall Score infinity Standard Score infinity C4D Score infinity Tracking Score 0. Clearly it's not running what should I check for?

Posted on 2018-12-18 13:39:35

Oops, that was my mistake when I did the last update. There is a new version that should fix that issue. Sorry about that!

Posted on 2018-12-18 17:18:35
Savage

Thanks Matt that did it!

Posted on 2018-12-18 23:53:05
Jeff

tried the test on the latest version of AE and it just kept looping through the test even though loop playback was off in the Ram Preview UI. Restarted several times with the same result. It did however write out a log and it looks like one time through the test took about 12 minutes. Wondering how that stacks up.

Posted on 2018-12-22 00:34:07

You sure loop is disabled for spacebar preview specifically? If it is, does it continue to loop anyways of you open a comp and manually hit spacebar? Only other thing I can think of is that somehow your "default" playback method is set to something besides spacebar, but I don't know how that could be done.

I would try resetting your preferences to default if you still can't get it to work right.

Posted on 2018-12-22 00:37:04
Enrico Geromin

This is very nice! Thank you for sharing yours develop & research softwares with us!
Will be very cool if in future there will be a Premiere version of this Benchmark. Thanks again for your work ppl!
My old 4930K @4.4ghz scored 539pts!

Posted on 2018-12-24 00:24:37

We have an internal benchmark for Premiere Pro, but it is likely going to be a while before we can make it public. Premiere doesn't have nearly the scripting capability that Ae or Ps have which means we have to use external utilities pretty much exclusively. That is completely fine for us, but it makes it finicky if you don't have the same screen resolution, DPI, workspace, etc. that it is designed for.

Working on it though!

Posted on 2019-01-02 19:54:14
Karl-Jason Mawdsley

Hey - just a score back of infinity for everything except tracking? Running on a Mac

Posted on 2019-01-02 13:16:52

Are you using the latest benchmark version (18.12.18)? There was a bug in earlier versions that caused that problem, but it should have been fixed in the latest.

Posted on 2019-01-02 19:52:20
Karl-Jason Mawdsley

Yeah I just downloaded it from the link above today. I had a few issues with it at first because I'm running Mojave so had to give MagnifytoFit and SendEnter access to control the computer in System Preferences. I'll try and run it again today and get back to you

Posted on 2019-01-03 08:59:37

I just downloaded the benchmark to the iMac Pro we have here and ran it to check it out. I didn't get any errors or any other issues, so my guess is that your problem is related to the Mojave security stuff. You might check out our video to make sure you have everything set properly there: https://youtu.be/mU3j0AM_fPM

After that, it may be a preference setting you have that is messing something up. You could try resetting the preferences to default to see if that fixes it: https://helpx.adobe.com/aft...

https://uploads.disquscdn.c...

Posted on 2019-01-03 18:55:00
Karl-Jason Mawdsley

When I ran it again it seemed to work fine

Posted on 2019-01-04 13:07:20

We typically see a variance of 3-5% between runs, but I don't think seeing a bit more than that across multiple machines would be too unusual. Especially since different systems might have difference applications running in the background, slightly different software configurations, etc., that can all affect benchmark performance.

Posted on 2019-01-07 18:25:52
Steve E Miller

I did this just for fun on my Lenovo Legion laptop that I got about 1.5 years ago. i7-7700HQ 2.8GHz, 16Gb ram. Took more than an hour to run it.
Results aren't too bad compared to the reference system considerinng the hardware differences.
Overall: 498.25
Standard: 55.6
C4D: 40.5
Tracking:: 47.6

Posted on 2019-01-07 19:32:06
RogueScholar

Thanks for making this available. My system is very similar to your base system. i9-9900k, 64gb RAM, Crucial 1TB SSD, Windows 10, but with an RTX 2080Ti. My scores were: Overall 896.5, Standard 95.9, C4D 95.1, Tracking 71.7. With our systems being so similar, and having the upgraded GPU, is there any way to narrow down why my scores weren't closer to the base, or even higher?

Posted on 2019-01-10 03:44:01

The GPU isn't going to make much of a difference in Ae - in fact, the 1080 Ti benchmarked a bit higher than the RTX cards in our testing: https://www.pugetsystems.co...

There is a lot that could be lowering your scores, ranging from the version of Windows you are on, driver/BIOS version, what motherboard you are using (some are more or less aggressive on Turbo settings), CPU cooling, and what applications you have running in the background. Things like antivirus can be especially bad depending on what software you are using.

Posted on 2019-01-10 18:02:03
Jc El

funny you said that. I am running a lower spec gpu and cpu and my overalls were better than this system and very close to the reference model. can you explain? I have a i7 8700k; 64gb ram and rtx geforce duel fan1070 super gpu.

Posted on 2020-05-09 08:43:13

I'm not seeing a popup after the tests, but I am getting a log file. Any ideas? Running AE Build 16.0.1

https://uploads.disquscdn.c...

Posted on 2019-01-17 14:44:41
zalmsalade

I have the same problem. No popup but I do get a log file. Also running Mojave and AE 16.0.1.

Posted on 2019-01-18 22:05:51
zalmsalade

The render times are mentioned in the log file and each project is 100 frames long, so you can calculate the framerates manually if you can't wait till this has been fixed.

Posted on 2019-01-20 19:06:52

You mind emailing the BenchLog_XXXXXX.csv and StatusLog_XXXXXX.txt files that are in the benchmark folder to labs@pugetsystems.com ? I haven't been able to replicate the issue on our Macs here, but that might give me enough information to figure out the issue. Thanks!

Posted on 2019-01-21 17:18:00

I have the same issue, run the test but no pop up. I am using After Effects v16.0.1 (Build 48) on a Windows 10 64bit PC.
I do get the log file but it looks like it may not have finished fully, the last entries of the log are:

10:6:1 - Done with RAM Preview Full Res
10:6:2 - Starting Final Render with projnum_render=1
10:6:3 - opening comp Phone Composite Project
10:6:3 - Moving playhead to frame 0
10:6:8 - Show render window
10:6:10 - Deleting render queue
10:6:11 - Add render to queue
10:6:13 - Starting Final Render
10:6:13 - Send Return
10:6:13 - Waiting for Final Render

Does this mean there was no result to the last benchmark, which would probably be the reason I did not get a pop up?

Thanks for your help,
Bradley

Posted on 2019-01-29 10:32:14

Interesting. Good to see it happens regardless of the OS.

One thing I'm wondering - are you leaving the system completely alone while it is running the benchmark, or are you doing other things at the same time? Browsing the web, etc. Basically, does the After Effects window always in focus?

Posted on 2019-01-29 17:17:07
Mark Donohoe

Im having the same problem,
18:0:2 - opening comp Cloud Lightning Project
18:0:2 - Moving playhead to frame 0
18:0:7 - Show render window
18:0:9 - Deleting render queue
18:0:10 - Add render to queue
18:0:12 - Starting Final Render
18:0:12 - Send Enter
18:0:13 - Waiting for Final Render

No final results are generated im on After Effects cc 2019 Windows 64bit PC

Posted on 2019-02-06 09:12:59

Is the render button showing, or had the render started? And again to confirm, the AE window has focus and you are not doing anything else on the system while the benchmark is running?

Posted on 2019-02-06 15:40:24
zalmsalade

Is there any news on this? Would love to test my machines. On both my Macs they stop at the final renders. AE window has focus and I'm not doing anything else. Followed all the instructions including the Mojave security settings.

18:55:56 - Starting Final Render with projnum_render=1
18:55:57 - opening comp Phone Composite Project
18:55:57 - Moving playhead to frame 0
18:56:2 - Show render window
18:56:4 - Deleting render queue
18:56:5 - Add render to queue
18:56:7 - Starting Final Render
18:56:7 - Send Return
18:56:7 - Waiting for Final Render

Posted on 2019-04-04 20:58:27
Joseph Redding

Since I ran the benchmark my after effects still goes fullscreen when I press play. How do I stop this? I've re-installed after effects which didn't solve.

Posted on 2019-01-30 03:44:22

That is one of the settings that is auto-set in order for the benchmark to work properly. It should have been reverted back to your original settings, but sometimes it doesn't for some unknown reason. You can change it in the Preview settings for Spacebar (or whatever key you use). Note that there are a bunch of other settings that are changed that you might want to doublecheck as well. They are all listed on the main "Benchmark Instructions" composition. Re-enabling disk cache and auto save are the two biggest ones I would confirm.

https://uploads.disquscdn.c...

Posted on 2019-01-30 17:47:19
diegosalinas

I have an overall Score of 706, where can I see the score with your machines? Is it possible to have better results with settings in after effects?

Overall: 706.25

Standars: 42.7
C4D 42.7
Tracking 77.6

i9-9920x 3.5ghz
64gb Ram
evga geforce 2080 ti

Thanks!

Posted on 2019-02-04 01:25:37

Right now, we have our results in our Ae hardware articles: https://www.pugetsystems.co...[]=After%20Effects . We only have one article at the moment using this benchmark, however: https://www.pugetsystems.co... .

Eventually we are planning on having some sort of public benchmark result viewer (maybe even allowing public uploads of results), but we have a long backlog of programming projects we need to get through before we can tackle that.

Posted on 2019-02-05 03:40:01
Rens Wegerif

so.... after the benchmark crashed, I decided to continue to work I was inititially doing. But now for every ram preview I make, the magnifitofit thingy is doing his work. Very annoying. How do I fix this?

Posted on 2019-02-05 13:33:17

The benchmark might still be running in the background (it shouldn't be, but programs crashing can be weird) . I would just do a quick reboot.

Posted on 2019-02-05 16:36:30
rdanielsdesign

Hi there. I get an error when I try to run the script:
Unable to execute script at line 15. Function d.getHours is undefined

Please assist. I'm on a Mac btw. Thanks!

Posted on 2019-02-05 17:09:16

That is weird, pretty sure that is a built-in command . You have any extensions or plug-ins installed? If so, you could try uninstalling them or resetting Ae back to default preferences (back them up first if you have anything set that you want to keep).

Posted on 2019-02-05 17:11:38
rdanielsdesign

I do have extensions installed. I'll take a stab at both suggestions and see if either of them fix the issue. Thanks!

Posted on 2019-02-05 17:15:18

Any solutions to this? I've had the same issue with some aescripts scripts in the past. I don't know if there's some sort of environment variable not set or what.

Posted on 2019-04-04 18:19:08

That is using built-in Javascript functions to get the current date and time, so if you get this as an error there is likely something wrong with your installation of After Effects. I would try doing a completely fresh Ae install. It is also possible you have some extension that is overwriting the "Date()" function somehow, so you could try uninstalling any extensions you have and add them back one by one. If you find the one that is causing it, definitely go yell at them since they probably have a function in their code called "Date()".

Posted on 2019-04-04 18:22:53

Yeah, I know. It doesn't make sense, especially when other scripts that use it are fine but I believe they all use their own sandboxed JS, sort of. Although there are collisions possible if you don't wrap your code, so I don't know. Anyway, I got it to run after a restart. I ran into some odd video card issues, but I've been having them for a little bit on my Hackintosh. It's run strong since the end of 2013, but it's time for something new. Anyway, thanks for the help!

Posted on 2019-04-04 19:36:00
Kathryn Thomson

I've tried running the script twice. Both times I get to the "cloud lightning project.mov" render, and it just stops. In status it says "Failed at 00000". 25.33GB available on the drive, so it isn't running out of space. Any ideas? Thanks.

I'm on OSX Mojave 10.14.2

Model Name: MacBook Pro
Model Identifier: MacBookPro11,4
Processor Name: Intel Core i7
Processor Speed: 2.8 GHz
Number of Processors: 1
Total Number of Cores: 4
L2 Cache (per Core): 256 KB
L3 Cache: 6 MB
Memory: 16 GB

Posted on 2019-02-07 18:40:00

25GB isn't a lot of free space. After Effects might require a certain amount of space in order to start the render. I would try to free up more space (at least 50GB if you can) and see if that solves the issue.

Posted on 2019-02-07 18:44:49
Tom Hall

But my MacPro 2013 kills the benchmark system x 200!
https://uploads.disquscdn.c...

Posted on 2019-02-16 02:39:18

Ha, that's odd. Looks like it didn't actually run the RAM Preview tests. Mind tossing the statuslog_.txt and benchlog_.csv files in an email to mattbach@pugetsystems.com ? I'm guessing it is a setting in your preferences that prevented the benchmark from running properly, but is may be able to figure out a check by reviewing those files.

