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

[reidak]

anyone have a good config for AMD 9800X3D? just got one, upgrading from an i7 10700k

 

[Earthplayer]

anyone have a good config for AMD 9800X3D? just got one, upgrading from an i7 10700k

Base config is exactly what you want with that one (8 threads total and picking first thread of each core). No reason to change anything if you have any of the x3D 8-core CPUs. (5700x3D,5800x3D,7800x3D,9800x3D)

 

[kakashisex]

Do you guys think the upcoming 9950X3D will perform better than the 9800X3D in VaM? Due to most likely better single core performance thanks to higher clock speeds

 

[Salazarkz]

i didnt know how to use the patch with my cpu so i did copy the AMD Ryzen 7800x(3d) my cpu is 9700x with 4070 super and 32 gb 5600 speed ram here is before and after i dont know if this is good or not i am new to vam in general
NO OC
Before

After

boot config :
gfx-enable-gfx-jobs=1
gfx-enable-native-gfx-jobs=1
wait-for-native-debugger=0
*********************************
SkinMeshPartDLL :
[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

 

[Yujujo]

System specs:
Processor: Intel(R) Core(TM) i9-9900K CPU @ 3.60GHz 3.60 GHz
RAM: 48.0 GB (47.9 GB usable)
GPU 1: NVIDIA RTX 4060ti 8GB VRAM
GPU 2: NVIDIA RTX 3090 24GB VRAM

Can someone help me configure the performance patch settings for VR use?

Current patch settings:
[threads]
computeColliders=16
skinmeshPart=16
applyMorphs=16
skinmeshPartMaxPerChar=16
applyMorphMaxPerChar=16
affinity=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16

[threadsVR]
computeColliders=16
skinmeshPart=16
applyMorphs=16
skinmeshPartMaxPerChar=16
applyMorphMaxPerChar=16
affinity=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16

[profiler]
enabled=0

It seems like there should be a patch or plugin that runs on a specified GPU specifically for Soft body physics. For instance, I have an extra Nvidia RTX 4060ti GPU that fits in my system along with my main GPU. It would make sense to be able to click on enable soft-body physics and be presented with an option to specify the computing device, CPU, or GPU, and which one. GPU All, 1, 2, etc... Dedicating a GPU to soft-body physics would resolve this. I am unaware of why this was not implemented during the early development of VAM...Thoughts?

 

[Evo]

Evo updated CPU Performance Patch (Up to 30% faster physics, up to 60% more FPS) with a new update entry:

Support for VaM 1.22.0.12 | Version 13beta1evo

- Rebuilt Version 13beta1 to include security fixes for VaM 1.22.0.12

Read the rest of this update entry...

 

[MDKSBG]

I am using an older version of vam. Do I need to use the latest CPU Performance Patch?

 

[Case]

Evo updated CPU Performance Patch (Up to 30% faster physics, up to 60% more FPS) with a new update entry:

Support for VaM 1.22.0.12 | Version 13beta1evo



Read the rest of this update entry...

Thanks a lot!

 

[a1942y]

How to set for 13700 4090 ??
Thank you for your response in advance.

 

[Evo]

I am using an older version of vam. Do I need to use the latest CPU Performance Patch?

The version of VaM the patch was built for needs to be used for stability. Mix matching may result in crashes. See this section for the versions of VaM for each version of the patch.

 

[Evo]

How to set for 13700 4090 ??
Thank you for your response in advance.

See this config.

If others have a CPU that is not listed on the page and have a configuration working well for you. Post it here and I will update the page.

 

[a1942y]

See this config.

If others have a CPU that is not listed on the page and have a configuration working well for you. Post it here and I will update the page.

Thank you for your response~
The list above does not include the value for 13700, so may need to see who can provide it.

 

[Bob Nothing]

I have a 9800x3D and a 4090 (trying to buy a 5090.) Is there a reason I should try the 13 beta 1 over 12?

 

[Evo]

Thank you for your response~
The list above does not include the value for 13700, so may need to see who can provide it.

13700K is listed there. The settings are:

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

 

[Evo]

I have a 9800x3D and a 4090 (trying to buy a 5090.) Is there a reason I should try the 13 beta 1 over 12?

