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

JN- wrote on 3/10/2020, 7:01 AM

@NickHope All updated now. I added your extra comments … (From above tests …) as an inserted comment in the spreadsheets, now in the ZIP.

In name order ...

Reyfox wrote on 3/10/2020, 8:20 AM

@JN- thanks for the hard work in putting this together.

A little off topic, but I found this video curious.

JN- wrote on 3/10/2020, 9:02 AM

@Reyfox Just finished looking at it, excellent, he put in a lot of work.

NickHope wrote on 3/10/2020, 12:31 PM

...thanks for the hard work in putting this together.

A little off topic, but I found this video curious.

The VEGAS social media team should be all over this 👆 Vegas rendering fastest out of all the competition @Bri-Hageman

Gary-Vrckovnik wrote on 5/13/2020, 8:43 PM

Hello … I'm a new VP17 user, still learning my way around as I've not used any video editor software for many years (my last version was Sony Vegas 8!!) ... I've got about 200 hours of family video (mostly MiniDV and 8mm tapes) that I'm finally organizing / editing. There are also some higher res videos from newer digital cameras (avchd files).

I'm considering upgrading my computer (i7-2600k processor, 8GB, Radeon HD6850 1GB video card) to enable faster rendering, smoother previews, etc. Currently I find that I can render the avi files to mp4 in roughly real time (with very few effects / transitions etc added).

I've just come across this 4K Sample benchmark and am very keen to use it to see how my 8 year old system compares to modern setups before making the investment in new hardware. But as a new user, I've got a few questions to ensure that I am using the benchmark properly ..

1. Is it truly as easy as unzipping the Benchmarking.zip file onto my SSD, loading the SampleProject 4K.veg file into VP17 and then doing Render As using the two templates illustrated in the FHD and UHD jpg files? After which I just note the two render times?

2. Where does one find the average FPS numbers quoted in column S of the results spreadsheets? I don't see a FPS number listed anywhere on my Vegas screen.

3. Where can I find the 4K Red Car benchmark that I have seen other threads in this forum reference? It might be good for me to run that one as well to gain additional comparison data. Or, has that test been superceded by this 4K Sample benchmark and is not worth running?

4. Finally, for my current project I don't plan to render to 4k. Currently I can render avi files to mp4 in roughly real time with minimal effects in the project. How much faster would the modern systems render this simpler content and output? Could I expect to see a 3-4x speedup in my render times for non 4k video compared with my current setup?

A HUGE thanks in advance to all for your help and suggestions...

NickHope wrote on 5/14/2020, 12:33 AM

@Gary-Vrckovnik

I'll answer a couple of them...

2. Where does one find the average FPS numbers quoted in column S of the results spreadsheets? I don't see a FPS number listed anywhere on my Vegas screen.

3. Where can I find the 4K Red Car benchmark that I have seen other threads in this forum reference? It might be good for me to run that one as well to gain additional comparison data. Or, has that test been superceded by this 4K Sample benchmark and is not worth running?

It can't do any harm, if you can find meaningful results to compare against. Web archive has it. It might be very slow to download, as in leave-it-overnight slow: https://web.archive.org/web/20131113061953/http://download.sonycreativesoftware.com/whitepapers/vp11_benchmark.zip

The formats you talk about are not very demanding in comparison to some others. You might find you don't need to upgrade. Work through this post to give your current system the best chance: https://www.vegascreativesoftware.info/us/forum/faq-how-can-i-make-my-video-preview-play-smoothly-in-vegas-pro--104624/

Howard-Vigorita wrote on 5/14/2020, 9:27 AM

@Gary-Vrckovnik Original Red Car and my 4K versions are on a google drive here. Thread on my 4K Redcar and comparative tables with Sample Project is at...

https://www.vegascreativesoftware.info/us/forum/4k-benchmarking-roundup--116449/

JN- wrote on 5/14/2020, 11:33 AM

I’ve now updated the screenshot on the home page of this post to clarify where to find the FPS readout for anyone else coming along.