Posted on 2019-02-16 03:27:27

Thanks for the log files! It looks like the render tests didn't properly run, so the benchmark thinks it was rendering at a couple thousand FPS. I've seen this a few other times, but haven't been able to figure out exactly why it happens. I have a few ideas to add to the next update that might help, however, or at least have it result in an error rather than a bogus score. Thanks again for the help!

Posted on 2019-02-16 03:41:57

Hi! I did a series of tests using this amazing benchmark, I mainly found out the performance gains from RAM frequency and dual channel, but the amount of RAM does not affect performance.
Here is my test result. https://docs.google.com/spr...

In addition, in some cases of 8G RAM, the RAM preview test data is abnormal. This benchmark seems to be 100 frames of comps. I hope to enhance the detection of RAM preview error and additional provide different time lengths comps for different amount of RAM to do stress test. Because there tests have been showing that the amount of RAM will affect the rendering time, I hope to confirm.
Sorry for my bad English, thank you!

Posted on 2019-02-18 22:44:42

In After Effects, more RAM doesn't really improve how fast AE is able to render each individual frame. What it does allow you to do, however, is to store more frames in RAM Preview. For example, if you have a comp that is 1000 frames long, but you only have enough RAM to store 500 frames, by the time you play through the entire comp and start over from the beginning, the frames at the start have already been bumped out of RAM Preview. In that case, you have to completely re-render those frames from scratch rather than just pulling the already rendered frame from RAM. If you have enough RAM to store all the frames, however, Ae doesn't have to keep re-rendering those frames.

This is why there really isn't an answer to "how much RAM do I need for Ae". It really is... as much as possible. It doesn't make Ae render any faster, but it lessens the chances that Ae has to re-render frames that it otherwise could have simply pulled from RAM.

Posted on 2019-02-19 00:52:18

Cool, this also be my conclusion, the amount of RAM will only affect the preview time length.
I made this with reference to Adobe's formula. https://helpx.adobe.com/aft...
https://uploads.disquscdn.c...
Actually it will be less than the theoretical value but it won't be worse, and the preview time length is usually the same. For example, the same test 32G RAM always stops at 1min 50secs (about 3300 frames) on different computers.

Posted on 2019-02-19 02:04:37
Bartek

Something is wrong with AE tracker part:
22:44:30 - Applying effect ADBE 3D Tracker
22:48:38 - Tracker BROKE ADBE 3D Tracker

Posted on 2019-02-24 21:54:30

I haven't seen that error come up before, so it may not be the benchmark but your install of Ae. Can you manually apply a 3D Camera Tracker on any footage, or does it break? If it breaks, you might want to reinstall Ae to see if that fixes it.

Posted on 2019-02-25 17:44:18
Bartek

Applying it to footage works fine. The issue happens all the time on OSX 10.12.6. This breaks tracker scores which results in 50% less score. Running benchmark on OSX 10.13.6 tracker works fine and results are almost the same as on Windows.

Posted on 2019-02-25 18:12:55
Darryn Rogers

Thanks for the great benchmark tool, guys. I am using it to figure out if it is worth upgrading my machine. I have a question regarding the CPU usage of AE's C4D renderer. I ran the test on my older 4 core machine and my friends brand new 8 core machine and my machine out performed his significantly in the C4D render test. Do you have any idea why this would be so? Does AE's use of the C4D render engine not multi core well? Here's the details:

My Machine - i7 6700K 4 core @ 4.0GHz
Overall Score - 655.25
Standard Score - 70.1
C4D Score - 53.2
Tracking Score - 68.7

My Friends Machine - i7 9800X 8 Core @ 3.8 GHz
Overall Score - 619.75
Standard Score - 74.1
C4D Score - 28.4
Tracking Score - 71.3

Posted on 2019-03-03 21:30:16

Your friend's score is definitely low for some reason. In our testing, the 9800X should get around 78 for the C4D score and an overall score of about 800 https://www.pugetsystems.co... . I'm guessing he has some program-wide setting that has been changed that the benchmark can't account for.

I think your scores are pretty accurate. I haven't tested a 6700K with this benchmark myself, but it lines up with what I would expect.

Posted on 2019-03-04 17:38:39
Darryn Rogers

Thanks Matt. We'll see if we can track down what's holding his machine back.

Posted on 2019-03-04 21:52:47
Alex S.

Hello, I was wondering : if we click every time on the render button by hand (instead of automatically with your scripts), does it alter the final results ? Thanks

Posted on 2019-03-13 22:03:41

Yes, if you have to click the render button by hand, it will skew the results a bit. I've got an update in the works that will add a bit more robustness to that portion of the benchmark though. It won't fix the issue (since the root cause isn't something I can control), but it will at least throw a warning that it couldn't click automatically and ask you to hit the button. Will have a bit more logic built in as well so hitting it manually won't affect the results anymore.

Posted on 2019-03-14 16:21:20

Dear Matt and team at Puget Systems, this assessment tool is very helpful. The test has a small hickup on my end. I’m on MacOS 10.13.6 and in the midst of the test I had to manually hit the Render Button for the Glitch Composition to make it continue. After that everything worked as expected. I used the latest version of the benchmark test from your website.

Posted on 2019-03-14 11:53:32

If it just didn't work once, most likely Ae wasn't the active window at that time. If you use the system while the benchmark is running, it can prevent it from starting some tests.

Posted on 2019-03-14 16:27:27

Thanks Matt, that was probably the case. I remember doing something in Finder.

Posted on 2019-03-14 17:33:25
Ali Khan

Hey Matt, i just saw the School Of Motion youtube video about After Effects machine, I live in Pakistan and unfortunately i can't purchase a custom PC from you guys as you don't ship internationally, Well currently i'm using Intel i7 8700k 4.0 ghz (unlocked) with 32 gb of 2000mhz crosshire rams, gtx 1060 6gb graphics card and 2 ssds and 2 standard hard drives.

I did the After effects bench mark test from the given project file on your website, During the test i got 'Hardware Display Acceleration' error but still i was able to complete the benchmarks and my total score is 660.5 which i think it's average.

I'm upgrading to 64 gb of rams and getting Rtx 2080 8gb also upgrading to 1 TB of SSD to move all of my assets to SSD from my standard drive.
For the i9 processor it's a bit out of budget for me.

My Question is that, will this new Add-on's will improve the overall performance in After Effects?

Posted on 2019-03-14 23:11:16

After Effects is really CPU limited most of the time, so the RTX 2080 may add a bit of performance, but probably not very much to be honest.

Moving your assets to an SSD definitely will help though. It likely won't show up in the benchmark, however, since it will mostly improve performance for things like importing media and opening projects. How much it helps will probably depend on the type of files you work with. Generally, a lot of smaller files will benefit more from an SSD than a few large files - although that depends a lot on what the random and sequential read/write performance of your current drive.

Posted on 2019-03-14 23:20:59
Ali Khan

Thanks for the quick reply, I have one more question.
I'm a vfx artist so i mostly use 3d Tracking, mocha pro AE, Elements 3D + C4D, Trapcode Particular, Keylight plugin.. tools for creating VFX work mostly.

Do you think the new RTX 2080 8gb (OC) will do help me with these plugins as they are GPU intense, i'm migrating from GTX 1060 6gb

Posted on 2019-03-14 23:28:04

Ah, yea, once you get into plugins things change completely. Right now, we don't do much performance testing with different plugins since that is just a giant rabbit hole that we could get sucked into. So I can't really tell you exactly how much a RTX 2080 would help, but like you said, some of those are GPU intensive so I would imagine it would help quite a bit.

Posted on 2019-03-14 23:36:21
Ali Khan

Thanks again for replying Matt, then i guess i should go for this upgrade then later on upgrade my CPU also, You guys should do a benchmark test for artists who use Plugins mostly for VFX work loads, it will really help us, one last thing..Can you tell me my Intel i7 8700k 4.0 ghz is a overall good processor, where i can see it's benchmarks with 64 gb rams

Posted on 2019-03-14 23:40:38

The 8700K is a really good CPU for After Effects. When the 9900K launched, we found that it was only about 15% faster in AE: https://www.pugetsystems.co...

Posted on 2019-03-15 00:02:41
Ali Khan

Thanks alot!

Posted on 2019-03-15 00:08:25
Ali Khan

Hey Matt again, I just ran the test again with old configrconfig, I'm getting this error why it's that? I have re installed after effects and also updated my graphics card drivers

Error: after effects out of memory (right after cinema 4d tracking test)

Error 2: display acceleration disabled https://uploads.disquscdn.c...

Posted on 2019-03-16 16:11:18

That could be caused by a lot of different things. Could be OS related, or you could have a bad GPU/RAM/motherboard. Figuring it out exactly is likely going to require in-depth troubleshooting to be honest. On the other hand, if the system is working fine for your everyday work, it may not be a big deal. A bit concerning, however.

Posted on 2019-03-18 16:14:23
Ali Khan

I use my system for daily with heavy projects in after effects, i don't get this error mostly, i'm guessing it's a GPU failure for sure, because Display Acceleration is related to GPU acceleration in after effects for sure, i have everything brand new with under warranty in my system expect for my graphic card which is 2-3 years old.

Posted on 2019-03-18 16:21:14
Justin Brown

All was going well until the final render. It failed! All that time waisted. I had several pop ups throughout asking for me to allow or open something. I always chose to open it. Any reason why it failed?

Posted on 2019-03-18 15:51:07

Did you follow all the setup directions? It sounds like you are on Mac and missed the Yellow message in the "How to Run the Benchmark" section: https://www.pugetsystems.co...

"If you are on a Mac using Mojave, you need to manually run the "MagnifytoFit.app" and "SendReturn.app" files prior to running the benchmark. Due to new security features in Mojave, these files will give an error and break the benchmark if they do not have permission to send commands to System Events. If you run them manually and click OK, you can enable permission in "Privacy and Security -> Accessibility". We have a video available showing this process that you can follow."

Posted on 2019-03-18 16:18:56
Justin Brown

Thanks Matt. I missed that! We are on Sierra due to our current servers, so I was able to set up everything with accessibility but we don't have the automation module in sierra. Hopefully it will work and I will try again. Thanks for the fast response.

Posted on 2019-03-18 17:05:03
Justin Brown

Looks like it still fails during the final rendering portion. I'm guessing Sierra OS is not supported for this?

Posted on 2019-03-18 20:37:37

Sierra should work, but to be honest we didn't do a huge amount of testing with that version. All the security stuff in Mojave was in the front of our minds when working on the Mac version. I'll write down a note to check on that as we work on the next revision.

Posted on 2019-03-18 22:22:34
Justin Brown

Great thank you Matt

Posted on 2019-03-19 11:54:49
Bruce Cameron

Matt Bach looks like a full time job for you right now just responding to this thread! One more query for the pile, doing this on a few machines. One is getting an error on the first comp that it cannot find the specific file. All assets are there and no problem with another machine. I cant find what the file in question is, no missing assets. If i click ok again it comes up again when it gets to the Electric comp. If i just ok through these errors will the results still be accurate?

Posted on 2019-03-18 16:38:46

It won't be 100% accurate since it is missing a file, but hard to know how far off it will be. Really strange that it is saying something it can't find something, especially since it isn't saying anything is missing. Seems like a bit of mixed messages going on there.

Posted on 2019-03-18 16:50:15
Greg Platt

Hey this is amazing! Thank you so much for putting this together. If I could offer one suggestion- try putting a test counter in each comp (eg. 1/16, 2/16, 3/16...etc) to give the user an idea of how far through the tests they are. I actually failed the benchmark a few times because I thought it was just stuck in a loop and didn't know how far in to the set I was. Many of the comps look very similar so it's tough to tell where you are. Otherwise, great work! The community has needed this for a long time!

Posted on 2019-03-19 22:10:20

That's a good idea. I can't put a counter in the UI without making the benchmark a full-blown extension, but I could probably add a text overlay on each comp that has that info. Technically it will lower performance a hair, but it would be minor and consistent so I don't think it would matter much.

Posted on 2019-03-19 22:15:25
Greg Platt

oh yeah that's exactly what I was thinking- just a single text layer in each comp so you have an idea of the order.
Or, saving that, just a list of the tests to be performed in the setup comp. :)
Or heck even in the comp name!

