LuxCoreRender Forums

Hey there!
I recently dared to switch from Windows to Linux finally. I decided to give PopOS a shot and i am happy with that by now, because almost everything seems to work out of the box.
But if i want to use BlendLuxCore with Cuda (Optix) i get this Error message:

OptixAccel CUDA program compilation error

E-Cycles however works fine with Optix.

Am i missing some dependencies needed for Luxcore? I just have installed latest Nvidia drivers but no Cuda and Optix packages.

System:
PopOS 20.10
Nvidia Driver 455.38
Asus 2080ti
Ryzen 1700x

Thanks!

johannes.wilde wrote: ↑Tue Jan 12, 2021 9:47 am But if i want to use BlendLuxCore with Cuda (Optix) i get this Error message:

OptixAccel CUDA program compilation error

Can you post a screenshot of the error ?

Can you check if there is some more information about the error on the console ?

Here is the Output from the Console:


[Engine/Viewport] New session
[Exporter] Creating session
[SDL][18.043] Define ImageMap: NamedObject
[SDL][18.044] Camera type: perspective
[SDL][18.044] Camera position: Point[14.7279, -6.50511, 8.01803]
[SDL][18.044] Camera target: Point[13.9091, -6.14344, 7.57226]
[SDL][18.044] Camera clipping plane disabled
WARNING: Material "Material": Missing node tree
[SDL][18.046] Material definition: 139660552593544
[SDL][18.046] Scene objects count: 1
[SDL][18.046] Light definition: 139660552519176
[SDL][18.046] Light definition: __WORLD_BACKGROUND_LIGHT__
[LuxCore][18.066] Configuration:
[LuxCore][18.066] path.pathdepth.total = 7
[LuxCore][18.066] path.pathdepth.diffuse = 5
[LuxCore][18.066] path.pathdepth.glossy = 5
[LuxCore][18.066] path.pathdepth.specular = 6
[LuxCore][18.066] path.hybridbackforward.enable = 0
[LuxCore][18.066] path.hybridbackforward.partition = 0.8
[LuxCore][18.066] path.hybridbackforward.glossinessthreshold = 0.049
[LuxCore][18.066] rtpathcpu.zoomphase.size = 4
[LuxCore][18.066] rtpathcpu.zoomphase.weight = 0
[LuxCore][18.066] renderengine.type = "RTPATHCPU"
[LuxCore][18.066] sampler.type = "RTPATHCPUSAMPLER"
[LuxCore][18.066] film.width = 2116
[LuxCore][18.066] film.height = 1227
[LuxCore][18.066] film.filter.type = "BLACKMANHARRIS"
[LuxCore][18.066] film.filter.width = 1.5
[LuxCore][18.066] lightstrategy.type = "LOG_POWER"
[LuxCore][18.066] scene.epsilon.min = 1e-05
[LuxCore][18.066] scene.epsilon.max = 0.1
[LuxCore][18.066] film.opencl.enable = 1
[LuxCore][18.066] film.opencl.device = 1
[LuxCore][18.067] path.forceblackbackground.enable = 0
[LuxCore][18.067] renderengine.seed = 1
[LuxCore][18.067] film.outputs.0.type = "RGB_IMAGEPIPELINE"
[LuxCore][18.067] film.outputs.0.index = 0
[LuxCore][18.067] film.outputs.0.filename = "RGB_IMAGEPIPELINE_0.png"
[LuxCore][18.067] film.outputs.1.type = "ALBEDO"
[LuxCore][18.067] film.outputs.1.filename = "ALBEDO.exr"
[LuxCore][18.067] film.imagepipelines.000.0.type = "NOP"
[LuxCore][18.067] film.imagepipelines.000.1.type = "TONEMAP_LINEAR"
[LuxCore][18.067] film.imagepipelines.000.1.scale = 1
[LuxCore][18.067] film.imagepipelines.000.2.type = "OPTIX_DENOISER"
[LuxCore][18.067] film.imagepipelines.000.2.sharpness = 0
[LuxCore][18.067] film.imagepipelines.000.2.minspp = 1
[LuxCore][18.067] film.imagepipelines.000.radiancescales.0.enabled = 1
[LuxCore][18.067] film.imagepipelines.000.radiancescales.0.globalscale = 1
[LuxCore][18.067] film.imagepipelines.000.radiancescales.0.rgbscale = 1 1 1
[LuxCore][18.067] batch.haltspp = 0 0
[LuxCore][18.067] batch.halttime = 0
[LuxCore][18.067] File Name Resolver Configuration:
Export took 0.0 s
[LuxCore][18.068] Film resolution: 2116x1227
[SDL][18.069] Film output definition: RGB_IMAGEPIPELINE [image.png]
[SDL][18.071] Image pipeline: film.imagepipelines.000
[SDL][18.071] Image pipeline step 0: NOP
[SDL][18.071] Image pipeline step 1: TONEMAP_LINEAR
[SDL][18.071] Image pipeline step 2: OPTIX_DENOISER
[SDL][18.071] Film output definition: RGB_IMAGEPIPELINE [RGB_IMAGEPIPELINE_0.png]
[SDL][18.071] Film output definition: ALBEDO [ALBEDO.exr]
[LuxRays][18.072] OpenCL support: enabled
[LuxRays][18.072] OpenCL Platform 0: NVIDIA CUDA
[LuxRays][18.072] CUDA support: enabled
[LuxRays][18.072] CUDA support: available