With 13beta1 there are fewer GC stutters for me, almost none.

 

[Salazarkz]

With 13beta1 there are fewer GC stutters for me, almost none.

hi is my setup was good for my build ? its above us i did send the benchmarks

 

[Evo]

hi is my setup was good for my build ? its above us i did send the benchmarks

Depending on your options in AMD settings, you could try not skipping the core numbers like this.

 

[Salazarkz]

Depending on your options in AMD settings, you could try not skipping the core numbers like this.

so should i chose 2 fastest core instead of 1.3.5 etc ? or should i overclook sorry i am new to this i do have smt enable if that what you mean

 

[Yujujo]

Is this how I should set this up for an intel cpu?

System specs:
Processor: Intel(R) Core(TM) i9-9900K CPU @ 3.60GHz 3.60 GHz

Current patch settings:
[threads]
computeColliders=16
skinmeshPart=16
applyMorphs=16
skinmeshPartMaxPerChar=16
applyMorphMaxPerChar=16
affinity=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16

[threadsVR]
computeColliders=16
skinmeshPart=16
applyMorphs=16
skinmeshPartMaxPerChar=16
applyMorphMaxPerChar=16
affinity=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16

[profiler]
enabled=0

 

[Slaanesh69]

Crash "too many heap sections" with 1.22.0.12. Your 7950x3d setting. No problems without the plugin.

7950x3d
64GB Ram
RTX 4090
Windows 11

 

[Evo]

@Yujujo
That setup is for 16 cores. Your 9900K has 8 real cores. Try the same settings I use on a 13900K (which also has 8 real cores, not Intel E cores)

@Slaanesh69
I also saw this before using the VaM updater to sync/repair core or a fresh VaM install. Then copy the patch files fresh. If it still happens after that, try v12evo

 

[Evo]

so should i chose 2 fastest core instead of 1.3.5 etc ? or should i overclook sorry i am new to this i do have smt enable if that what you mean

See the notes in FAQ on AMD settings and reasons where you may not see much improvement.

 

[a1942y]

13700K is listed there. The settings are:

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

Can the relevant values also be applied to the 13700?
Thank you for your response.

 

[supplovr]

@Yujujo
That setup is for 16 cores. Your 9900K has 8 real cores. Try the same settings I use on a 13900K (which also has 8 real cores, not Intel E cores)

@Slaanesh69
I also saw this before using the VaM updater to sync/repair core or a fresh VaM install. Then copy the patch files fresh. If it still happens after that, try v12evo

same here on a 14900K/96GB/4090: crashes on larger scenes with "too many heap sections".
Repair didn't work but crawling back from 13b1evo to 12evo worked rendering VAM stable again.
I experienced this with the original patch too and also remained on 12.

Regarding the GC stutters. Having the 1st line below in my boot.ini helped. Didn't tune the value of 3... so maybe tehre are better values depending on a particular system.

gc-max-time-slice=3
gfx-enable-gfx-jobs=1
gfx-enable-native-gfx-jobs=1
wait-for-native-debugger=0

 

[darkesthours443]

Your newest V13 for Vam xx.12 works perfect with both a Ryzen 5600x and a Ryzen 5950x, identical to your previous version of V13 for vam xx.03

[threads]
computeColliders=6
skinmeshPart=6
skinmeshPartMaxPerChar=6
applyMorphs=6
applyMorphMaxPerChar=6
affinity=1,2,3,4,5,6,7,8
engineAffinity=4,5

[threadsVR]
computeColliders=6
skinmeshPart=6
skinmeshPartMaxPerChar=6
applyMorphs=6
applyMorphMaxPerChar=6
affinity=1,2,3,4,5,6,7,8
engineAffinity=4,5
[profiler]
enabled=0

If people are getting errors with similar cpu's to the ones I posted above, I'd reckon they have a BAD vam installation, as I cannot imagine they would have a "bad cpu" instead.

In my 5000+ hours of using vam with various versions of CPU patch, I have NEVER gotten a "too many heap sections" error. And I can safely say I have seen it almost all in terms of VaM scenes and what not, which suggests to me it's an error that is related to very specific CPU's.