Posted on 2019-03-20 02:20:55
Ali Khan

Hey Matt Bach i wanted to update about about my benchmark, i finally upgraded my rams uptp 64 gb, bought RTX 2080, Moved all my data to 1 TB SSD and 256 gb m.2 as cache drive, just need to upgrade my processor, but i got amazing results for now and my after effects is faster now.

https://uploads.disquscdn.c...

Posted on 2019-03-24 20:17:56

Nice!

Posted on 2019-03-25 16:37:25
Ali Khan

Thank you so much for over all guidance.

Posted on 2019-03-25 16:41:44
Jeff B

Tested this on my 5k Retina iMac and iMac Pro: interesting results
https://uploads.disquscdn.c...
Interesting that the higher clock speed on the Retina 5k resulted in lower scores than the 2.3GHz Xeon.
Kinda the opposite of the results from this video https://www.youtube.com/wat...
where it says higher clock speed is more important. Must be an architectural difference?

Posted on 2019-03-29 15:25:47
Mickosan

Imac 5K 27" 2017
i7 4,2GHz / 40Go DDR4
500Go SSD / Radeon Pro 580 8Go
Overall Score : 617 / Standard : 57,3 / C4D : 54,2 / Tracking : 78

Posted on 2019-03-30 09:34:59
Rebellion

Guys... Is this normal
https://uploads.disquscdn.c...

Cause overall I dont really feel good with the performance of my iMac Pro...

Posted on 2019-03-31 08:54:41

Usually, that kind of score means you didn't disable the disk cache. Make sure that is off, and try running it again.

Posted on 2019-03-31 15:24:05
Tobias Lamp

Hi, how do I get the result?
The test was running, but i didn't got any result.

Posted on 2019-04-01 11:45:36

If it never gave a result window, then it must not have completed. I know a lot of people with Macs get stuck on the Render tests (due to MacOS security settings I believe). You can check the "StatusLog_XXXXXX.txt" file in the benchmark folder to see what the last task that was run.

Posted on 2019-04-01 19:26:48
Tobias Lamp

0:11:55 - Done with Final Render in 957.87 seconds
0:12:17 - Starting Final Render with projnum_render=4
0:12:57 - opening comp Glitch Project
0:12:57 - Moving playhead to frame 0
0:13:5 - Show render window
0:13:46 - Deleting render queue
0:14:26 - Add render to queue
0:15:7 - Starting Final Render
0:15:7 - Send Return
0:15:8 - Waiting for Final Render

this are the last lines. would this be the last action?

edit: I answered it myself it's not the last action and the renderd Glitch file also is missing.

Posted on 2019-04-03 09:21:19

Yea, it probably isn't starting that render. Are you making sure to keep After Effects as the active window? If you click off of it and start using another program, the benchmark might not be able to start the render tests. We have to use an outside program to click the render button since there is no way to script the start of a render without it freezing the entire program for the duration of the render.

Posted on 2019-04-03 16:17:33
Tobias Lamp

Hi, how do I get the result?
The test was running, but i didn't got any result.

working on a MacPro 2010
OS 10.13.6

here are the last lines of the StatusLog

13:31:54 - Show render window
13:31:56 - Deleting render queue
13:31:57 - Add render to queue
13:32:0 - Starting Final Render
13:32:0 - Send Return
13:32:0 - Waiting for Final Render
13:46:39 - Done with Final Render in 878.902 seconds
13:46:41 - Starting Final Render with projnum_render=3
13:46:42 - opening comp Cloud Lightning Project

Posted on 2019-04-01 12:09:03
Cedric

i ran the benchmark, got overall score of 1014.5
please answer this question because i am confused: 1TB NVME (OS+Programs), 1TB NVME(Project+Footage), 1TB 860SSD (Scratch Media Cache)
i saw in one youtube video that you recommended them to use the normal SSD-> OS and Programs, NVME ->Scratch Media Cache. On your site, when selecting the premiere pro 4k workstation, on "third hard drive" there is no option to pick NVME, only first and second harddrive, so i assume you want me to use those 2 NVMEs for os,applications, footage....and normal SSD for scratch etc?
My main programs are AAE, VEGAS PRO, DavinciR., PremiereP
i use the Designare mainboard by Gigabyte with 9900k, C-State settings in BIOS ON or OFF? (if so why)

Posted on 2019-04-02 00:32:27
Justin Brown

Quick question, I know caching is disabled, but did you run the AE benchmark off of the M2 SSD? I am assembling a similar build, but want to know where to run the benchmark from.

Posted on 2019-04-02 19:19:55

Our reference score is off a 960 Pro NVMe M.2 drive, but it shouldn't affect performance at all. If you go down to a platter drive the render tests may be a hair slower, but anything SSD-level and above should give nearly identical scores.

Posted on 2019-04-02 19:25:05
Justin Brown

Awesome thanks! Your reply was just as fast as your systems! lol

Posted on 2019-04-02 19:48:28
Jerome Bryan Anonuevo Retslov

Thanks Matt and Puget Systems for this awesome test.

My system and scores:
i7-9700K @ 4.9 GHz - Air cooled
32 GB Corsair Vengeance LPX 3200 MHz
Gigabyte Vega 64 OC - 8 GB
500 GB Samsung 970 Evo NVMe

https://uploads.disquscdn.c...

I can confirm that the performance increase from more threads on the i9-9900K is very limited. Real cores with high clock speeds seems to be better than virtual threads it self.

Keep up the good work. I had fun with the benchmark.

Cheers

Posted on 2019-04-03 22:51:48
Justin Brown

I have recently updated my system to one very close to yours. I'm using the i9 9900k with the Gigabyte Z390, SSD etc. I'm also cooling my CPU with a coolermaster ML240. I had a local company assemble for me. My question is about CPU temps with regards to heavy renders inside AE. I am peaking at low 80s (C) at times while it's probably averaging lower to upper 70s rendering heavy comps. Idle is somewhere in the mid 30's to 40. The freq tops around 4.8ghz so it's using the turbo on CPU. Does that sound right for heavy AE renders and what temps have you noticed with rendering inside AE and the 9900k? BTW thanks to your sites advice, my best bench score came out to 855 which I'm very happy with considering it was 490 before the upgrade.

Posted on 2019-04-08 13:56:18

80C is a perfectly fine temperature. Intel CPUs don't start throttling until 99C, and we typically don't worry too much until the CPU starts to get over 90C in sustained, heavy loads. An AE render isn't actually all that high of a load (unless you are using some sort of batch render like RenderGarden), but keep in mind that the CPU temperature is only half the equation - CPU fan speed is the other half.

CPU fans ramp up/down depending on the temperature, and most motherboards by default don't really ramp up the fan speed until around 75-80C. So most of the time, 80C is what you will get under any significant load, but the temperature shouldn't go up much beyond that since the fans will ramp up to compensate. So really, if you are worried about long-term cooling and making sure the system will be fine even once it gets a fine layer of dust inside it is to check out the CPU fan speed and make sure there is a decent amount of headroom there.

Posted on 2019-04-08 14:13:11
Justin Brown

Thanks for the response! The AE scene I'm rendering is pretty heavy with effects, lens blurs, high res images and textures so it def pushes the CPU. The fans will ramp up pretty high and then go down. (Assuming it's preparing the next frame to render) Then it goes up again. I can hear the fans really well. It's kind of a cycle on each frame with an up and down on the fans. I'm assuming that's normal? I think the highest it's ever peaked was maybe 83C, but just watching the temps live, it's bouncing up and down from anywhere between 50-80 rendering.

Posted on 2019-04-08 14:25:02

Yep, fans ramping is normal and good.

Posted on 2019-04-08 14:27:22
Justin Brown

Awesome, well thanks again for the help!

Posted on 2019-04-08 14:58:47
Josiah

I am having the same problem ComputahNerd had 4 months ago. I just built a machine with basically the same specs your recommended build had, and I am running after effects 16.1. I have tried unintalling after effects, tried resetting preferences as your website recommends, even tried running the test with the loop issue, but it just looped endlessly until I manually ended it. Is there something I am missing?

Posted on 2019-04-13 20:14:21
Aquaman

So ran the test a few weeks ago and got something around 650 (should have written it down). Just ran it again today after upgrading ram and hard drives and got:

Overall: 685.25
Standard Score: 71.2
C4D Score: 70.1
Tracking Score: 61.6

To me that seems quite low for the system i have, but i could be wrong?

Sytem:
Ryzen 1700x 8 core 3.4GHz (OC 3.8GHz)
Nvidia 1070ti 8gb (OC +200MHz)
4x8gb G.Skill Trident RGB 3000mhz CL16 (running at 2933MHz)
M.2 Samsung 960 Evo 500GB (Boot)
M.2 Sansung 970 Evo Plus 500GB (Cache)
SSD Samsung 860 QVO 1TB (Project & Media)
MSI x370 Gaming Pro Carbon Motherboard

Now i know my system isn't quite top of the range but I'd at least expect it to beat a 2018 High Spec Macbook pro that got 760 overall. Does this suggest something is wrong with my system? Or is that actually quite a reasonable result for the above? Any suggestions on what else i should upgrade? I want to be pushing at least close to 800 overall.

I'm either off to buy more parts or taking it to get repaired. Any input is more than welcome. Thanks in advanced.

Posted on 2019-04-15 17:49:27

That score is pretty much what I would expect - maybe a hair low but that could just be due to background applications/tasks that are running. A Ryzen 2700X with a 2080 Ti scores about 730 points ( https://www.pugetsystems.co... ), so 6% lower than that feels right. Unfortunately, AMD simply isn't great for Ae since it relies so heavily on single-threaded performance which is where Intel is stronger than AMD. If you want to hit 800 points with hardware that is currently available, you are pretty much going to need to go with Intel unless you get into heavy overclocking with Threadripper.

Posted on 2019-04-15 19:06:59
crackedeggs

Thanks for the benchmark! I just did it from CGDirector PC-Builder Tool. I wanted to know if this score is accurate? The C4D score seems a tad high https://uploads.disquscdn.c...

Processor: Intel i9 9900K
CPU Cooler: be quiet! Dark Rock Pro 4 1151
Graphics Card #1: NVIDIA RTX 2080Ti
Memory (RAM): 64GB Corsair Vengeance LPX Kit DDR4-3000, CL16
SSD (OS & Apps): Samsung SSD 860 EVO 1TBGB SATA
PCIe-SSD (Project and Footage): Samsung SSD 970 EVO 1TB, M.2
Mainboard: ASUS Prime Z390-A ATX 1151
Power Supply: Corsair RMx 650W ATX 2.4

I followed all instructions on the file idk what i could've done wrong. thanks for the help! (:

Posted on 2019-04-16 07:05:45

Nope, something must not have been set right. Most likely, the disc cache wasn't disabled so the export t tests are using cached frames rather than generating new ones from scratch.

Posted on 2019-04-16 07:08:16
Roei T

Hi, I noticed that when it renders, the benchmark file is using my default preset (which so happens to be H.264 mp4 using after codecs plugin). Does the output module that eventually renders affect the benchmark score? I am obviously trying to get the most neutral result . Thanks!

Posted on 2019-04-22 14:03:21

It should be setting the render settings manually, but if it is for some reason using your default preset that will change the results a bit. Probably not to a huge degree, but enough to throw off the results slightly.

Posted on 2019-04-25 04:35:27
Roei T

Thank you. What render format should be set there? If it’s Ae’s default then QT Animation. I want to make sure I am getting accurate results because after upgrading my HD I am getting the same results as before (switched my projects drive from HDD to SSD)

Posted on 2019-04-25 07:59:36

I'm out of town at the moment so I can't check the code, but I'm pretty sure it is just the default QuickTime AVI "best" preset

Posted on 2019-04-25 08:01:25
Roei T

Will try the default thanks

Posted on 2019-04-25 08:02:34
loresuto

Hi Matt, I'm still having issues that others have reported below. I get the message about needing to set Space Bar to 'Play Once' no matter how many times I make sure its set to play once and restart. If I skip this message, I get another error that says "unknown".

I'm using 19.4.4 version of the test. Is there a newer version that addresses this?
AE 16.1.1 (Build 4) Win64 Pro.
Thanks for putting this test together.

Posted on 2019-04-25 19:53:52
Abdelrahman

I have ae 16.1.1 bulid 4
and the latest version of the benchmark. I have the settings right and i did this many times with the same results. It renders the first comp and after finishing it, ae crashes.
Please help!

https://uploads.disquscdn.c... https://uploads.disquscdn.c...

