• Hi Guest!

    We are extremely excited to announce the release of our first Beta1.1 and the first release of our Public AddonKit!
    To participate in the Beta, a subscription to the Entertainer or Creator Tier is required. For access to the Public AddonKit you must be a Creator tier member. Once subscribed, download instructions can be found here.

    Click here for information and guides regarding the VaM2 beta. Join our Discord server for more announcements and community discussion about VaM2.
  • Hi Guest!

    VaM2 Resource Categories have now been added to the Hub! For information on posting VaM2 resources and details about VaM2 related changes to our Community Forums, please see our official announcement here.
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]
Hello :)
First of all thanks for your hard work.
I really wanted to make your plugin works on my computer and i tried a lot of differents configuration but i didnt see any improvment (in desktop or in VR)
Maybe i miss something, some tweak in my bios ? i dont know.

Here is my config
I7 12700KF, 32 gb RAM, SSD, RTX 3080

and all the .ini files i tried :

Code:
[threads]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8
affinity=1,3,5,7,9,11,13,15
engineAffinity=1,3

[threadsVR]
computeColliders=4
skinmeshPart=2
applyMorphs=2
skinmeshPartMaxPerChar=4
applyMorphMaxPerChar=2
affinity=1,3,5,7,9,11,13,15
engineAffinity=1,3

[profiler]
enabled=0

Code:
[threads]
computeColliders=6
skinmeshPart=1
affinity=1,3,5,7,9,11,13,15

Code:
[threads]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8

affinity=0,2,4,6,8,10,12,14
engineAffinity=0,2

Some benchmark captures (i also tried different resolutions).


Without patch

Benchmark-20251202-202525.png

With patch

Benchmark-20251202-174222.png


without patch :


Benchmark-20251202-133204.png

with patch :

Benchmark-20251202-125650-without.png
 
I figured out the issue that was causing occasional too many heap sections crashes with the patch -

Resolved by rolling back to v12evo (patched to include security fixes for VaM 1.22.0.12 and removing job-worker-count=12, which was effectively crippling the patch and causing instability, from the boot.config file).

I did extensive benchmark testing with MacGruber's Benchmark to baseline Desktop and VR (My Optimal Settings and MacGruber's Official Benchmark settings and compare to patched.

I also benchmarked VAM Startup time and heavy scene load times (initial and reload) baseline and patched.

Rig: 9800X3D, TUF 5090 (OC/UV), 64GB KF Beast 6000-CL30 tuned to CL26. CPU is most relevant here.

Results (Benchmarks attached, Summaries below):VR Notes - Some of the VR Benchmarks show my VR HMD is an Oculus Quest2, but I am running a MeganeX SL8K native SteamVR setup via SBoys3's driver. Also my OS is Windows 11 Pro not Windows 10. All other info is accurate.

In SteamVR I am running 150% Steam Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On (I've tested both and quality/performance is slightly better with ASF On, this helps when render res upscaling).

My Optimal Custom User Preferences Settings: (Render Scale 1.00, Mirrored Surfaces Enabled, Realtime Reflection Probes off, Generate Depth Texture off, Soft Body Physics on, Shader Quality High, MSAA 4X, Pixel Light Count 6, Smooth Passes 2, Process Priority High, Physics Rate 90 HZ, Physics Update Cap 1, High Quality Physics on, Glow Effects off, Desktop Vsync off). Tested baselines and these were the most optimal settings with best quality to performance. Mirrored Surfaces enabled is optional.

