Not even I am sure of what I mean. lol
I had this thought: I saw your post, I supposed you managed to get GPULightmass working on 4.23 by building it from the source with GPULightmass source. So I believed that you had the resulting dlls and exes from the build that everybody else just needed to overwrite in the launcher version of 4.23 in order to use GPULightmass’ files compatible with it.
Is that it? Because if its, I guess if you share the same dlls and exes that come in the binary versions of GPULightmass, but built with your 4.23 source, they could be used to replace the launcher version files. Unless you modified 4.23 prior to building, of course.
I managed to port GPU Lightmass to 4.23.0-preview-6. Here’s the download for the binary engine integration:
And here’s my UE4 fork with GPU Lightmass cherry-picked from 's code on top of 4.23:
https://.com/irobot/UnrealEngine/tree/GPU-Lightmass-cherry-pick-4.23.0-preview-6
Updated for UE4.23.0-preview-7. Binaries download link is the same.
Source branch: https://.com/irobot/UnrealEngine/tree/GPU-Lightmass-cherry-pick-4.23.0-preview-7
Great news! Is there any way you can provide a patch so that people can integrate it into their source UE4 builds (for example I am using Oculus fork of UE4 and I could use patch instead of weeding out changes from the entire stock UE4 and implementing them into Oculus fork of UE4) ?
If you look at the commit log, you’ll notice that the only difference from the original EpicGames source are the last three commits, which I simply cherry-picked from 's fork (with some merge conflicts resolved, obviously). You could do the same for the OculusVR fork.
Has anyone encountered this problem? I’m not sure what am I doing wrong. Tried all settings already and the result is still the same. 4.22.3, gtx 1070. Just applied 4.22 version onto mine, then replaces StaticLightingProductionQuality in baselightmass with lines:
NumShadowRaysScale=32
NumPenumbraShadowRaysScale=64
ApproximateHighResTexelsPerMaxTransitionDistanceScale=9
MinDistanceFieldUpsampleFactor=7
NumHemisphereSamplesScale=100
NumImportanceSearchPhotonsScale=6
NumDirectPhotonsScale=32
; Decrease direct photon search distance so that we will have more accurate shadow transitions. This requires a higher density of direct photons.
DirectPhotonSearchDistanceScale=.5
NumIndirectPhotonPathsScale=32
; Need a lot of indirect photons since we have increased the number of first bounce photons to use for final gathering with NumImportanceSearchPhotonsScale
NumIndirectPhotonsScale=64
NumIndirectIrradiancePhotonsScale=32
; Decreasing the record radius results in more records, which increases quality
RecordRadiusScaleScale=.45
InterpolationMaxAngleScale=.75
IrradianceCacheSmoothFactor=.75
NumAdaptiveRefinementLevels=3
AdaptiveBrightnessThresholdScale=.25
AdaptiveFirstBouncePhotonConeAngleScale=2.5
AdaptiveSkyVarianceThresholdScale=.5
Sorry if I’m repeating this problem, thanks in advance!
Why is my GPU usage remains 16% when Im using GPU lightmass bake? While CPU is 3X% and higher?
Also my memories usage most of the time remain 75-95%, and when memories drop to 50% the cpu usage will go up to 40-50%, does it mean my PC memories is relatively low that holding the CPU back in the rendering process?
Thanks again for those who script such a powerful GPU baking tool!
In the bottom of BaseLightmass, which you might have gotten rid of accidentally if you copied it in from a different project/engine version.
You can use https://.com/sgeraldes/LightmassConfiguration/ to make tweaking that a bit more practical, since you can just hop between quality presets between bakes easily with it.
Yes, correct. Those pretty much align with a ‘low/medium quality’ preset, there.
Bumping them up a decent bit, like with the LightmassConfigurator should give more samples and smooth out the final result a lot.
Bumping it up to 64/16 or 128/32 would give you a high/very high quality setting, pretty much.