Benchmarking results continued

JN- wrote on 1/18/2020, 12:56 PM

Continued from the now locked “Benchmarking” thread ...

Moderator's Note August 2021:

Project sample for this benchmarking project can be downloaded here:

https://www.dropbox.com/s/3i4xgnhl9fmc0je/BenchmarkProject4K.zip?dl=0

(Project works in Vegas Pro 16 and newer).

The participant's benchmark entry has been changed via Google Form:

Submit your result here : https://forms.gle/2gnzmv93Ar8P1tkx8

And view the result here: https://docs.google.com/spreadsheets/d/1b3ggVifKsuT-cp2kQHjum_KnQ4-2jBmUIvzmu7BQZ34/

Thank you user @RogerS for continuing mantaining this benchmark system.

Please message him if you have questions or problems regarding the render benchmark.

----------------------

Note that the render time results are for the full project duration, the average FPS value is for the “Region 1” duration only. Please enter only the same "Average FPS" value for all of your different encoder and frame size entries.

Don't change the Project properties. Whether checking your FPS/rendering UHD or FHD, leave as is.

Templates and user data input with examples … Use similar for Nvenc, QSV, VCE, and CPU encoders.

The data rates of the 2 templates below is 24/12 for FHD and 50/28 for UHD. So irrespective of which encoder you're using, just use those data rates..

Set the Intel QSV quality slider to 4, default. If this range of 1 to 7 changes in the future leave it at whatever the default is.

My rendered out file sizes on disc. You should get somewhat similar, but not necessarily the exact same results.

FHD size = 48.9 mb ..... Data rate = 9,583 kbps ... Nvenc encoding

FHD size = 50.8 mb ..... Data rate = 9,971 kbps ... Cpu encoding

UHD size = 114 mb ..... Data rate = 22,611 kbps ... Nvenc encoding

UHD size = 117 mb ...... Data rate = 23,308 kbps ... Cpu encoding

If your output file sizes and data rates are not similar, then check that you are using the correct render template settings.

Sample User Data Input with example, text file is in the ZIP file …

Use the instructions in the screenshots below to get the Average FPS value …

If possible set the Dynamic ram Preview to zero while doing so.

Note that the region 1 section above is only used to get the average FPS value, do render out the full length project to get the various render times.

The older "Benchmarking" thread.

Back to post

Comments

RogerS wrote on 12/29/2020, 7:54 AM

Great, thanks for updating it. I think frameserving to x.264 is more recent than the benchmarking. Maybe it's still too niche for most users but I see some of the most experienced ones here using it. Thanks for keeping this list going.

JN- wrote on 12/29/2020, 7:58 AM

@RogerS “Maybe it's still too niche for most users” I agree. This started in July 2019 in the older, now locked Benchmarking thread.

MP- wrote on 1/26/2021, 12:09 PM

User name       :Miltos Pavlou
   Machine           :OB
   VP version       :18
   VP Build number       :434
   CPU           :i9-10980XE
   Cores           :18
   GPU           :AMD RX-6800XT
   iGPU           :none
   OC           :No
   Render frame size   :UHD
   HW Acc.           :AMD
   Render As       :Magix AVC
   Encode mode       :VCE
   Render time       :01:00s
   FPS           : 17.5
   Comments       : MB AUS X299 Prime A-II

Source, scratch and rendering delivery disks are 3 Samsung 970 Plus, OS Disk is Samsung Evo 840 SSD

XFX Radeon 6800XT recognised by VP as gfx 1030 (at Preferences>Video> tab)

i/O sees the AMD UVD/VCN in the HW decoder box.

CPU usage remains too low (5-10%, even with highest priority, unless some peaks in the end of rendering up to 80%). no other programmes running to occupy CPU

Until 25% of rendering this file, fps=1.69 after this it accelerates to reach 6.76 average fps

Rendering time with Mainconcept AVC instead of VCE = 3 min. With a strange CPU Usage pattern of 5% plateaus and 20% spikes. fps always around 6fps

CPU - Graphic card are high end but rendering times are too slow even compared to much weaker CPUs.