Desktop Optimal Settings Baseline:
Benchmark-20251207-Desktop- OPTIMAL Baseline.png


Desktop Optimal Settings Patched:
Benchmark-20251208-Desktop-OPTIMAL-CPU Patch.png
Desktop Optimal Settings Baseline to Patch Comp: ~22% avg FPS gain, ~45% avg TotalTime improvement


Desktop Official Benchmark Baseline:
Benchmark-20251207-Desktop-Official Baseline.png


Desktop Official Benchmark Patched:
Benchmark-20251208-Desktop-Official-CPU Patch.png


Desktop Official Benchmark Baseline to Patch Comp: ~10% avg FPS gain, ~40% avg TotalTime improvement


VR Optimal Settings Baseline (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
Benchmark-20251207-VR- OPTIMAL- Baseline-150% RR ASF On.png


VR Optimal Settings Patched (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
Benchmark-20251208-VR-OPTIMAL-CPU Patch - 150% RR ASF On.png

VR Optimal Settings Baseline to Patch Comp: FPS was pegged at ~90fps for both (expected due to 90Hz Physics Rate), but TotalTime Avg improved by 10-20%


VR Official Benchmark Baseline (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
Benchmark-20251207-VR Official Baseline -150% RR ASF On.png


VR Official Benchmark Patched (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
Benchmark-20251208-VR-Official - CPU Patch - 150% RR ASF On.png

VR Official Benchmark Baseline to Patch Comp: All tests FPS was pegged at 90fps avg except for Mirror - 62fps improved to 67fps, baseline 3 and simpler physics tests were both pegged at 45fps avg, but again - TotalTime Improvement of ~10-20%

VR Benefits - Biggest benefits here are reduced Avg/Min/Max TotalTime, reduced drops, more stability at 90fps under heavy load

VAM Startup/Heavy Scene Load times - Baseline vs CPU Performance Patch Comparison:

VAM startup time baseline - 30.28 seconds
VAM startup time patch - 30.00 seconds (No patch improvement)

Here's where it gets more interesting...

VAM very heavy scene (1) load time baseline (TGC VS Reloaded 02.12.25 (Patreon Full Latest) - 3:35.16 initial and reload (almost identical)
VAM very heavy scene (1) load time patch (TGC VS Reloaded 02.12.25 (Patreon Full Latest) - 3:04.38 initial, 2:40.03 reload (14.3% faster initial load, 25.6% faster reload!)

VAM heavy scene (2) load time baseline (CuddleMocap 010-Lap-Dance-01.2) - 1:05.87 initial and reload (almost identical)
VAM heavy scene 2 load time patch (CuddleMocap 010-Lap-Dance-01.2) - 42.02 initial, 36.45 reload (36.2% faster initial load, 44.7% faster reload!!)

VAM Startup - unchanged
Initial load of heavy scenes - 14-36% faster load
Reload of heavy scenes - 25-45% faster load


My Optimal patch settings for the 9800X3D (on patched v12evo)

SkinMeshPartDLL.ini:
[threads]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8
affinity=1,3,5,7,9,11,13,15

[threadsVR]
computeColliders=4
skinmeshPart=2
applyMorphs=2
skinmeshPartMaxPerChar=4
applyMorphMaxPerChar=2
affinity=1,3,5,7,9,11,13,15

boot.config:

gfx-enable-gfx-jobs=1
gfx-enable-native-gfx-jobs=1
gfx-disable-mt-rendering=0
wait-for-native-debugger=0
gc-max-time-slice=3
single-instance=1

Hope this helps.
 
I figured out the issue that was causing occasional too many heap sections crashes with the patch -

Resolved by rolling back to v12evo (patched to include security fixes for VaM 1.22.0.12 and removing job-worker-count=12, which was effectively crippling the patch and causing instability, from the boot.config file).

I did extensive benchmark testing with MacGruber's Benchmark to baseline Desktop and VR (My Optimal Settings and MacGruber's Official Benchmark settings and compare to patched.

I also benchmarked VAM Startup time and heavy scene load times (initial and reload) baseline and patched.

Rig: 9800X3D, TUF 5090 (OC/UV), 64GB KF Beast 6000-CL30 tuned to CL26. CPU is most relevant here.

Results (Benchmarks attached, Summaries below):VR Notes - Some of the VR Benchmarks show my VR HMD is an Oculus Quest2, but I am running a MeganeX SL8K native SteamVR setup via SBoys3's driver. Also my OS is Windows 11 Pro not Windows 10. All other info is accurate.

In SteamVR I am running 150% Steam Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On (I've tested both and quality/performance is slightly better with ASF On, this helps when render res upscaling).