Posted on 2019-04-30 11:09:43
Isidore Isou

Hey,
I didn't see any popup window, maybe I missed it, here is the last lines of the log, is the benchmark completed?
Can you find the score in the log or somewhere else?
Thanks!

19:50:49 - Starting Half Res playback
19:50:51 - Half Res playback for is in progress
19:50:56 - Done with Half Res playback in 7.199 seconds
19:50:58 - Starting RAM Preview Half Res with projnum_half=9
19:50:59 - opening comp C4D Text Draft Project
19:50:59 - Moving playhead to frame 0
19:51:4 - Setting Comp to Half Res
19:51:6 - Setting Comp to magnify to fit
19:51:8 - Starting Half Res playback
19:51:10 - Half Res playback for is in progress
19:51:21 - Done with Half Res playback in 12.401 seconds

Posted on 2019-05-11 19:53:37

I also did not see a pop up window both times and both of my logs ened on the exact same line:

"16:25:54 - opening comp Phone Composite Project
16:25:54 - Moving playhead to frame 0
16:26:0 - Setting Comp to Half Res
16:26:2 - Setting Comp to magnify to fit
16:26:4 - Starting Half Res playback"

Posted on 2019-05-12 22:37:51
Gustavo Favoreto

That was interesting.
My macbook was:

Overall: 536.75
Standard: 49.4
C4D: 54
Tracking:: 61.9

2.9GHz Intel i9
32GB 2400 DDR4
Radeon Pro 560x

And my PC was:

Overall: 594.25
Standard: 75
C4D: 26.3
Tracking:: 61.4

Asus X299-DELUXE
Intel Core i9-7900X
1TB 960 EVO NVMe
GeForce GTX 1080 Ti
64GB 3200 DDR4

I thought that PC was way more powerful.

Posted on 2019-05-19 05:57:29

That is a really low "C4D" score for a 7900X, but the "Standard" and "Tracking" scores look right. I would recommend resetting the AE preferences on that system and running the benchmark again - it could be that you have some setting for the C4D renderer set to something that is different from the default.

Posted on 2019-05-20 18:43:57
Gustavo Favoreto

Thank you for the answer!
AE was fresh installed on that machine, I just did the setting that is require to do the test. I thought that score for C4D was really slow.

Thanks, Gus

Posted on 2019-05-21 22:56:34

Interesting. You might want to do some other benchmarks real quick to make sure you don't have some sort of hardware issue. Cinebench is probably one of the best for easily checking raw CPU performance:https://www.maxon.net/en/pr... . From what I can find, it looks like a 7900X should score somewhere around 6000 points

Posted on 2019-05-21 23:00:59
Gustavo Favoreto

I did the Cinebench. The score was 5128. I think I need change my CPU.

Posted on 2019-05-22 02:50:58
Gustavo Favoreto

I did the Cinebench, the score was 5128. I think that I should change my CPU.

Posted on 2019-05-22 02:58:22
Gustavo Favoreto

I did the Cinebench, the score was 5145 and I did a standard overclock that I have via Bios and got 5599 but the system crashed all the time. I could complete the benchmark. Maybe the CPU is not doing well.

Posted on 2019-05-22 04:06:29
Gustavo Favoreto

Hahaha that was Spam!!
I was trying to respond thru the PC, and you can see that it wasn't doing well.
I am so sorry for the spams.

Posted on 2019-05-22 17:26:13
Luke

Hey Everyone!

I just ran the test to check where my work station is at and I was surprised how low the score was, considering we spent a considerable amount of money in the last year to build it. The specs seemed pretty comparable to the higher end recommendations. My system specs are below

CPU - Intel i9-7900x CPU @ 3.30GHz
Ram - 32 GB DDR4 (2666 MHz)
GPU - Nvidia GTX 1080 Ti
SSD - Samsung 970 EVO 250GB PCI-e 3200MB/s
HDD - Seagate Baracuda 3TB 7200rpm 64MB Cache

Benchmark Scores

Overall Score: 579.25
Standard Score: 71.3
C4D Score: 34.9
Tracking Score: 54.2

Do these numbers look correct or should I perhaps be troublehsooting a larger issue with the system thats causing slowdown past the core components?

Thank you in advance for any advice!

Posted on 2019-05-30 17:35:12

The C4D score seems a bit low, but I've seen multiple people with that generation of CPU getting low results there, so that may just be how they perform. Keep in mind that the 79__X series processors are a couple years old now and isn't going to be as fast as Intel's high-end "mainstream" processors since the X-series is more about core count than single-core performance (which is what AE prefers). The newer X-series CPUs makes up for this a bit with the improved Turbo Boost 3.0 speeds, but they are still a good ~10% slower than the less expensive Intel 9th Gen CPUs.

Posted on 2019-05-30 17:45:09
Luke

Wow, that was a fast reply haha. Thank you so much for the info Matt! In the near future I am looking to upgrade my workstation at home, so if I can target where my rig at the office falls short, I can make sure I'm buying the best parts for my own rig.

Posted on 2019-05-30 17:47:20

Ha, that's what I get for having auto email notifications!

Definitely make sure you check out our AE hardware articles when you get to buying parts: https://www.pugetsystems.com/all_articles.php?filter[]=After%20Effects or our Hardware Recommendations for AE page: https://www.pugetsystems.co...

And of course, if you end up not wanting to build yourself (either for home or work), keep us in mind!

Posted on 2019-05-30 17:51:24
Luke

Will do, thank you again for your help!!!

Posted on 2019-05-30 17:52:32
Per Weum

I have an older system at work and i'm wondering if there is any point in upgrading the graphics card. Or if there is any other upgrade that would be worth considering. Im mainly working in AE and Blender.
System:
Xeon E5 2695v3
64gb DDR4 2400mhz
GTX 980Ti
Samsung Evo pro 500GB
https://uploads.disquscdn.c...

Posted on 2019-06-06 12:18:08

In AE, it really depends on how many GPU-based effects you use (they are marked with symbol of three horizontal lines and a play symbol). There are a few things like fractal noise that may benefit from a faster GPU, but to be honest, you are likely CPU bottlenecked.

Blender is a different matter if you use it's rendering engine - Cycles. That uses the GPU and typically a newer/faster GPU will give you a decent bump in performance for that.
https://uploads.disquscdn.c...

Posted on 2019-06-06 17:14:55
yezhacker

Is the Z390 platform using 64 or 128 GB of RAM?
Thanks in advance!

Posted on 2019-06-11 02:22:25

The reference system for the scoring was using 64GB of RAM, but as long as you have 12GB free for Ae to use, the amount of RAM in the system shouldn't affect performance. More RAM just lets you cache more frames into RAM Preview.

Posted on 2019-06-11 18:37:09
Hwgeek

Important news for Ryzen owners: Latest Win10 Update+AMD's Chipset drivers fixed the scheduling bug! please update and test.
https://www.anandtech.com/s...
https://uploads.disquscdn.c...

Posted on 2019-06-11 07:01:11

Our testing going forward will be with the 1903 update, so we'll see how it affects things. That said, I know we've done some preliminary testing and so far the improvement is anywhere between non-existent to around 7% faster. Definitely not bad, but don't expect to see a universal uptick in performance.

Posted on 2019-06-11 18:31:41
tomdarch

Obviously, we're all massively looking forward to your Ryzen 3xxx benchmarking!

Posted on 2019-06-11 13:53:36
Krzysztof Karnicki

Great update!
Unfortunately the benchmark did not calculate the Render section on my machine.
The suspects:

1. The first list of tests all got a red cross in front - something definitely did not compute.

2. In the Renderlog file I found these:
a) "Unable to read VR path registry from C:\Users\...\AppData\Local\openvr\openvrpaths.vrpath
b) "aerender ERROR: No output module template was found with the given name"

Any ideas?

Posted on 2019-06-11 15:32:56

We use just the default render templates, so likely you either have some plug-in that is overwriting it, or some issues with your Ae install. What I would try would be to clear your media cache, reset preferences to default, disable any plugins, and try running it again. If you don't want to lose your current preferences, you can backup the folders in %appdata%\Adobe\After Effects

Posted on 2019-06-11 18:48:01
Aleksandar Aleksic

Great UPDATE :D Tested on Macbook Pro 2017
System: https://uploads.disquscdn.c... https://uploads.disquscdn.c...

Posted on 2019-06-11 18:10:21

That is the previous version, not the new V0.5 BETA that got uploaded yesterday. And that is definitely not a valid score, wayyyy too high. The new version has some code in place to try to combat the most likely causes of that.

Posted on 2019-06-11 18:55:37
Aleksandar Aleksic

I will then put to test v0.5 BETA and post the result. To add, it was strange to me for high score

Posted on 2019-06-11 20:17:53
Matthew Sanchez

I can't find the "Play Once" section. Any help would be great!

Posted on 2019-06-11 22:57:28

Added an image to the "How to run the benchmark" section, but it should be in the "Preview" window.

Posted on 2019-06-12 00:03:12

Just got my new comp running today. Seems aight? https://uploads.disquscdn.c...

Posted on 2019-06-12 00:06:07

Yep, pretty much spot on! Same CPU as the reference system, and the 2080 Ti I wouldn't expect to be much faster than the 2080 we used in the reference system for After Effects (and AE doesn't really use multiple GPUs).

Posted on 2019-06-12 00:10:04

That's a relief. I have this fear of spending a lot of hard earned $ then having the system not run much faster than my last. Seems like I actually upgraded this time. The 2080ti's are for Octane :)

Side note: I dug through a lot of your articles and they were very helpful with configuring this setup so I really appreciate all your hard work. All of the Puget posts are very insightful and it's great to have an AE benchmarker. I ended up getting a PLX board to have the best of both worlds AE/Octane. Not experiencing any "latency" issues. So far everything is running solid. Hopefully it stays that way with future software/hardware updates :)

Posted on 2019-06-12 00:45:41
Kennon Fleisher

I just posted my results above, scored slightly higher... but one thing I found interesting is that my late-2018 i9 6 core MacBook Pro scores a lower benchmark, but in practical use it beats my PC in render time by 10-15% consistently across various projects, every time. There was no instance where my PC was faster, which kind of makes me sad. It also makes me wonder about the new 8 core MacBook Pro...

EDIT: I've found a few comps that took the laptop 3x longer to render than the PC. It seems like there is more work to be done in determining what causes AE To run faster in some instances on the MBP than on the PC, and why the PC blows some of these tasks out of the water compared to the MBP.

Posted on 2019-06-14 01:14:44
Kennon Fleisher

I built your Johnny Cache / recommended machine myself and upon running the benchmark, I only scored a 1510 overall - and a whopping 0 on my render score.

I have no clue why this is happening or why this machine is running so slow, but this setup is atrocious as is...

Posted on 2019-06-13 20:30:02

If it gave a 0 for the render test, then something is wrong with the setup - either for the benchmark or for After Effects in general. Are you using the newer V0.5 version of the benchmark we put up just the other day?

Posted on 2019-06-13 20:33:08
Kennon Fleisher

Yes I did. And I ran through the steps and made sure everything was right. My CPU seems to be only running at 13% while rendering from AE... everything seems to be hooked up correctly and properly cooling, too.

Posted on 2019-06-13 20:36:11

Can you post the actual results? In fact, both the AEBenchResults and AEStatusLog files would let me figure out what went wrong. My guess is that one or multiple render tests failed (either to not having enough RAM/VRAM or something else) and it failed in a way I don't have a catch programmed for yet.

Posted on 2019-06-13 20:44:08
Kennon Fleisher

Here it is: https://uploads.disquscdn.c...

I know, I know. I should have just had you guys build this machine for me instead of trying to do it on my own :(

Posted on 2019-06-13 20:46:45

Yea, the render tests completely failed. There should be a Renderlog.txt file in the export folder that will have more information on why it failed.

Also, the C4D results are wayyy to high. It may not be correctly linking to the C4D project. My bet is that if you figure out why the renders are failing, that will fix both issues though.

Posted on 2019-06-13 20:59:29
Kennon Fleisher

After updating all drivers, I got a lot better scores... I think? Let me know if this seems right. I've got the i9 9900k, Z390 Designare, 64 gigs of DDR4 ram, running on 2 NVM-e M2 drives (3500 mbps) and OS is on a standard Samsung SSD. https://uploads.disquscdn.c...

I can email you the actual results; what's a good email?

