Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > High Efficiency Video Coding (HEVC)

Reply
 
Thread Tools Search this Thread Display Modes
Old 18th July 2019, 20:38   #61  |  Link
RanmaCanada
Registered User
 
Join Date: May 2009
Posts: 328
So stupid question, with the 3950x being released in September, and the current 3900 available now, does the advice in this thread work well with those processors, or is it even needed? Looking to upgrade to one of them come black friday, and just curious!
RanmaCanada is offline   Reply With Quote
Old 18th July 2019, 23:58   #62  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by RanmaCanada View Post
So stupid question, with the 3950x being released in September, and the current 3900 available now, does the advice in this thread work well with those processors, or is it even needed? Looking to upgrade to one of them come black friday, and just curious!
The 3950X should have about 33% more performance than the 3900X for 50% more money. Whether it's worth it or not is up to you.

Last edited by Stereodude; 19th July 2019 at 00:01.
Stereodude is offline   Reply With Quote
Old 21st July 2019, 19:59   #63  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
Quote:
Originally Posted by Atak_Snajpera View Post
Always use UMA. Video encoders unlike games do not need super low latency when accessing memory.
I used numa in my test and on the default settings x265 also chose numa.
Since I only have 1 16 core I'll try UMA next time with --pool 52 or something like that.
blublub is offline   Reply With Quote
Old 21st July 2019, 20:00   #64  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
Quote:
Originally Posted by Stereodude View Post
The 3950X should have about 33% more performance than the 3900X for 50% more money. Whether it's worth it or not is up to you.
Only if you are able to utilize all of those cores. If not then the 12c is likely the much better buy.
blublub is offline   Reply With Quote
Old 21st July 2019, 20:01   #65  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
Quote:
Originally Posted by benwaggoner View Post
The number of frame threads can have a significant impact on quality. It was pretty overwhelming on early versions of x265, and it's still ideal to run at -F 1 if possible for high quality/efficiency encoding.

Lookahead threading is reduced in slower presets, but I'm fuzzy on if and how it actually impacts quality much.

Other than that, I don't think that thread count impacts quality so much.

For Intel dual-socket systems at least, I've often found that pinning a job to just one socket ala --pools "-,+" has minimal impact on performance, although that probably varies with encoded frame size and number of cores per socket.

Since at default frame-thread was 5 and I am currently setting it to 2 or 3
I am already improving quality - so with frame-thread 2 or 3 I doubt that anyone will see a difference.
blublub is offline   Reply With Quote
Old 21st July 2019, 22:14   #66  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by blublub View Post
Only if you are able to utilize all of those cores. If not then the 12c is likely the much better buy.
Well, you can make the same argument until you're using the 6c or 4c model.
Stereodude is offline   Reply With Quote
Old 21st July 2019, 22:21   #67  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
Quote:
Originally Posted by Stereodude View Post
Well, you can make the same argument until you're using the 6c or 4c model.
No u can't.
With x265 it's easy to saturate 8c 16t but it's hard to do 16c 32t at 100% - if u leave the defaults you are only utilizing 60 to 70% so about a 10c 20t CPU on UHD resolution.
With tweaking u can get close to 100% but it's not yet clear if that degrades image quality or not - I can't see a difference but that doesn't mean there isn't any.
blublub is offline   Reply With Quote
Old 22nd July 2019, 00:46   #68  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by blublub View Post
No u can't.
Sure I can.
Quote:
With x265 it's easy to saturate 8c 16t but it's hard to do 16c 32t at 100% - if u leave the defaults you are only utilizing 60 to 70% so about a 10c 20t CPU on UHD resolution.
I can saturate any number of cores, even 64, with x265 without degrading quality.
Stereodude is offline   Reply With Quote
Old 22nd July 2019, 05:19   #69  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
With default settings? Which software and material are u using?
blublub is offline   Reply With Quote
Old 22nd July 2019, 12:29   #70  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by blublub View Post
With default settings? Which software and material are u using?
I didn't say with a single x265 instance. Break your source up into pieces and encode them at the same time. Or, run more than one encoding job at the same time. Think a little outside the box.
Stereodude is offline   Reply With Quote
Old 22nd July 2019, 14:58   #71  |  Link
Atak_Snajpera
RipBot264 author
 
Atak_Snajpera's Avatar
 
Join Date: May 2006
Location: Poland
Posts: 7,806
Quote:
Originally Posted by blublub View Post
With default settings? Which software and material are u using?
Just activate Distributed Encoding in Ripbot264 and done


This way I can saturate hundreds of threads!
Atak_Snajpera is offline   Reply With Quote
Old 22nd July 2019, 15:54   #72  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by Atak_Snajpera View Post
Just activate Distributed Encoding in Ripbot264 and done