@Gary-Vrckovnik I’ll update your results when posted here. Best of luck with your testing.

Gary-Vrckovnik wrote on 5/17/2020, 6:13 PM

What template should I use for my rendering using Benchmarking.zip? I have an old Radeon video card (HD6850) so I can't use the NVIDIA NVENC templates that are suggested in the Benchmarking.zip file / shown on the home page of this post.

The two quick tests I've done so far used:

MAGIX HEVC/AAC MP4 … with Internet HD 1080p 25 fps (Intel QSV) and resulted in a Render Time of 8:00

MAGIX AVC/AAC MP4 … with Internet HD 1080p 25 fps and resulted in a Render Time of 3:04

Please let me know which Render template you'd suggest for as accurate a comparison with modern system results as possible.

Thanks again!

Gary

JN- wrote on 5/17/2020, 7:39 PM

@Gary-Vrckovnik Hi Gary, just use the same templates, don't change anything except the "encode" mode. Change it to Mainconcept, that'll just use your CPU. For QSV and CPU encoding use the templates data rates of 24/12 for FHD or 50/28 for UHD.

Gary-Vrckovnik wrote on 5/17/2020, 10:23 PM

Many thanks JN. I believe I've got things working properly now and here are my results for the Benchmarking.zip test using the System I built back in Jan 2012... (the values might be useful for others thinking of upgrading old hw) ...

User name  :GV
Machine   :OB
VP version  :17
VP Build number  :421
CPU   :i7-2600K CPU @ 3.40GHz
Cores   :8
GPU   :Radeon HD6850
iGPU   :-
OC   :No
Render frame size :FHD
HW Acc.   :
Render As  :Magix Avc
Encode mode  :Mainconcept AVC
Render time  :06:25s
FPS   :0.8
Comments  :System Built in 2012 with 8GB RAM

User name  :GV
Machine   :OB
VP version  :17
VP Build number  :421
CPU   :i7-2600K CPU @ 3.40GHz
Cores   :8
GPU   :Radeon HD6850
iGPU   :-
OC   :No
Render frame size :UHD
HW Acc.   :
Render As  :Magix Avc
Encode mode  :Mainconcept AVC
Render time  :15:00s
FPS   :0.8
Comments  :System Built in 2012 with 8GB RAM

Looking at your results spreadsheet, it seems like a substantial performance increase will be obtainable by moving to modern hardware - not much of a surprise of course! I'm currently leaning towards a Ryzen 7 or 9 family CPU. Not sure about the GPU yet.

In your spreadsheet, UserName fifonik lists a Ryzen 5 3700X processor - to my knowledge the Ryzen 5 family tops out at a 3600X processor not a 3700X. Wherease the Ryzen 7 family has a 3700X. Any idea if the CPU is actually a Ryzen 7?

JN- wrote on 5/18/2020, 6:06 AM

@Gary-Vrckovnik Hi Gary, can you complete your user data input fields for iGPU and HW Acceleration.

The HW Acc. options are None/Off, Nvidia, Intel, AMD. See Options\Preferences\Video tab\ GPU Acceleration of Video processing

iGPU if N/A just enter None.

JN- wrote on 5/18/2020, 6:22 AM

@Gary-Vrckovnik Looking at the Intel specs for the i7-2600K it was a 4 core, not an 8 core CPU, it uses the HD 3000 iGPU, so I just now need your HW Acc. value.

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

FYI The 2013 Benchmarking project also had spme i7-2600K CPU's.

JN- wrote on 5/18/2020, 7:09 AM

@fifonik Can you clarify this, thanks, then i’ll update the results here accordingly.

