Author Topic: How to Optimize PL for iRacing (or other titles that use just 2 cores)  (Read 2916 times)

Offline Globespy

  • Basic
  • **
  • Posts: 18
I am an avid fan of the motorsports simulation 'iRacing'.

It's still using an old engine - they only updated from DX9 to DX11 in late 2016 as many of their users were on WinXP and of course it does not support DX11 so they had to wait until they could 'live without' these members if they chose to not upgrade their 15+ year old OS!

It uses 2 CPU cores, one for all the physics calculations and the other for everything else (rendering), yet it's really just 1 core doing most of the work.
It's usual to see CPU usage (i7 8086K @ 5.3Ghz on all cores) maxing out at 20-25%.
I race exclusively in VR, and similarly my GPU (1080ti) rarely sees above 45% utilization.

iRacing developers use the usual DX11 excuse saying that DX11's design is the limiting factor - sure it's not truly multi-threaded like DX12.
But if I look at other DX11 titles like 'Assetto Corsa' or 'Project Cars 2' they use at least 4 cores and will max out the GPU usage, so iRacing is partially correct.

So, is there anything I can do with PL to optimize this title (the main cores doing all the work)?

Thanks!
« Last Edit: January 24, 2019, 03:27:47 PM by Globespy »



Offline edkiefer

  • Volunteer User Moderator
  • Member#
  • *****
  • Posts: 1789
  • Gender: Male
Re: How to Optimize PL for iRacing (or other titles that use just 2 cores)
« Reply #1 on: January 24, 2019, 04:53:18 PM »
I am an avid fan of the motorsports simulation 'iRacing'.

It's still using an old engine - they only updated from DX9 to DX11 in late 2016 as many of their users were on WinXP and of course it does not support DX11 so they had to wait until they could 'live without' these members if they chose to not upgrade their 15+ year old OS!

It uses 2 CPU cores, one for all the physics calculations and the other for everything else (rendering), yet it's really just 1 core doing most of the work.
It's usual to see CPU usage (i7 8086K @ 5.3Ghz on all cores) maxing out at 20-25%.
I race exclusively in VR, and similarly my GPU (1080ti) rarely sees above 45% utilization.

iRacing developers use the usual DX11 excuse saying that DX11's design is the limiting factor - sure it's not truly multi-threaded like DX12.
But if I look at other DX11 titles like 'Assetto Corsa' or 'Project Cars 2' they use at least 4 cores and will max out the GPU usage, so iRacing is partially correct.

So, is there anything I can do with PL to optimize this title (the main cores doing all the work)?

Thanks!
I answered in other thread how you can micromanage affinity to get better results with "multiple" software running at same time, but PL cannot alter how the application is coded. you can only act on all the processes it uses, not being able to alter usage or GPU usage. If your 6 core is getting really low usage across all cores you could limit the game to 4 cores say and usage should go up, just have to see if it doesn't hurt gameplay by limiting cores.
Then you would have free cores for whatever you want to run alongside.
Bitsum QA Engineer