This way I can saturate hundreds of threads!
How does Ripbot264 decide where to splice? Does it correctly handle breaking up a qpfile between the segments?
Stereodude is offline   Reply With Quote
Old 22nd July 2019, 15:57   #73  |  Link
Atak_Snajpera
RipBot264 author
 
Atak_Snajpera's Avatar
 
Join Date: May 2006
Location: Poland
Posts: 7,806
Splitting is done at keyframes reported by FFMS.
Atak_Snajpera is offline   Reply With Quote
Old 29th July 2019, 12:03   #74  |  Link
blublub
Registered User
 
Join Date: Jan 2015
Posts: 118
Mhh gotta try that
blublub is offline   Reply With Quote
Old 3rd May 2021, 17:30   #75  |  Link
nakTT
Registered User
 
Join Date: Dec 2008
Posts: 415
Quote:
Originally Posted by benwaggoner View Post
There can be quality regressions when using frame threading. Although that can be turned off directly via -F 1.
Hi, how can i do that in MeGUI? Currently I'm just using the option in the GUI menu to set the frame-threads to 1 to force it to just 1 thread because i don't know how to actually disable the frame threading altogether. All this while i only use GUI menu, not sure how to do it manually with MeGUI. Many thanks in advance.

Last edited by nakTT; 3rd May 2021 at 17:34.
nakTT is offline   Reply With Quote
Old 3rd May 2021, 21:21   #76  |  Link
benwaggoner
Moderator
 
Join Date: Jan 2006
Location: Portland, OR
Posts: 4,752
Quote:
Originally Posted by nakTT View Post
Hi, how can i do that in MeGUI? Currently I'm just using the option in the GUI menu to set the frame-threads to 1 to force it to just 1 thread because i don't know how to actually disable the frame threading altogether. All this while i only use GUI menu, not sure how to do it manually with MeGUI. Many thanks in advance.
-F 1 is the same as --frame-threads 1, so you're probably getting what you wanted. Actually turning the number of encoding threads down to 1 would be --pools "1" Which you definitely don't want to do outside of some specific multi-bitrate encoding scenarios. Even without frame-threads, there is plenty of parallelization available to x265 from WPP and such, which scales with frame size. If you're finding only a fraction of your available cores are being used, you should try --pmode.

Also, the potential quality regressions are proportional to the number of frame threads. I've found -F 2 as being generally safe while offering a big perf boost compared to -F 1
__________________
Ben Waggoner
Principal Video Specialist, Amazon Prime Video

My Compression Book
benwaggoner is offline   Reply With Quote
Old 4th May 2021, 07:59   #77  |  Link
Boulder
Pig on the wing
 
Boulder's Avatar
 
Join Date: Mar 2002
Location: Finland
Posts: 5,718
Before you start lowering the frame-threads parameter in all your encodes, I suggest you test how it looks. It's very likely that you won't be able to see the difference between -F 1 or -F 4.
__________________
And if the band you're in starts playing different tunes
I'll see you on the dark side of the Moon...
Boulder is offline   Reply With Quote
Old 4th May 2021, 18:29   #78  |  Link
benwaggoner
Moderator
 
Join Date: Jan 2006
Location: Portland, OR
Posts: 4,752
Quote:
Originally Posted by Boulder View Post
Before you start lowering the frame-threads parameter in all your encodes, I suggest you test how it looks. It's very likely that you won't be able to see the difference between -F 1 or -F 4.
True that. I've not really seen any substantial issues in a couple of years with higher frame threads. Doing more basic settings works reliably. When new features are introduced (RADL for example), there may be frame threading issues initially that get fixed later.
__________________
Ben Waggoner
Principal Video Specialist, Amazon Prime Video

My Compression Book
benwaggoner is offline   Reply With Quote
Old 9th January 2022, 09:32   #79  |  Link
tonemapped
Video Fanatic
 
tonemapped's Avatar
 
Join Date: Jul 2021
Location: Surrey
Posts: 89
Sorry to dig up an old thread, but I've found the most straight forward way to utilise 12 core is to run two instances (and use 2 for frame threads). That results in constant 99.5%+ CPU utilisation. Without that, and with pmode, I'd only see ~65-70% utilisation.
__________________
PC: R9 5900X | 32GB 3600 MT/s RAM | 2*1TB NVMe | RTX 3080 | water-cooled

NAS: SM 48-bay 240TB+ storage | Xeon 1220 | 32GB DDR4 ECC

HTPC: Pentium J5005 | 16GB RAM | 256GB SSD | 15W
tonemapped is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 06:58.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.