Gary-Vrckohnik ...”In your spreadsheet, UserName fifonik lists a Ryzen 5 3700X processor - to my knowledge the Ryzen 5 family tops out at a 3600X processor not a 3700X. Wherease the Ryzen 7 family has a 3700X. Any idea if the CPU is actually a Ryzen 7”

Gary-Vrckovnik wrote on 5/18/2020, 8:20 AM

My corrected results JN ...

User name  :GV
Machine   :OB
VP version  :17
VP Build number  :421
CPU   :i7-2600K CPU @ 3.40GHz
Cores   :4
GPU   :Radeon HD6850
iGPU   :HD 3000
OC   :No
Render frame size :FHD
HW Acc.   :AMD
Render As  :Magix Avc
Encode mode  :Mainconcept AVC
Render time  :06:45s
FPS   :0.8
Comments  :System Built in 2012 with 8GB RAM, Render Time an average of 4 runs (low 5:58, hi 7:17)

User name  :GV
Machine   :OB
VP version  :17
VP Build number  :421
CPU   :i7-2600K CPU @ 3.40GHz
Cores   :4
GPU   :Radeon HD6850
iGPU   :HD 3000
OC   :No
Render frame size :UHD
HW Acc.   :AMD
Render As  :Magix Avc
Encode mode  :Mainconcept AVC
Render time  :15:11s
FPS   :0.8
Comments  :System Built in 2012 with 8GB RAM, Render Time an average of 2 runs (15:00 and 15:23)

Interesting to watch the Task Manager CPU%... on the last UHD test I ran, it peaked at 97.3%, but most of the time it was bouncing around in the 30-50% range.

JN- wrote on 5/18/2020, 9:51 AM

@Gary-Vrckovnik Your results are now updated, thanks.

Gary-Vrckovnik wrote on 5/18/2020, 9:57 AM

Any guesses how much faster than real-time a simple render of MiniDV avi file footage to mp4 would be with modern processors? I can't find any references as everyone has moved on to higher source resolutions. My current setup does these in effectively real time.

JN- wrote on 5/18/2020, 10:12 AM

@Gary-Vrckovnik Hi Gary, if no one else chips in I can give you that later, unfortunately my PC is currently tied up doing a very large file copy that will take several hours. I have plenty of DV25 footage and can give you a figure easily.

I can do a CPU and an Nvenc render to .mp4 at the same PAL resolution as source. i.e. 720 x 576.

Gary-Vrckovnik wrote on 5/18/2020, 10:36 AM

That would be hugely appreciated JN - I'm 95% decided it's time to upgrade my system - that info will give me the final nudge!!

john_dennis wrote on 5/18/2020, 11:44 AM

@Gary-Vrckovnik

I have some tapes from the 1980s that I captured to DV. I rendered one while I'm waiting for a P2V VM conversion to run. On the system from my signature, which is due for upgrade in Q1 2021, I can encode video like yours in less than half of real time with AMD VCE.

When using GPU encoders, the work moves to the GPU and the CPU loafs. There just isn't enough going on with this source to demand more CPU. If you turn off GPU things change.
You won't get any benefit from GPU decode with this source media.
If you use GPU assisted fX, you might get higher utilization.

"For this use case, I'd buy the GPU first if you plan to use GPU encoders." he said knowing that he will upgrade every four years whether he needs to or not.

JN- wrote on 5/18/2020, 1:16 PM

@Gary-Vrckovnik Hi Gary. I used a 60 second piece of DV25 720 x 576 25 fps PAL source.

I rendered out to an MP4 file, same resolution

#1 Magix Nvenc render completed it in 11s … Data rate 7,454 kbps = ~ 5.5 times faster than real time

#2 Magix CPU render completed it in 14s … Data rate 8,011 kbps = ~ 4.3 times faster than real time

#3 Happy Otter Render+ Nvenc render completed it in 10.2s … Data rate 8,348 kbps = ~ 5.9 times faster than real time.

I used this #3 because unlike Magix Nvenc you can bump up the data rate somewhat, Magix Nvenc is limited to ~ 7,454 kbps, but it's probably perfectly adequate for this type of media.