Whole system has good benchmarks: Passmark = 6521.4 /CPU Mark=33069, CinebenchR15 = 3640

and I reinstalled clean Vegas 18 to check. This happened last week. Previously shorter rendering times were achieved.

Your input, ideas are very welcome!

@JN @therhino

JN- wrote on 1/26/2021, 1:36 PM

@MP- Thank you for that.

#1. "Until 25% of rendering this file, fps=1.69 after this it accelerates to reach 6.76 average fps"

Is the 6.76 fps value you give for the small region 1? Or is it for all of your render? Unfortunately, I know that it's not a very exact methology, but the average FPS, best eyeball estimate, of region 1 is required.

#2. Do you want me to post two results, the 2nd one with all of the same items as the 1st. one except these 2 ... Mainconcept AVC instead of VCE = 3 min. That is Render As Formats = Magix Avc, and Encode Mode = CPU and time of 3 min. If not, just post the complete 2nd set of values, if you wish.

MP- wrote on 1/26/2021, 1:59 PM

@MP- Thank you for that.

#1. "Until 25% of rendering this file, fps=1.69 after this it accelerates to reach 6.76 average fps"

Is the 6.76 fps value you give for the small region 1? Or is it for all of your render? Unfortunately, I know that it's not a very exact methology, but the average FPS, best eyeball estimate, of region 1 is required.

#2. Do you want me to post two results, the 2nd one with all of the same items as the 1st. one except these 2 ... Mainconcept AVC instead of VCE = 3 min. That is Render As Formats = Magix Avc, and Encode Mode = CPU and time of 3 min. If not, just post the complete 2nd set of values, if you wish.

#1 6.76fps is the average of the entire rendering. The 1st slow bit reaches about 1.6-2fps.

#2 please use the detailed VCE result. I will post separately the mainconcept AVC. Its CPU usage pattern is unique and strange. Will post a screenshot.

Overall, I was alarmed since I had to render for 25hours a 2h 4k video I had previously rendered in 8hours with my older CPU i7 5820.

I upgraded to this system exactly to shorten rendering times and instead things got worse, while the system is excellent. For example it flies in Ultra res the MS flight simulator 2020..

Thanks again

JN- wrote on 1/26/2021, 2:10 PM

@MP- #1 6.76fps is the average of the entire rendering. The 1st slow bit reaches about 1.6-2fps.

Ok, so what’s required is the average FPS for region 1 only, as marked out in the screenshot in my very 1st. post in this thread. You can update your 1st. post here with those values when you are also posting your 2nd. set.

MP- wrote on 1/27/2021, 1:27 AM

@MP- #1 6.76fps is the average of the entire rendering. The 1st slow bit reaches about 1.6-2fps.

Ok, so what’s required is the average FPS for region 1 only, as marked out in the screenshot in my very 1st. post in this thread. You can update your 1st. post here with those values when you are also posting your 2nd. set.

@JN Region 1 only, is 8.226 fps - updated above.

My RX6800XT video card at Preferences>Video> tab is recognized as optimal - AMD, (gfx 1030)

MP- wrote on 1/27/2021, 1:38 AM

User name       :Miltos Pavlou
   Machine           :OB
   VP version       :18
   VP Build number       :434
   CPU           :i9-10980XE
   Cores           :18
   GPU           :AMD RX-6800XT
   iGPU           :none
   OC           :No
   Render frame size   :UHD
   HW Acc.           :AMD
   Render As       :Magix AVC
   Encode mode       :CPU
   Render time       :01:37s
   FPS           : 17.5 fps
   Comments       : ASUS X299 Prime A-II

Source, scratch and rendering delivery disks are 3 Samsung 970 Plus, OS Disk is Samsung Evo 840 SSD

@JN @therhino

RogerS wrote on 1/27/2021, 1:39 AM

A search indicates GFX 1030 seems correct for your card. It's also known as Navi 21. Link

JN- wrote on 1/27/2021, 2:28 AM

@MP- Ok, thank you for the extra UHD set of render times. Do you wish to also include two sets for FHD, i.e. a 3rd. FHD using VCE and a 4th. FHD using Magix CPU? If so i’ll wait until you post same.

