CPU Performance Patch (Up to 30% faster physics, up to 60% more FPS)

Plugins + Scripts CPU Performance Patch (Up to 30% faster physics, up to 60% more FPS)

Download [1.97 MB]
Did you scroll through all of the settings on the home page of the CPU Performance Booster? Did you search this thread for the settings?

I would try these on that CPU as it looks similar to yours:

These are from the recommended settings on the OVERVIEW page of this CPU Performance Boost Plugin.

Intel 13900K/13800K/13700K
(we are using the performance cores only)


Spoiler
HI! THANKS FOR THE REPLY. I’VE ALREADY TRIED ALL OF THE DEFAULT SETTINGS ABOVE, WHICH IS WHY I POSTED HERE FOR HELP.
 
I'm seeing 98 threads on my Intel i7-8700K CPU. Is this optimal? Should I see a log file called SkinMeshPartDLL.log being created at started-up. I don't see one. Can anyone recommend optimal settings for SkinMeshPartDLL.ini and boot.config?
 
HI! THANKS FOR THE REPLY. I’VE ALREADY TRIED ALL OF THE DEFAULT SETTINGS ABOVE, WHICH IS WHY I POSTED HERE FOR HELP.
Which setting gives you the best performance? Best FPS? I am not sure what level each thread is running in my install, I just played around until I had a nice improvement in performance that was stable and left it at that. There are too many different processors out there for the developer to test and post the "official" settings for all of them.

I use the plugin because with it I can run a complex scene, in VR, with four characters at an acceptable FPS and without it, I can't. And now, I can do that same thing with DLSS5.

Find what works and go with it.
 
Anyone know why this happen ? I have use this setting...
[threads]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8
affinity=1,3,5,7,9,11,13,15

[threadsVR]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8
affinity=1,3,5,7,9,11,13,15

[profiler]
enabled=0

Ummmh ... what do you mean? The fact that first core ist boosting hardest and only the physical cores are boosting? Afaics, that's the processor doing what you're telling it to do? You've got a Zen 7 5700x 8 core, Zen architecture has one "virtual core" for every real physical core. Also, you'd want to use your two best cores for heavy workloads like VaM. Performance plugin accounts for both those properties with the "affinity" and "engine affinity" commands:
* "affinity = ..." command lists which cores Windows can assign VaM to. So "affinity = 1,3,5,7,9,11,13,15" is telling Windows to only use physical cores for VaM (i.e. the odd-numbered ones in your graph), not the virtual ones (i.e. the even numbered). And it looks like Windows does as its told. That's how turtlebackgoofy meant the plugin to work.
* Another feature of Perf Boost is to assign VaM to your first two most performant cores with the "engineAffinity= ..." - Now, you haven't used enginAffinity, so I think the fact that it uses your first and fourth physical core (numbers 1 and 7 in your graph) the most is because your Windows THINKS those are your best cores. Windows isn't always right about that. If you aren't sure whether those two are actually your best cores, you can find out with RyZen Master or similar monitoring tools - Headsup: Some monitoring apps enumerate the cores differntly: They start counting from 0 rather than from 1, and they might only list physical cores. TL;DR - Find out two best cores with some monitoring tool (RyZenMaster etc.) and use engineAffinity to tell perf plugin which those are - if, e.g., it tells you that your second and fourth cores are your best, then you'd tell performance plugin: engineAffinity= 3,7 (ie. the numbers corresponding to your physical cores 2 and 4). Yeah, that's a bit confusing.

TL;DR - Afaics, this is your processor doing exactly what you've told it to do?

P.S.: This is my old Zen 7 5800x with
affinity=1,3,5,7,9,11,13,15 <- "Use only physical cores"
engineAffinity=3,7 <- "Use 2nd and 4th core"

PerfMon_engineAffinity_3_7.png
 
Last edited:
Back
Top Bottom