Posted on 2019-06-13 21:55:29

Yea, those score look perfect. Slightly higher than the reference, but that is probably because you are on the latest version of Windows 10 (1903) and the reference was on one version back.

Glad you got your system working and performing as it should!

Posted on 2019-06-13 22:01:09
Kennon Fleisher

Thank you!

Something fairly interesting to note... I've been running side by side renders with my core i9 6 core 2.9ghz MacBook Pro all day, and the MBP has beat the PC in render time by about 10-15% each time. I find it a bit odd because it's benchmarking so well. I'll keep running more tests to see what happens.

EDIT: I've found a few comps that took the laptop 3x longer to render than the PC. It seems like there is more work to be done in determining what causes AE To run faster in some instances on the MBP than on the PC, and why the PC blows some of these tasks out of the water compared to the MBP.

Posted on 2019-06-14 00:52:22
Kennon Fleisher

Apologies for continuing to post here, but I'm wondering if this seems like a normal amount of CPU usage for a heavy render out of AE... it seems like it's running slower than it should be, right?

This is the CPU process while rendering out a 10 minute comp which takes 2 hours and 45 minutes on the 6 core MBP and 2 hours and 30 minutes on the i9 9900k.

Seems like it should be faster, right? I benchmarked the MBP at 530 earlier tonight so it seems like this should be twice as fast. https://uploads.disquscdn.c...

Posted on 2019-06-14 05:20:27

That is pretty common for After Effects - a good portion of the things it is doing is going to be limited to a single core. I think a lot of that is due to the highly threaded tasks being offloaded on to the GPU, which leaves only the single or moderately threaded tasks left to be run on the CPU.

That said, I would expect a MBP to be quite a bit slower, especially for long renders where throttling is likely going to come in to play on the laptop. Are you sure the render settings are the same between the two systems? AE will use the disk cache as well if available, so it may be that you have caches files on the laptop and not on the desktop.

You may also look in to using something like RenderGarden https://www.mekajiki.com/re... to improve render times. This lets you spawn additional render nodes on your system to better utilize all the CPU cores. Each instance does increase RAM/VRAM usage, however, so depending on the resolution of your project there is limit to how many you can spawn. If you are comfortable with scripting, you could also make your own version of rendergarden using aerender https://helpx.adobe.com/aft...

Posted on 2019-06-14 17:10:16
Andy Hay

2017 iMac 5k i7 64GB RAM: 660.

Not bad.

Posted on 2019-06-14 13:45:43
potahto

I'm running this on a machine with a GTX 660 GPU. The benchmark crashes on the Cloud Lightning Project -- log says "Return on error: error code -17, ..\..\Src\IntelMSDK\IntelD3D11Context.cpp 297". When I open the comp, I get AE's error message "After Effects has encountered a failure related to GPU-enabled effects on this frame. This is likely because your GPU is out of memory. . . "

Switching to Software Only in Project Settings doesn't work because (I assume) the script sets it back.

Is there a way to run the benchmark so it switches CUDA off when the card can't handle it, the way I would in real life?

I'll eventually get a new card for this machine, but it'd be nice to be run the rest of the benchmark for a "before" snapshot.

Posted on 2019-06-15 19:32:53

Right now, we have to force set it to CUDA if available, followed be Metal, then OpenCL, then software only as a fallback. I might at some point add a config option to let you change that priority, but I'm not sure when I would be able to add that feature.

Posted on 2019-06-17 16:02:26
Nick Hamilton

i7-5960X (3.0GHz), GTX 980Ti, 64GB RAM and SSDs: Score of 570.

This was a top of the line machine not even three years ago haha. Gonna shortly be upgrading to a new machine with an i9-9900K & RTX 2080. Hoping to almost double that score of 570!

Posted on 2019-06-24 14:09:00
Tyler

I have tried everything to get this test to work After Effects Version 1.6.1.2 (Build SS) and will not work on clean install on new machines. :( Hopefully an update is coming?

Posted on 2019-06-25 12:51:51

Do you mean Ae version 16.1.2 (Build 55)? I'm not aware of any issue with that version and the latest version of our benchmark (.51 BETA). Can you be more specific about what problem you are running in to? Is there a StatusLog_____.txt log file being generated that you can link?

Posted on 2019-06-25 17:21:29
CASEYV

I'm receiving a Dynamic Link error on the tracking stabilizing test. I think the 2nd last one? Has anyone else had this issue? It means the tracking score fails and i dont get an overall score. https://uploads.disquscdn.c...

Posted on 2019-07-04 02:16:53

Got this too. I just started the benchmark again and all worked.

Posted on 2019-07-06 10:46:37

Thanks for this benchmark!

I tested it on my Hackintosh and also on Windows, both with the same hardware and the same AE version.
Just switched from macOS to Windows after ~14 years because I really don't want to wait for a silent AMD GPU or years of waiting for new NVIDIA drivers (sure there wont be any for the next years) o_O Moved from my last iMac to a Hackintosh about 3 years ago.

My current setup:
* i7-8700k
* 32GB
* GTX 1080

macOS @ 4.9 GHz / out of the box
Overall: 794.25 / 703.25
Render: 82.8 / 71.1
Preview: 77.9 / 67.8
Tracking: 78.2 / 71.3

Windows 10 @ 4.9 GHz
Overall: 910
Render: 86.9
Preview: 84.9
Tracking: 101.1

Posted on 2019-07-06 10:41:00
Nick Perry

Thank you for sharing the test

Just built my first PC workstation after moving to
Mac 10 years ago.

Intel i9900K + NZXT Kraken X62
Gigabyte GeForce GTX1070 8192 MB
Corsair Dominator Platinum DDR4 64 GB (4 x 16 GB)
3000 MHz C15 XMP 2.0
Samsung MZ-V7E1T0BW 970 EVO 1 TB V-NAND M.2
Samsung SSD 860 QVO 2TB
Asus Prime Z390-A

Overall: 742
Render: 73.1
Preview: 73.3
Tracking: 76.

Posted on 2019-07-08 13:50:22
Meshraz

Hello! I am having trouble opening the project on the 2009 Mac Pro of the company I work for.
Version of AE is 2018 (can't update it further because the machine (or the OS) does not support newer versions)
Specs of the machine are:
OSX El Captain
CPU: 2 x 2,26 GHz Quad-Core Intel Xeon
GPU: Radeon HD 4870 512MB

https://uploads.disquscdn.c...

Posted on 2019-07-29 20:59:38
Kromaqi

Thanks for this tool! I actually put together a system similar to the one you guys currently spec out for After Effects, with an i9 9900k and a Noctua nh-u12s, but I'm running into some overheating issues. It seems to spike into the 90s on occasion and I'm not sure why. I've already re-seated the cooler and re-applied thermal paste, but it made no change. No overclock or anything. With that processor and cooler, what temps do you guys normally get?

Posted on 2019-07-30 14:59:29

I just checked a system with the 9900K, Noctua cooler, and a RTX 2080 Ti Founders Edition (a card that pumps a lot of heat back into the system) - and it tended to be in the 70s and 80s under load, with a couple peaks into the 90s only during Linpack - a very stressful test, where it reached 96C at the very highest. I wouldn't really want to see a CPU sit in the 90s for extended periods of time, but if it is just rare spikes then that may not be a big deal. You could try to set the fan to a more aggressive profile if you want, or maybe add a second fan (the NH-U12S can mount one on each side of the heatsink, if you want). More overall airflow through the system could potentially help as well, so that the air passing through the Noctua is cooler to begin with and can therefore pull more heat away from the heatsink.

Posted on 2019-08-08 18:42:51
Mauricio Garcia

well mine failed, Imac 2015 what should i do? https://uploads.disquscdn.c...

Posted on 2019-08-04 06:54:42

Most likely, that test needs more VRAM than your GPU has. I think that card is only 2GB, which probably isn't enough.

Posted on 2019-08-05 17:53:09
Meshraz

Does anybody have a benchmark result for the Ryzen 9 3900x?

Posted on 2019-08-26 20:58:17

Hey Meshraz, the 3900X is included in our recent CPU Roundup testing for After Effects: https://www.pugetsystems.co...

Posted on 2019-08-26 21:32:40
Meshraz

Thank you for letting me know! I checked the chart. Huge thanks for you guys!

Posted on 2019-08-28 19:01:37
Meshraz

Oh, there is something else that maybe you could help. From the results, my PC should be faster than the iMac Pro 10 cores, but the newly bought iMac Pro in the company I am working in renders MUCH faster than my PC. Any ideas why?

My PC specs are:
Ryzen 7 1700 locked @3.7Ghz
2933 Mhz 32GB RAM
GTX 1080TI
HyperX Savage SATA III SSD

I think the iMac Pro has a 1TB SSD, which I assume is a PCIe NVME because the write speeds reach 3GB. Does that make a difference for rendering or is it just that the Mac OS is better suited for rendering in After Effects?

Posted on 2019-08-28 19:15:20

The Ryzen 7 1700 is a few generations old at this point, I'm not surprised a modern iMac Pro is faster. It has been a long time since we tested the first generation Ryzen CPUs, but if I had to guess, I imagine a 10-core iMac Pro to be somewhere around 15-25% faster than your PC.

The storage speed could be a part of it, but at least with the render settings out benchmark uses, it shouldn't be a significant concern. In the real world, a faster drive will help with RAM Preview disc caching, but that shouldn't show up in render performance. Your CPU is really what is going to hold back performance in AE.

Posted on 2019-08-28 19:24:33
Meshraz

Ryzen 1000 is just about 5% slower than Ryzen 2000 from what I remember.
I`ll have to do some more tests, but the Imac rendered a 7680x2160 project which was only using shapes and a basic starfield particular TWICE as fast as my PC, but there was also another test project that rendered almost as fast as the iMac Pro.

Posted on 2019-08-30 22:09:05
Steven

I just finished a Ryzen build, with the Ryzen 9 3900x. 64GB of G-skill Neo 3600-16 RAM. I had an upgraded but aging MacPro 5,1 that is being phased out. My results went from 393 overall (LOL), to 943.
https://uploads.disquscdn.c...

Posted on 2019-08-27 05:28:30
Meshraz

Thanks for sharing!

Posted on 2019-08-28 19:00:49
michel molder

Hi! Thanks for all this! I just completed this test with my newly build hackintosh and generated a score of 843. Eventhough I hoped it to score better, both After Effects and Premiere perform terribly slow, way beyond what would feel like that score. Do you guys know of some settings that need to be different for those programs to work in an optimal way?
Thanks!

Posted on 2019-09-08 08:50:47

Performance is always going to depend on exactly what you are doing including what codec your media is and what effects you use. Our benchmark gives you an idea of how well your system can perform in these applications, but there are a ton of things it doesn't measure. For example, if your cache files are on a slow drive, that can make importing media really slow but that isn't something our benchmark looks at. Similarly, if you are using a codec that is poor for editing (high bitrate H.264 for example), that is going to result in pretty slow performance.

Posted on 2019-09-09 17:44:04
michel molder

Thanks for your reply Matt. Crazy thing is that I try to nail down what the problem is with my new Hackintosh and it doesn't make much sense. So I disconnected from the internet and closed down all my other apps. Editing in native Pro Res 4K in AE while doing nothing crazy is giving me very slow results, while the benchmark scores a 903. Cashe is on an empty and seperate SSD, my whole system is in RAID 2.0 on 2 different SSD's and all my projects are running from an internal SSD as well. I really feel that there is something wrong with my machine but the benchmark is telling me something else. Any extra thoughts? Would appreciate any out of the box solutions or extra tests.

Posted on 2019-10-10 16:33:22

Out of the box solution? Buy one of our After Effects workstations :)

But really, the fact that it is a Hackintosh really makes it difficult. If it was me, I would probably dual boot (or just use that empty cache drive temporarily) Windows 10 and try your project there. At least that way, you can determine if your OS is what is causing the problem. If you get the same performance in Windows, maybe see if you have someone you can send a test project to and have them try on their machine. It could be that the performance is simply as good as it is going to be so you need to look into things like proxies or other workflow adjustments..

Posted on 2019-10-10 16:38:35
trjones123

https://uploads.disquscdn.c... I tried to run the benchmark yesterday and got error pop ups related to memory leaking? See attached image. An error window popped up for the first 10 out of the 32 tests. The benchmark did complete, but I am just not sure if the score is still valid.

