Handbrake for High Core Count Utilization

[intel]

Description of problem or question:

Handbrake unable to utilize high core count CPU


Steps to reproduce the problem (If Applicable):

None-


HandBrake version (e.g., 1.0.0):

1.2.2

Operating system and version (e.g., Ubuntu 16.04 LTS, macOS 10.13 High Sierra, Windows 10 Creators Update):

Windows Server 2019 Datacenter Core with Desktop Experience


HandBrake Activity Log ***required*** (see How-to get an activity log)

Code: Select all

Please replace this text with the contents of your log file between the two code tags - OR -  provide a pastebin URL in place of these 3 lines.

So I'm currently encoding using a server, which has 96 vCPU, but the current handbrake version only utilize up to 48 cores, is there any way to unlock handbrake's ability to utilize all CPU cores available?

[rollin_eng]

Could you please post your HB logs, instructions can be found here:

https://handbrake.fr/docs/en/latest/hel ... y-log.html

[Woodstock]

Run multiple copies.

There are limits to how much the various encoders will "spread the load". And not all filters allow more than one thread, which also imposes speed limitations.

Not every problem gets "better" with more cores. Beyond a certain point (which varies with the task), the overhead of keeping things synchronized exceeds the benefit of spreading the load.

[BradleyS]

I believe we need libnuma for multi-socket systems, which is in consideration. What Woodstock said is still relevant.

[JohnAStebbins]

I believe we need libnuma for multi-socket systems, which is in consideration. What Woodstock said is still relevant.

Not just mult-socket these days. AMD Epyc and some Ryzen are multi-module in a single socket. These also benefit from numa.

[mduell]

Encode higher res content or use more threaded filters.

If you don't need either of those, run multiple copies in parallel.