Best of luck with your new build.

Gary-Vrckovnik wrote on 5/18/2020, 2:11 PM

Huge thanks to both of you (JN, John Dennis) for your metrics. Build time it is!!

fifonik wrote on 5/18/2020, 11:59 PM

Fresh results from my system

Changes from my previous report:

GPU: RX 470 => RX 580
GPU Drivers: 19.9.2 => 20.2.2 
Source Drive: Samsung 850 Evo 256MB SSD => Samsung 970 Evo+ 500MB NVMe
Video | Maximum number of rendering threads: 8 => 16

When I got results I noticed quite big difference in preview framerate from my previous report(5-10 => 5-6)

After investigation I've found that internal option 'Enable so4 Component Reader for AVC/M2TS' affects the preview speed quite a bit (looks like this was already reported in this thread by @NickHope).

I'm usually working with the option set to NON-default value (FALSE) as with default value (TRUE) VP does not decode correctly my AC3 5.1 footage (known bug, based on Magix reply they will not fix it).

My previous report was also with the option set to FALSE.

Results with "Internal | Enable so4 Component Reader for AVC/M2TS" = TRUE (default)

    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :FHD
    HW Acc.            :AMD
    Render As        :Magix Avc
    Encode mode        :VCE
    Render time        :00:53s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :UHD
    HW Acc.            :AMD
    Render As        :Magix Avc
    Encode mode        :VCE
    Render time        :01:51s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :FHD
    HW Acc.            :AMD
    Render As        :Magix Hevc
    Encode mode        :VCE
    Render time        :00:53s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :UHD
    HW Acc.            :AMD
    Render As        :Magix Hevc
    Encode mode        :VCE
    Render time        :01:48s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :FHD
    HW Acc.            :AMD
    Render As        :Magix Avc
    Encode mode        :CPU
    Render time        :00:58s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :UHD
    HW Acc.            :AMD
    Render As        :Magix Avc
    Encode mode        :CPU
    Render time        :02:21s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :FHD
    HW Acc.            :AMD
    Render As        :Magix Hevc
    Encode mode        :CPU
    Render time        :01:02s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)


    User name        :fifonik
    Machine            :OB
    VP version        :17
    VP Build number        :421
    CPU            :Ryzen 7 3700X
    Cores            :8
    GPU            :AMD RX580
    iGPU            :none
    OC            :Partial
    Render frame size    :UHD
    HW Acc.            :AMD
    Render As        :Magix Hevc
    Encode mode        :CPU
    Render time        :02:13s
    FPS            :5-6
    Comments        :Win10 Pro 1909, GPU Driver 20.2.2, GPU Model: MSI RX580 ARMOR 8G OC (factory overclocked)

With "Internal | Enable so4 Component Reader for AVC/M2TS" = FALSE (NOT default) the fps is 6-12. Render times are pretty much the same. I do not know why, but I did every test twice so pretty sure that results are correct. My only guess -- the affected region is too small.

JN- wrote on 5/19/2020, 3:31 AM

@fifonik Thats very interesting all right. So to summarise, given that most other users would have left that internal value to default=true, (previous fps avg.=7.5=false) your previous fps avg should be ~ 5.5, I say about ~ because of the gpu change to your system. I can update your previous cpu info, pointed out by Gary, + change your previous fps avg to 5.5.

So what would you like to do ...

A) Just modify those two items (Avg. fps + Cpu type) in your previous ss data.

B) Add in all of the new info with different gpu etc as an extra 2nd. set of data, leaving previous data untouched.

C) Remove your previous data and add in the new data you have now just posted.

D) Modify your previous data with the 2 items, cpu + fps change and add in 2nd. set of data.

E) Just modify your CPU info and add a note to your comments re: internals setting set to false.

F) Do nothing and let your comments here suffice.

1 … 3 4 5 … 15

of 15