Posted on 2019-10-04 11:18:45

That looks like an issue with either AE itself or your system overall. You might want to try reinstalling After Effects because of there really is a memory leak, that isn't really something you want.

As for the benchmark scores, I'm not sure if it will affect them or not. As long as the benchmark completes and actually gives results, I don't think a moru leak should affect them. But I'm not 100% certain on that.

Posted on 2019-10-04 14:17:31
Nick Beardo Frewin

getting the same and have tried earlier versions and resinstalling - same problem

Posted on 2019-10-10 17:45:14
Nick Beardo Frewin

hi - just got a a new system (i9 with 64gb ram and a RTX 2080ti) - when i try and run i get an error saying unable to confirm ram is available for AE to use (have 60mb assigned to AE). any clues?

Posted on 2019-10-09 22:38:15

Hopefully you mean 60GB! That should just be a warning, so if you know you have enough RAM allocated you should be able to just continue.

The most likely cause of the issue is if you are not using After Effects CC 2019. A lot of the benchmark doesn't work on older versions.

Posted on 2019-10-09 22:57:04
Nick Beardo Frewin

yeah sorry 60gb assigned and have CC2019 16.1.3 - it's an i9-9960x - getting a score of 735, Render 81.8, Preview 63.7, Tracking 75.1
not giving amazing results - in comparison my laptop (i7-8750H 32gb ram. GTX 2070) is getting 695!

Posted on 2019-10-09 23:48:49

An Intel Core i9 9960X with a RTX 2080 Ti should be giving you a score around 945 - that is virtually the same specs used in one of our recent articles: https://www.pugetsystems.co...

Sounds like your new system is under-performing for some reason. If you have one of our workstation, I would reach out to our support department.

Posted on 2019-10-09 23:59:37
Nick Beardo Frewin

hmm strange - i am also getting the message about the loop to one shot not being able to be found and also some memory leak messages.
i guess it's an AE bug, and presume it could all be affecting my scores

Posted on 2019-10-10 17:44:33
michel molder

Hi! Thanks for all this. I tried your benchmark and my new Hackintosh systems scores 903. Problem is that it's extremely slow performing with 4K video in After Effects. I'm kinda lost having quite a fast machine that does not work properly. Is there a list with proper settings that can guide me or do you guys have any experience with good benchmarks and slow results in reality?

Posted on 2019-10-10 12:07:46

That score is pretty decent (within 10-15% of the fastest hardware), so your system itself is probably just fine. Typically, in situations like this it is one of two things:

1) Your general workflow has some sub-optimal aspects. This could by anything from using a codec like H.264/H.265, trying to stay at full res preview when it you should turn it down to 1/2, 1/4 or even 1/8, working at a higher bpc than you need, etc.
2) Simply ambitious expectations. After Effects isn't a video editor, so it is going to be really difficult for it to play almost any sort of video (especially 4K) in realtime no matter how much money or hardware you throw at it. You probably will need to work with lower res proxies, or if it is longer video do most of the edits in Premiere and use dynamic link to only do the VFX/motion graphics work in AE.

As far as settings go, there isn't all that much to try if you are getting a good benchmark score. Only thing I can think of is to move the disk and media cache to a fast, dedicated cache drive.

Posted on 2019-10-10 16:16:37
Niichan !

I tried to open it on After Effects CC 2018 but it says that it cannot be opened on my current version...

Posted on 2019-10-26 14:03:06

Yes, this benchmark is for Ae 2019 only. It technically should work in older versions if you open it in Ae 2019 and save it in an older format, but that isn't officially supported since we haven't tested the current benchmark with older versions of Ae.

Posted on 2019-10-28 17:04:55
Yoav Koplowich

Hi, I have a windows 10 machine and runing latest AE version.
during the tests, a black windows open and says:
2019-11-08 14:17:11:677 : ERROR *** IVCAutoReleasePool: Object 0000000023477BE0 of class vcfoundation::data::DataWrapper autoreleased with no pool in place - just leaking
also for some reason in the begining of the test it gives a warning about not being able to recognize how much ram is alocated for AE and an error about the spacebar loop, even though I changed it to play once and restarted AE

Posted on 2019-11-08 12:20:36
John Ryan Preece

When i try to run the benchmark, I get a prompt saying AE can't verify the amount of ram dedicated to AE....

Posted on 2019-11-08 16:31:57
Luiz Amoasei

Dear friends: The AE Bench use only Metal API in Mac? Can I perform the same tests with OpenCL?

Posted on 2019-11-13 19:35:13

Right now that isn't an option - we have to make sure we set it since it would otherwise default to Software Only since we tend to create the projects using CUDA. The ability to specify OpenCL/Metal/CUDA is planned in a future version, however!

Posted on 2019-11-13 19:38:33
matinciel

Hello, so happy to see this !!!

Try to run the script on Windows 10 (haven't check but latest version probably) - AE 17.00 build557 (just made a clean install of AE during last upgrade).

i7 6800k
GTX1070
64RAM
Asus X99E

I got a first warning message saying that the script cannot check for the amount of RAM (64 / 6 reserved in my pc), click yes anyway.

Then the script starts the test, and error message : " After Effect Warning: ERROR: The system cannot find the file specified". (seems to happen at step 1/32 phone composite project, at least only this arrow appear in blue, no other arrow).

Let me know if I could help debugging (providing the details step :-) ).

Posted on 2019-11-18 21:29:04

I get this error too - which means that I cant benchmark at the moment.

Posted on 2019-11-19 11:45:32

The benchmark currently doesn't support After Effects 2020. We are working on an update, but there are a bunch of other improvements we are making that need a bit more testing before we are comfortable putting it up for download. I would estimate about 2-3 weeks until we have the next version available.

Posted on 2019-11-19 18:02:07
沈漢迪

I got this error when I hit run
It means After Effects warning ERROR : The system cannot find the file specified
I'm using CC 2019 tho (16.0.0 Build 235)
https://uploads.disquscdn.c...

Posted on 2019-12-11 12:11:21

You are running AE 2019 -the benchmark needs AE 2050 to run.

Posted on 2019-12-11 15:54:06

Not only did I typo (meant After Effects 2020), but I forgot that I actually did include support for After Effects CC 2019! So that is a bug that I will get fixed right now. Pretty sure it is related to Adobe dropping the term "CC" from their product names with the 2020 launches, so it should be an easy fix. I should have a .081 BETA version up in the next hour or so.

Posted on 2019-12-11 17:32:45
Cklein404

Hi!

if I update my version 0.5 for 0.8, the scores obtained on the configs benches at 0.5 will be comparable to the scores obtained on other machines with version 0.8?

Or is the score no longer calculated in the same way and the results are no longer comparable?

Posted on 2019-12-13 10:40:42

The tests themselves have not changed, so the scores between those versions should be identical (assuming you are using the same version of After Effects of course).

Posted on 2019-12-13 16:46:05
Neooooo

Hello! thanks alot for this very useful project ! but at every turn when my computer finish a video , then popup a window tell me my computer didnot finish render, but actually i did , i can see all the video i rendered in export and play well ..... , any ideas? https://uploads.disquscdn.c... https://uploads.disquscdn.c...

Posted on 2019-12-20 17:06:17

Are you running the benchmark from network drive? Either that or something abnormal in the folder path is the only reason I can think of it not being able to verify that the render actually finished.

Posted on 2019-12-20 17:13:40
Neooooo

nope, i run this in a very local drive ...
I have friend who encounter same problem, we both use multiple graphic cards, I dont know if this is reason
https://uploads.disquscdn.c...

Posted on 2019-12-20 17:32:58

It may be a language thing. Can you toss the latest "aestatuslog.txt" and the log file in the export folder in an email to mattbach@pugetsystems.com?

You might also try changing the Windows language to English (and AE to English if that has a separate setting for language) and seeing if that clears it up. My guess is that something in the script may not be able to account for non English characters.

Posted on 2019-12-20 18:05:26
Neooooo

Sure, i will send evey logfile in my folder to you , thanks a lot !

Posted on 2019-12-20 18:46:34
Nikolay

Having the same issue when I start the benchmark. Is there a sollution to this already? : /

Posted on 2020-01-01 19:31:56

Thanks for the logs! Not sure why it is having problems, but maybe there is some issue exporting to a non-C: drive? The log files definitely help and I'll see if I can get some potential fixes in place for the next version.

Posted on 2020-01-02 19:15:01
Bjorn Peters

I did everything right, step by step. When I run the script file I get an error that says: Unable to confirm how much RAM is available for AE to use. In my preferences I have 16 GB available for other apps, 48GB for AE so I have no clue why this isn't working :(

Posted on 2019-12-23 15:19:36

Usually that means you are not using a supported version of AE or that the OS/app language is not English. It you know it is set right, however, that is just a warning that allows you to proceed. It doesn't hard stop the benchmark.

Posted on 2019-12-23 15:48:46
Daniel Walker

Well, I am trying to see my score but the test stops running after the "glitch" test. That comp never closes and nothing else appears to be happening. I don't expect to have a great score, but I do have 64GB of ram with 58 or so dedicated to AE. Running on a 16core AMD though I know the core count doesn't matter much. Do have an RTX 2070 Super in here though.

Posted on 2019-12-26 14:14:18

I've been having a bit of an issue with the playback tests not completely correctly on the new SUPER cards as well. I'm not sure why (and maybe it has nothing to do with the SUPER cards at all, but one of the recent AE updates), but I am working on making that part of the test more robust.

Posted on 2020-01-02 19:08:22
Daniel Walker

That's cool! It would be nice to know where I stand here! Plan on doubling my RAM sometime soon I hope!

Posted on 2020-01-04 14:43:34

Just a heads up, we just posted a new version (8.2 BETA) that should resolve the issue completely. Still not sure why the method we've been using for so long decided to suddenly stop working, but the AE dev team helped us work out much, much better methods to initiate and log playback and render times. Give the new version a shot!

Posted on 2020-01-17 23:27:57
blindcheesecake

https://uploads.disquscdn.c... Tried to run the benchmark but I'm getting this error:

Posted on 2019-12-27 16:34:54
Michael Payne

I have the exact same error. Have you figured out a solution to it yet?

Posted on 2020-02-17 23:43:22
Usman Aslam

hi sir i am using hp z420 workstation E5-1650 0 @ 3.20GHZ 6 CORES WITH 16GB ddr3 ram and 1TB HDD and 240GB hp ssd and radeon rx 5708GB GPU. but its not working even bit fine on after effect but little bit better in premiere pro. SO please guide me to do changes to have best PC for after effect.

Posted on 2019-12-29 12:00:10
Nikolay

Hey Matt, I am trying to run the benchmark on my system but each of the stages says - InitializeFileLogging Failed

Do you have any idea why that happens? I have set all settings before the render as they should be.

This is my setup:

AE 2020

Aorus Master X570
Ryzen 9 3950 X
GeForce RTX 2080 Ti
Corsair MP600 SSD 2TB
128BG RAM

Posted on 2019-12-30 20:04:07
Boris Dimitrov

Hey Matt, massive thanks to you and the team for your invaluable work & the insights you bring to the community!

Heres the results I got on our speced out 2019 5K iMac:
Intel Core i9 9900K 8 Core
128GB RAM
512GB Internal Flash drive
1TB Samsung T5 SSD (Puget Benchmark files ran from here)
Radeon Pro Vega 48 8GB

Should I attribute the ~20% lower performance vs the benchmark 9900K machine to the more limited cooling of the 5K iMac and the geared-for-less-noise control of the CPU by the motherboard, or are there other factors at play as well? Do you think I can try and boost the performance by overriding the fan to run more intensely?

Thanks again!

https://uploads.disquscdn.c...

Posted on 2020-01-06 11:21:29

It could be a lot of different factors, but if I had to guess it is probably a combination of CPU throttling due to heat and the simple fact that it is MacOS. We haven't done direct comparisons between MacOS and Windows for AE, but from the testing we have done versus the 14 iMac Pro in some of the recent articles, I would estimate that with the same hardware, you would get about 10% higher performance in Windows.

Those two things combined could very well cover the 20% lower performance you are seeing versus a Windows-based system with a 9900K

Posted on 2020-01-06 18:01:01
Boris Dimitrov

Good to know! I wasn't aware that the are performance differences between the two operating systems, thats another really useful insight to keep in mind. After the deadline crunch we are in right now, I'll do some research and experimenting with overriding the fan operation to see whether I can boost the benchmark scores and report back in case you or others are interested in the results.