RE: the average FPS PLAYBACK value for Region 1. I can only ever input a single same value for this, for all of your testing. So will I enter 8.226 or 15.5, which is the avg. of 24 and 7. Thanks.

The recommended render template for FHD is also available in my 1st. post in this thread.

MP- wrote on 1/27/2021, 2:55 AM

@MP- Ok, thank you for the extra UHD set of render times. Do you wish to also include two sets for FHD, i.e. a 3rd. FHD using VCE and a 4th. FHD using Magix CPU? If so i’ll wait until you post same.

RE: the average FPS PLAYBACK value for Region 1. I can only ever input a single same value for this, for all of your testing. So will I enter 8.226 or 15.5, which is the avg. of 24 and 7. Thanks.

The recommended render template for FHD is also available in my 1st. post in this thread.

In the mainconcept AVC render, the fps varied between 7 and 24, but 24 were the momentaneous peaks.

Most of the time the fps were juggling around 7.8.

Here are the FHD data:

User name       :Miltos Pavlou
   Machine           :OB
   VP version       :18
   VP Build number       :434
   CPU           :i9-10980XE
   Cores           :18
   GPU           :AMD RX-6800XT
   iGPU           :none
   OC           :No
   Render frame size   :FHD
   HW Acc.           :AMD
   Render As       :Magix AVC
   Encode mode       :VCE
   Render time       :00:30s
   FPS           : 17.5
   Comments       :

Source, scratch and rendering delivery disks are 3 Samsung 970 Plus, OS Disk is Samsung Evo 840 SSD

avg fps during rendering: 17.26fps

MP- wrote on 1/27/2021, 2:56 AM

User name       :Miltos Pavlou
   Machine           :OB
   VP version       :18
   VP Build number       :434
   CPU           :i9-10980XE
   Cores           :18
   GPU           :AMD RX-6800XT
   iGPU           :none
   OC           :No
   Render frame size   :FHD
   HW Acc.           :AMD
   Render As       :Magix AVC
   Encode mode       :CPU
   Render time       :00:35s
   FPS           : 17.5
   Comments       : ASUS X299 Prime A-II

Source, scratch and rendering delivery disks are 3 Samsung 970 Plus, OS Disk is Samsung Evo 840 SSD

JN- wrote on 1/27/2021, 4:19 AM

@MP- I really don’t know. Of course this is only a single benchmarking test project. So the media in use is only a small portion of what users in everyday use would use VP for, a small snapshot as it were. So for example your 16 core cpu will of course be excellent in other programs, besides VP, in multitasking etc. VP may not be fully utilising all 16 cores? Whether other NLE's do or not, i’m not sure.

Before I post all 4 of your results can you give me one single average fps value to enter for all 4, as I only enter one single AVG. PLAYBACK value, the same for all 4, no matter how many render tests that you do.

Other users have also at times, entered multiple fps values. The way I look at it is this. Irrespective of what type of render test that is done, fhd, uhd, different encoding methods, the project properties for the project are fixed. So for region 1 there can only be one single fps playback value.

MP- wrote on 1/27/2021, 4:29 AM

@MP- I really don’t know. Of course this is only a single benchmarking test project. So the media in use is only a small portion of what users in everyday use would use VP for, a small snapshot as it were. So for example your 16 core cpu will of course be excellent in other programs, besides VP, in multitasking etc. VP may not be fully utilising all 16 cores? Whether other NLE's do or not, i’m not sure.

Before I post all 4 of your results can you give me one single average fps value to enter for all 4, as I only enter one single AVG. PLAYBACK value, the same for all 4, no matter how many render tests that you do.

Other users have also at times, entered multiple fps values. The way I look at it is this. Irrespective of what type of render test that is done, fhd, uhd, different encoding methods, the project properties for the project are fixed. So for region 1 there can only be one single fps playback value.

8.226 fps playback should cover all

Opened a support ticket with Magix, in case they can help. It really does not make sense that a most advanced CPU is under used.

Ran also Prime95 and it seems the CPU works fine and fully.