My Optimal Custom User Preferences Settings: (Render Scale 1.00, Mirrored Surfaces Enabled, Realtime Reflection Probes off, Generate Depth Texture off, Soft Body Physics on, Shader Quality High, MSAA 4X, Pixel Light Count 6, Smooth Passes 2, Process Priority High, Physics Rate 90 HZ, Physics Update Cap 1, High Quality Physics on, Glow Effects off, Desktop Vsync off). Tested baselines and these were the most optimal settings with best quality to performance. Mirrored Surfaces enabled is optional.

Desktop Optimal Settings Baseline:
View attachment 547945

Desktop Optimal Settings Patched:
View attachment 547946Desktop Optimal Settings Baseline to Patch Comp: ~22% avg FPS gain, ~45% avg TotalTime improvement


Desktop Official Benchmark Baseline:
View attachment 547947

Desktop Official Benchmark Patched:
View attachment 547948

Desktop Official Benchmark Baseline to Patch Comp: ~10% avg FPS gain, ~40% avg TotalTime improvement


VR Optimal Settings Baseline (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
View attachment 547949

VR Optimal Settings Patched (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
View attachment 547950
VR Optimal Settings Baseline to Patch Comp: FPS was pegged at ~90fps for both (expected due to 90Hz Physics Rate), but TotalTime Avg improved by 10-20%


VR Official Benchmark Baseline (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
View attachment 547951

VR Official Benchmark Patched (150% SteamVR Render Resolution, Motion Smoothing Off, Advanced Supersample Filtering On):
View attachment 547952
VR Official Benchmark Baseline to Patch Comp: All tests FPS was pegged at 90fps avg except for Mirror - 62fps improved to 67fps, baseline 3 and simpler physics tests were both pegged at 45fps avg, but again - TotalTime Improvement of ~10-20%

VR Benefits - Biggest benefits here are reduced Avg/Min/Max TotalTime, reduced drops, more stability at 90fps under heavy load

VAM Startup/Heavy Scene Load times - Baseline vs CPU Performance Patch Comparison:

VAM startup time baseline - 30.28 seconds
VAM startup time patch - 30.00 seconds (No patch improvement)

Here's where it gets more interesting...

VAM very heavy scene (1) load time baseline (TGC VS Reloaded 02.12.25 (Patreon Full Latest) - 3:35.16 initial and reload (almost identical)
VAM very heavy scene (1) load time patch (TGC VS Reloaded 02.12.25 (Patreon Full Latest) - 3:04.38 initial, 2:40.03 reload (14.3% faster initial load, 25.6% faster reload!)

VAM heavy scene (2) load time baseline (CuddleMocap 010-Lap-Dance-01.2) - 1:05.87 initial and reload (almost identical)
VAM heavy scene 2 load time patch (CuddleMocap 010-Lap-Dance-01.2) - 42.02 initial, 36.45 reload (36.2% faster initial load, 44.7% faster reload!!)

VAM Startup - unchanged
Initial load of heavy scenes - 14-36% faster load
Reload of heavy scenes - 25-45% faster load


My Optimal patch settings for the 9800X3D (on patched v12evo)

SkinMeshPartDLL.ini:
[threads]
computeColliders=8
skinmeshPart=8
applyMorphs=8
skinmeshPartMaxPerChar=8
applyMorphMaxPerChar=8
affinity=1,3,5,7,9,11,13,15

[threadsVR]
computeColliders=4
skinmeshPart=2
applyMorphs=2
skinmeshPartMaxPerChar=4
applyMorphMaxPerChar=2
affinity=1,3,5,7,9,11,13,15

boot.config:

gfx-enable-gfx-jobs=1
gfx-enable-native-gfx-jobs=1
gfx-disable-mt-rendering=0
wait-for-native-debugger=0
gc-max-time-slice=3
single-instance=1

Hope this helps.
can I ask you what boot.config file you're talking about?, where is it located? and where can I find the tached v12evo?
 
Back
Top Bottom