Thanks again for being so awesome!

Posted on 2020-01-07 12:11:30

Thanks so much for providing this tool, and thanks also for including it for Mac users.

I'm a longtime AE user, but haven't set up the software from scratch in quite some time (I was using AE2017 on a 2010 Mac Pro 5,1 for years).

I recently upgraded to a new Mac Pro 7,1 (3.2 GHz 16-Core Intel Xeon W, 384 GB 2933 MHz DDR4 RAM, AMD Radeon Pro Vega II 32 GB, and two Sonnet M.2 4x4 PCIe cards, each one set up as 8TB SSD RAID 0 arrays, which are home to my projects, PSD project files, ProRes stock footage, and disk cache).

I've been having some strange RAM preview issues (slow/stuttering playback -- around 8FPS for 24FPS comps) and freezes. I'm guessing it has something to do with my prefs setup, but in trying to troubleshoot it I decided to take the benchmark test -- here's what I got: https://uploads.disquscdn.c...

Posted on 2020-01-08 23:37:42

Hi Matt -- quick questions -- with the amount of RAM I have installed on my system (384 GB), how much would you recommend I reserve for other applications in my prefs? I currently have it set to 1/4 (96 GB)?

I also have my Maximum Disk Cache Size set at 1000GB (on a Sonnet M.2 4x4 PCIe card, set up as an 8TB SSD RAID 0 array), and my Previews/Fast Previews/GPU Information/Texture Memory is set at 13100 MB for my AMD Radeon Pro Vega II 32 GB).

Posted on 2020-01-10 15:53:23

That all really depends on what else you are using the system for. If you are just using Adobe apps (they share the allocated RAM setting), then you really don't need all that much reserved for the rest of the system. 96GB is probably more than you need, but ~300GB is also pretty overkill for Adobe apps, so I suppose you might as well.

If you also run something like Resolve, C4D, or other non-Adobe apps, then - depending on how many of those apps and how much resources they need - you could adjust the reserved RAM. With how much RAM you have though, I doubt you would see any advantage to changing anything unless you are running into problems.

Posted on 2020-01-10 18:31:15
Cosmo Collider

Using AE 2020. The test isn't completing for me. It gets through the list of 10 or so, then initiates the next round of tests where you see the phone composite, the clouds, the glitch and then the rain test. It then just sits there and doesn't move to the next test. I removed all 3rd party plug-ins, purged all prefs multiple times. Then reinstalled (purging prefs too). But doesn't seem to work. Just hangs on the Rain Project footage window. Previously, it hung on the glitch window. But then after purging it was getting one step further repeatedly.

I put all the logs in a dropbox here:
https://www.dropbox.com/sh/...

A sample log:
5:49:14 - Done with Full Res preview of Glitch Project in 13.241 seconds
5:49:14 - Full Res Preview Glitch Project result - 7.6 FPS
5:49:16 - Starting RAM Preview Half Res with projnum_half=5
5:49:17 - opening comp Realistic Rain Project
5:49:17 - Making sure we are using the Classic 3D Renderer
5:49:17 - Moving playhead to frame 0
5:49:22 - Setting Comp to Half Res
5:49:24 - Starting Half Res playback

Posted on 2020-01-09 15:33:10

Yea, this looks like some issue with the latest update for AE 2020. It looks like AE randomly decides to halt scripts while previewing a comp, which messes up the playback detection. I think I have a workaround, but need to make sure it always works before I push out a new version.

Posted on 2020-01-09 18:01:20
Cosmo Collider

Thanks for the response!

Posted on 2020-01-09 18:38:49

(Not sure if you will see my response to Matt's comment, so I'll post it here too)

New version (8.2) actually just went up that should resolve the issue completely. Still not sure why the method we've been using for so long decided to suddenly stop working, but the AE dev team helped us work out much, much better methods to initiate and log playback and render times. Give the new version a shot!

Posted on 2020-01-17 23:27:03
Matt O'Donnell

I'm having the same issue in both 2020 and 2019. Any way to download the previous version of the benchmark?

Posted on 2020-01-17 20:41:28

New version (8.2) actually just went up that should resolve the issue completely. Still not sure why the method we've been using for so long decided to suddenly stop working, but the AE dev team helped us work out much, much better methods to initiate and log playback and render times. Give the new version a shot!

Posted on 2020-01-17 23:26:34
Kenneth Lee

Hi Matt, thanks so much for releasing this bench marking tool! It is very helpful. At my design agency two of us motion designers upgraded from the old trashcan Mac Pros to brand new Alienware Aurora R8 PCs over the summer. The performance has been an incredible boost from the mac pros, but i felt as though we werent getting enough out of it. I ran the benchmark and it turns out we arent. Our machines are the same, running the same hardware. (Windows 10, Intel Core i9-9900K, Alienware brand liquid CPU cooler, 64GB 2666 MHz RAM, NVIDIA GeForce RTX 2080 Ti, 850 Watt Power Supply,) I spec'd our machines out based on the "Building the Ultimate After Effects Computer" article from School of Motion that worked together with Puget Systems. We have almost the exact same hardware from that machine. I'm sure this could be a loaded question but why do you think my score is so low? I was not running any other programs while running the bench. I've been trying to troubleshoot my PC for a little while now and have not been able to come up with any valid reason why i'm not getting the proper performance. One thing I did notice is on my results image for my GPU it says "LogMeIn Mirror Driver" I wonder why it says that there. Log me in is the app I use to remote into my desktop when working outside of the office.

Sorry to send such a loaded question but if you have any thoughts for how to help me I would greatly appreciate it. Thanks so much!

https://uploads.disquscdn.c...
https://uploads.disquscdn.c...

Posted on 2020-01-20 20:08:43

Our latest articles use AE 2020 (version 17.0), so that could be part of it, but I doubt that is the whole picture. My guess is that either their BIOS is set to be very conservative on the Turbo speeds, or the motherboard/cooling is not able to maintain Turbo speeds like it should. You could run something like HWinfo or Coretemp and see if the CPU is hitting ~5GHz like it should in single threaded workloads. You probably won't see exactly 5GHz since there are always background tasks going on, but if it is any lower than ~4.5GHz, that is probably the issue.

Actually, I just noticed that the CPU was overclocked to 4.7GHz. So, you should just see 4.7GHz solid under any load. Honestly, the overclock is probably bad and the CPU is having to throttle in order to keep the system from crashing. I have no idea how their BIOS works, but you could work with their support to remove the overclock and see if performance improves.

You might also want to run some additional benchmarks to check as well. Cinebench R20 ( https://www.maxon.net/en-us... ) has both single and multi-threaded tests, so it is a good benchmark to check the theoretical performance in single and highly threaded workloads. I believe a normal score for the 9900K is about 500 points for the single threaded benchmark, and ~4850 for the multithreaded. That will tell you at least if it is a fundamental issue, or a problem specific to After Effects. At that point, you should have some ammunition to take to Alienware to get them to figure out what is going on.

Posted on 2020-01-20 21:00:50
Kenneth Lee

Thank you so much for the fast informative reply! I will use what you said to look into it.

Posted on 2020-01-20 21:30:49
Dan Hopkins

Hey! So I want to make sure i'm understanding the scoring correctly. Is a "perfect score" based on the setup listed at the bottom? I have a system that I thought would be on par, or better, and i'm scoring an 850 overall. I'm a numbers nerd, so i'm just curious. Is my GPU the limiting factor?

I have the following setup:

Intel Core i9 10920x 12 Core
NVIDIA GeForce GTX 1070 8GB
64GB of 3200MHz RAM
SanDisk SSD PLUS 1TB Internal SSD
Windows 10 (1809)
Adobe After Effects CC 2020

Thanks!
Dan

Posted on 2020-01-24 18:21:37

There is no such thing as a perfect score. You can get 100 for the RAM Preview tests if you were able to play them all at full res at full FPS (which isn't happening for at least several years with the test project we are using), but since you can technically export at a higher FPS than the project's FPS, you could get a Render score of 100, 200, or 1,000,000 in theory.

We typically see an overall score of about 1,000 on our 10920X tests. You GPU may be a small part of it, but you might also try setting the Windows power mode to "High Performance". We've had issues with some X-series platforms not giving full performance on the balanced power mode, and switching to high performance can sometimes give a pretty big boost to performance. Updating to the latest version of Windows 10 can also help a little bit.

Posted on 2020-01-24 18:26:42
Dan Hopkins

Yeah, I'm almost positive my power mode is on "High Performance" but i'll double check. I've checked for Windows updates and it looks like i'm all set there. Are you testing with the 10920x running stock, or OC'ed? In my bios, my AI Tweaker is set to "auto", which sets the base frequency at 3.8GHz, a middle frequence (I forgot what it's called) at 4GHz and a max turbo at 4.8GHz. When monitoring during the high CPU load parts of this benchmark (the C4D part), it looks like it's revving up to about 4.4GHz, so that's pretty good. I'll make sure my power mode is set to High Performance I suppose!

Posted on 2020-01-24 18:34:06

The latest version of Windows 10 is 1909, so there definitely is a newer version. I doubt that will get you more than a few percent, however, so it is more likely there is something else holding back your system. You might want to run something a bit more synthetic like Cinebench R20 to see how your single and multi-threaded performance is. That would answer whether it is some setting in After Effects that is the problem, or a more system-wide issue.

Posted on 2020-01-24 18:42:48
Dan Hopkins

Yeah, I ran that, it was around 6700 points. Maybe it has to do with disabling "disk cache" ? I also would be curious if CC19 is faster right now.

Posted on 2020-01-24 18:46:34

We tell you to disable disk cache because it can make the results inconsistent. Since AE writes to the disk cache whenever it thinks it can sneak it in during what it thinks are "idle" periods, it results in sometimes making no difference in the score, but other times making a big difference. If you did not disable the disk cache like the instructions say to, definitely do that and give it another shot.

Posted on 2020-01-24 18:56:52
Dan Hopkins

I did disable it as the instructions said to. I know it's much more important for real world tests to be fast, but It still makes me wonder what's going on. Considering my Cinebench R20 scored pretty much exactly where it should have, I'll just chalk it up to a weaker GPU, or just something in AE 2020. I'd be curious what would happen if I swapped an RTX 2080 in!

Posted on 2020-01-24 19:16:14

It says "version 8.2" instead of .82 in the update section near the top of the page.

Posted on 2020-02-05 20:08:04

Oops, thanks for the report!

Posted on 2020-02-05 20:11:43

You bet!

Posted on 2020-02-05 20:39:15
Grant R

Thanks for this article, it was great to see where my machine is at in terms of performance. Unfortunately I was disappointed when I received my final score. I realize that a lot goes into the final output but my components are fairly similar to the baseline. Full disclosure, I am very new to the PC world and it will take me awhile to understand how each component effects performance. That said I would have expected a higher score based on what is in my PC build. I would appreciate any help in this as I have done a bunch of searching and it seems like a case by base basis. I have included a breakdown of the main components Im using. I can provide more info on my system if needed but like I said before, I don't know exactly what info everyone needs. Thank you in advance!

After Effects CC 2020
OS: Microsoft Windows 10 Home
CPU: Intel i9-9900k @3.60Ghz
Mobo: PRIME z390-A
RAM: 64 GB (4x16GB) 2133MHz
GPU: Intel UHD Graphics 630, NVIDIA GeForce RTX 2080
MPE: Open CL

Posted on 2020-02-18 15:55:22
Grant R

For whatever reason it will not let me upload a picture of my score but it is as follows

Overal: 669
Render Score: 54.5
Preview Score: 56.3
Tracking Score: 90

Posted on 2020-02-18 15:56:34

It looks like you have an issue that is preventing AE from using the RTX 2080 in CUDA mode. The benchmark automatically tries to set MPE to CUDA if it can (or METAL on Mac), and the fact that it had to run in OpenCL means that AE doesn't see CUDA as an option. If you go into the project settings, is CUDA an option? Usually, that is just an issue with the GPU driver not being up to date, so you might want to update that and run it again.

Posted on 2020-02-18 18:03:29
Grant R

Hey Matt, I appreciate the fast reply. I will try this out later today and see what happens. Thank you!