Given that VP does use 100% of the CPU int the first seconds and the last 75% of rendering, it looks to have to do with the VP rendering process itself and integration of the CPU i9 10980XE skylake chip in this process

Thanks again

JN- wrote on 1/27/2021, 4:31 AM

@MP- Aok, i’ll post your results later, and notify you here when done.

JN- wrote on 1/27/2021, 7:38 AM

@MP- I’m really no expert at reading those graphs, you can probably make more from them than I can. There might be a broader response to this issue if you opened a new discussion thread on it? I tend to just post the test results here in this Benchmarking Results Thread.

MP- wrote on 1/27/2021, 7:40 AM

@MP- I’m really no expert at reading those graphs, you can probably make more from them than I can. There might be a broader response to this issue if you opened a new discussion thread on it? I tend to just post the test results here in this Benchmarking Results Thread.

ok, thanks I deleted them. There is a separate thread: https://www.vegascreativesoftware.info/us/forum/low-cpu-usage-20-rendering-at-highest-priority--126837/?page=1

thanks again

JN- wrote on 1/27/2021, 8:55 AM

@MP- Your results are now updated, available at the screenshots in the 1st. post in this thread and also in the download zip, spreadsheets and screenshots.

MP- wrote on 1/28/2021, 2:24 AM

@MP- Your results are now updated, available at the screenshots in the 1st. post in this thread and also in the download zip, spreadsheets and screenshots.

Thank you.

I tried also separate parts rendering of the benchmarking project and I realised that in different sections of rendering the CPU usage is very different - corresponding to the different stages of the CPU usage pattern (pics I uploaded showing a distinctive patter). However, I don't see any particular differencies in the project FXs applied.

03.18 (Work fast appears) to 07.15 (NEW, Dynamic... apepars) --> CPU usage drops to 5-7% with spikes pattern)

07.15 - end --> CPU goes to 86% until the end.

No matter at which point I start the rendering I get this CPU usage in different parts of the project.

Can this be an indication of specific issues with rendering elements? I cannot see very different FX applied.

In fact, more FX are applied in the 2nd very fast rendering part.

In case you know much better than me the benchmarking project you can flag to me any particular differencies between different parts of the clip.

Many thanks again

JN- wrote on 1/28/2021, 3:45 AM

@MP- “In case you know much better than me the benchmarking project you can flag to me any particular differencies between different parts of the clip.”

Not really. The project is Magix's own FHD upscaled to UHD by another forum user JV. The only standout item I saw was how very little difference there was between your VCE HW encoding and CPU encoding, in FHD, a mere 4s improvement.

I would have expected the VCE HW encoding to be much faster. If you have done everything correctly, then it is what it is.

MP- wrote on 1/30/2021, 1:08 PM

@JN

I managed to considerablly improve rendering times and updated my relevant posts above with the new timings.

what I did:

- flashed a new BIOS on my ASUS X299 Prime A-II

- updated Windows (Windows 10 Insider Preview 21301.1000 (rs_prerelease)

No clue what the problem was. I suspect the motherboard and i was thinking of moving to a better one anyway

JN- wrote on 1/30/2021, 6:46 PM

@MP-

#1 FHD AMD Magix Avc VCE ... 00m : 30s

#2 FHD AMD Magix Avc CPU ... 00m : 35s

#3 UHD AMD Magix Avc VCE ... 01m : 00s

#4 UHD AMD Magix Avc CPU ... 01m : 37s

2# for a CPU only encode appears to be especially good, indeed very close to the #1 VCE test.

There is no other user with similar CPU and GPU to compare to yet, so I have to assume that your results are exceptionally good, well done, simply excellent rig.

OK Miltos, can you confirm for me that these are the final changes you wish me to update to? There is quite a bit of work involved in updating/adding in results, ideally best to get it right this time.

I remember one user previously entered really fantastic results but he subsequently realised that he was only rendering out the loop region section, not the full project.

Hopefully that's not the case here, if so, going forward, that combination of your HW is something to aim for.

RogerS wrote on 1/30/2021, 8:49 PM

Those numbers are stunning. Glad Vegas can take advantage of high-end hardware after all!

JN- wrote on 1/31/2021, 4:17 AM

Indeed, just need @MP- to confirm first.

1 … 5 6 7 … 15

of 15