Posted on 2020-02-18 19:50:17
Grant R

I tried that and CUDA was not an option. Although I did discover that my monitor is plugged into my motherboard which I'm finding out is a rookie mistake. In my control panel it says my NVIDIA card has an error 43. So I am assuming that AE is using the integrated graphics on the CPU, and I have an expensive GPU paperweight in my machine not doing anything. At the same time though the score I received seems a bit high for just using integrated graphics no? I'm going to switch my monitor to play from my GPU later today and see if that resolves my issues. Im curious of your thoughts and appreciate any feedback!

Posted on 2020-02-19 15:22:25

AE benefits from having a GPU, but it doesn't scale as much as you might expect. So your results seem about what I would expect with onboard graphics.

Hopefully you can get your GPU working with just a clean driver install. If that doesn't work, trying completely removing the NVIDIA driver with DDU cleaner ( https://www.guru3d.com/file... ) before installing the driver. Good luck!

Posted on 2020-02-19 17:20:14
Daniel Elfe

Hello
I guess there's now way to bench my AE CC 2018?
Thank you

Posted on 2020-02-20 11:14:01

Not at the moment. Once we hit full release and have everything working as we want it to, we may go back and add support for older versions. But there are sometimes features/effects we are testing that simply don't exist on older versions, so we can't always do that.

You can install newer versions of After Effects side-by-side with older versions, however. That way, you can install 2019/2020 in order to run the benchmark, without affecting any of your current projects.

Posted on 2020-02-20 18:09:37
Henrik Santana

https://uploads.disquscdn.c... Hey Mat,
I run the AE test, but got a error at final test C4D. Help please ;-)

Posted on 2020-02-21 18:54:48
Andrew Hoeveler

Thanks for this, but for some reason I am getting N/A on Overall, Render, and Preview scores. I have set everything including the spacebar preview change which isn't mentioned in the project instructions anymore, but IS mentioned above. I've tried running on both 2020 and 2019 AE, and also looking at the log it appears there might be some issue with javascript vs. extendscript, so I've gone into the project settings and switched the code type and ran it both ways... same thing. Here's attached the png screenshot, happy to share log files too.
https://uploads.disquscdn.c...

Posted on 2020-02-23 22:14:56

Looks like one of the C4D projects is performing so low that the FPS is rounding down to 0. That is likely causing the benchmark to assume that test broke and prevents it from generating the render/preview/overall scores.

It should be throwing an error if that is the case, though. I'll have to check the code and likely add in some additional code for the next update.

Posted on 2020-02-24 00:29:43

Yea, I'm pretty sure we just didn't code for the possibility of an actual 0 FPS result. I'll get that fixed for the next update, but I manually calculated the scores for you and it would be:

Overall Score: 621
Render Score: 61.7
Preview Score: 59.6
Tracking Score: 64.9

Posted on 2020-02-24 16:31:28
motionato

So I got this interesting comparison, I know that both processors are very old but when I looked at the result it made me more confused. First Rig is Dual Processor X5670 which is 2.93 turbo to 3.33 the other is Single E5-1620 V0 at 3.6 to 3.8 turbo. I expected a very big performance difference since Ae uses lower in # and faster cores better than many low speed cores. But the results were nearly the same with a higher score for the E5-1620 V0 any idea why this happened would be really helpful. https://uploads.disquscdn.c... https://uploads.disquscdn.c...

Posted on 2020-03-04 00:04:36

In some of the tests (the Phone Composite for example), the E5-1620 is significantly faster - around 4x for the half res preview. The scores ended up fairly even because of the NVIDIA GPU in the X5670 system (which helps many of the test, but especially the GPU stress project) and because the C4D tests are a bit more multi-threaded than the rest of After Effects.

You are also on different Windows versions and that can impact performance in some cases and background tasks that are running can affect the results as well. But I think the biggest thing is the NVIDIA GPU.

Posted on 2020-03-04 00:11:52
motionato

What about the cloud lighting project, results are nearly the same. Dose that mean a faster processor shouldn't always behaves fast? or there is another reason that explain this project result?

Posted on 2020-03-04 09:34:05

It is really impossible to say exactly without swapping hardware around. But my guess is that the better GPU in the X5670 system is offsetting the lower performance of that CPU, and it just happens to be exactly the right amount to result in almost identical performance from both systems.

Posted on 2020-03-04 17:06:03
CakeInMotion

Thanks for all the help.

Posted on 2020-03-04 17:16:37
Suraj

Hey Matt
I'm going to Build My new Pc for After effects.

Processor - Ryzen 5 3600
Motherboard - I'm Confused Which Chip set would be better B450 Or X570 Please Suggest me For That

Posted on 2020-03-28 05:21:04
Torjus Aanderaa

Im trying to run the Bench on my old computer for comparison, but run in to the "After Effects has encountered a failure to GPU-enabled effects on this frame. This is likely because the GPU is out of memory"

whether i press ignore or "use mercury software only" the script stops running.

This machine has an old Geforce GTX 770 with 2 GB VRAM. Is that simply not enough for this benchmark or is there another workaround ?

Posted on 2020-03-31 14:54:19

Yea, a 2GB GPU is right at the minimum specs for After Effects, and since some of the tests are a bit heavy on the GPU, 2GB just isn't going to be enough. You could try updating to the latest NVIDIA Studio driver, but more than likely there isn't going to be much you can do.

Posted on 2020-03-31 16:27:21
Basem Brimo

https://uploads.disquscdn.c...
Hello
I'm having this problem
What should I do ?

Posted on 2020-04-08 22:48:49

Go through everything in the troubleshooting section: https://www.pugetsystems.co... . If you are still getting that error, you might want to reinstall After Effects as well.

Posted on 2020-04-08 22:50:01
Aperiso

Hi, I just tried running the benchmark for AE and after the cellphone render/export test it goes to load the polygon comp, but i get an error saying "unable to find polygon composition". It's strange, because if i open the benchmark project, it's all there, but after the cellphone render/export all the comps disappear completely from the project window.

Posted on 2020-04-13 09:04:54

Are you running that latest 0.9 version of the benchmark? It almost sounds like the entire project is closing after the "Phone Composite Project" tests run, but I don't know how that could be possible. You might want to clear your preferences in case there is some weird conflict with something like a custom keyboard shortcut. You can use our preference backup/clear utility if you want to restore your preferences after you run the benchmark: http://puget.systems/go/Pug...

Posted on 2020-04-13 16:36:22
stooovie

Hi, Im unable to complete the test. One or another comp gives me the "Error: Only X frames played out of total...". I'm on 2015 5k iMac with 24 GB RAM and nothing else running. Everything fully updated and running on a brand new user account. Lease advise.

Also, the benchmark seems to have hard-coded the startup Volume. I had to rename it to Macintosh HD for the test to begin.

Posted on 2020-04-19 21:20:43

If AE is not able to render all the frames, you don't have enough RAM available. If it isn't giving you a warning when you first start the benchmark, your AE settings should be fine, but you may have too many other programs running that are eating up your RAM. Try closing down as much as you can, or turning up a bit the amount of RAM that is reserved for AE (Preferences -> Memory).

Can you clarify what you mean by "hard-coded the startup Volume"? I haven't heard of anyone having an issue like that and that also isn't something we have encountered ourselves when running the benchmark on the Macs we have.

Posted on 2020-04-23 22:55:49
stooovie

I ran it on a 24 GB iMac with Catalina, new user account with nothing else running. AE is alotted 21 GBs out of those 24.

As for hard-coded volume, the benchmark simply doesn't begin if the startup volume is not named "Macintosh HD". I like to name it simply "Mac" to not have to deal with the space in the filename. The error is: "Error: Could not get system information for Motherboard: //Volumes//Macintosh HD/usr/sbin/system_profiler SPHardwareDataType. Continuing the benchmark, but will not be able to accurately log system specs."

When I rename the volume to the default "Macintosh HD", the benchmark starts.

Posted on 2020-04-24 09:30:07

Ah, I see, that is it trying to get all the system information. I didn't think about that issue, but I'll add it to the list of things to fix in the next update. Thanks for the heads up!

Posted on 2020-04-30 03:31:10
Gitt Phuckt

872
Metabox i9 9900
RTX 2070
64gb RAM
2 x Intel 660P 2TB SSD
Wouldhave liked to have cracked the 900 but im happy with the result

Posted on 2020-05-02 12:56:02
Pedro Marques

Hello, the following error appears:
https://uploads.disquscdn.c...

Any idea what I'm doing wrong?
The configuration is correct. I use a Mac Pro 5.1, macOS 10.13.6

Posted on 2020-05-30 18:11:14
Daliovic Dali

I am not being able to even start the benchmark on my 2015 15" MBP
https://uploads.disquscdn.c...
As soon as the benchmarks starts I get 3 error popups telling me
Error: Could not get system information for System: //Volumes//Macintosh HD/usr/sbin/system_profiler SPHardwareDataType . Continuing the benchmark, but will not be able to accurately log system specs
It says the same for 2 different files : system_profiler (Twice, for 2 commands) and sw_vers.
https://uploads.disquscdn.c...
After I click on OK on each of the popups nothing happens.
I made sure to give all of the privacy permissions to anything related to After Effects.
What exactly am I doing wrong?

Posted on 2020-06-10 11:27:18

You probably have your drive renamed from the default. We have to use an absolute path for the system information commands, and I haven't figured out yet how to get the primary drive name so that we can make it handle renamed drives on MacOS. It is a quirk of running commands from within Adobe applications that we need to work around.

Posted on 2020-06-10 17:59:43
DàLy DàLy

Thank you Matt, I will try renaming the drive and see if it will work.

Posted on 2020-06-10 18:29:19
Hồ Minh Nhật

I'm trying to investigating why my tracking score is so low .
https://www.pugetsystems.co...

Intel(R) Core(TM) i9-9900K CPU @ 3.60GHz
NVIDIA GeForce RTX 2080 Ti
64GB (4x16GB) 3000MHz

Posted on 2020-07-30 07:34:58
Kcls

Hi! (I posted this originally on the wrong test page) I'm trying to run this on my Macbook pro to see if my new or old macbook is faster with After Effects (old laptop has a higher clock speed, but less ram, new laptop has a lower clock speed but higher RAM.) I run into an error on my old computer.

Laptop Specs:
Macbook Pro 15in 2017
3.1 GHz Quad-Core Intel Core i7
16 GB 2133 MHz LPDDR3
Radeon Pro 555 2 GB

On this computer: The test plays fine, until it hits the Cloud Lightning Full Res test, and it errors out every time. The error says: "Error: Only 1 frames played out of the total 100 for Cloud Lightning Project - Full Res" I have 13gb of RAM allocated to after effects, the most it allows me to allocate on this computer. I can confirm that if I click out of the benchmark, and attempt to ram preview the Cloud Lightning file at full res, it plays one frame and then the preview stops. I can manually scrub the timeline to render each frame, and then if I ram preview again, it'll play the frames I've rendered, but once it hits an unrendered frame it stops again. I called into your tech support and they recommended I post my logs here for the test. I see the logs (there are 6 total files), but they all only mention the polygon project and the phone composite project, nothing about the cloud lightning.

Any ideas on how to get this to run beyond the full res cloud lightning test?

Posted on 2020-09-03 16:04:22
Zoe Earnshaw

Hi, Just trying to run the AE benchmark on brand new Macbook Pro 16, 64GB ram, updated OS, newly installed AE. Followed all the setup instructions but get error message after the Phone Composite Project first test ends with 'Error: Only 99 frames played out of the total 100 for phone composite project- half res' and benchmark stops...can you help?

Posted on 2020-09-21 14:20:56

It looks like the latest update to AE changed something. Probably something like they switched from frames being 0-indexed to 1-indexed on the API side for some reason. I'm working on an update now, but a quick work-around would be to roll back to AE 17.1.3

Edit: New version is uploaded that should resolve the issue.

Posted on 2020-09-21 16:57:59
Zoe Earnshaw

Tried it again and worked fine- many thanks for the speedy solution :)

Posted on 2020-09-21 21:32:28
David Rembi

Hey guys,

I used this on my last set-up and it worked great. now i'm trying to run it, and in the first test it keeps getting stuck telling me that it only rendered 99 frames... i tried uninstalling, downloading the files again, nothing seems to work.. any thoughts? (Im working with the latest AE version)

Posted on 2020-09-21 15:16:40

Looks like the latest update to AE changed something. Probably something like they switched from frames being 0-indexed to 1-indexed on the API side for some reason. I'm working on an update now, but a quick work-around would be to roll back to AE 17.1.3

Edit: New version is uploaded that should resolve the issue.

Posted on 2020-09-21 16:57:41