[LuxRays][18.072] CUDA driver version: 11.10
[LuxRays][18.072] CUDA device count: 1
[LuxRays][18.072] Optix support: available
[LuxRays][18.072] Device 0 name: Native
[LuxRays][18.072] Device 0 type: NATIVE_THREAD
[LuxRays][18.072] Device 0 compute units: 1
[LuxRays][18.072] Device 0 preferred float vector width: 4
[LuxRays][18.072] Device 0 max allocable memory: 17592186044415MBytes
[LuxRays][18.072] Device 0 max allocable memory block size: 17592186044415MBytes
[LuxRays][18.072] Device 0 has out of core memory support: 0
[LuxRays][18.072] Device 1 name: GeForce RTX 2080 Ti
[LuxRays][18.072] Device 1 type: OPENCL_GPU
[LuxRays][18.072] Device 1 compute units: 68
[LuxRays][18.072] Device 1 preferred float vector width: 1
[LuxRays][18.073] Device 1 max allocable memory: 11016MBytes
[LuxRays][18.073] Device 1 max allocable memory block size: 2754MBytes
[LuxRays][18.073] Device 1 has out of core memory support: 0
[LuxRays][18.073] Device 2 name: GeForce RTX 2080 Ti
[LuxRays][18.073] Device 2 type: CUDA_GPU
[LuxRays][18.073] Device 2 compute units: 64
[LuxRays][18.073] Device 2 preferred float vector width: 1
[LuxRays][18.073] Device 2 max allocable memory: 11016MBytes
[LuxRays][18.073] Device 2 max allocable memory block size: 17592186044415MBytes
[LuxRays][18.073] Device 2 has out of core memory support: 1
[LuxRays][18.073] Device 2 CUDA compute capability: 7.5
[LuxRays][18.074] Creating 16 intersection device(s)
[LuxRays][18.074] Allocating intersection device 0: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 1: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 2: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 3: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 4: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 5: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 6: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 7: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 8: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 9: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 10: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 11: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 12: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 13: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 14: Native (Type = NATIVE_THREAD)
[LuxRays][18.074] Allocating intersection device 15: Native (Type = NATIVE_THREAD)
[LuxCore][18.074] Configuring 16 CPU render threads
[LuxRays][18.075] Preprocessing DataSet
[LuxRays][18.075] Total vertex count: 24
[LuxRays][18.075] Total triangle count: 12
[LuxRays][18.075] Preprocessing DataSet done
[LuxRays][18.075] Adding DataSet accelerator: EMBREE
[LuxRays][18.075] Total vertex count: 24
[LuxRays][18.075] Total triangle count: 12
[LuxRays][18.079] EmbreeAccel build time: 2ms
[LuxCore][18.282] Film hardware image pipeline
[LuxCore][18.580] Film hardware device used: GeForce RTX 2080 Ti CUDAIntersect (Type: CUDA_GPU)
[LuxRays][18.580] [Device GeForce RTX 2080 Ti CUDAIntersect] IMAGEPIPELINE buffer size: 30425Kbytes
[LuxRays][18.585] [Device GeForce RTX 2080 Ti CUDAIntersect] ALBEDO buffer size: 40567Kbytes
[LuxRays][18.592] [Device GeForce RTX 2080 Ti CUDAIntersect] Merge buffer size: 40567Kbytes
[LuxRays][18.592] [MergeSampleBuffersOCL] Compiler options: -D LUXRAYS_OPENCL_KERNEL -D SLG_OPENCL_KERNEL -D LUXRAYS_CUDA_DEVICE -D LUXRAYS_OS_LINUX --use_fast_math
[LuxRays][18.592] [MergeSampleBuffersOCL] Compiling kernels
[LuxRays][18.606] [MergeSampleBuffersOCL] CUDA program compilation error:
nvrtc: error: failed to open libnvrtc-builtins.so.
Make sure that libnvrtc-builtins.so is installed correctly.
Traceback (most recent call last):
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/engine/base.py", line 124, in view_draw
viewport.view_draw(self, context, depsgraph)
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/engine/viewport.py", line 207, in view_draw
framebuffer.update(engine.session, scene)
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/draw/viewport.py", line 229, in update
luxcore_session.GetFilm().GetOutputFloat(self._output_type, self.buffer)
RuntimeError: MergeSampleBuffersOCL CUDA program compilation error
[LuxCore][18.610] [GeForce RTX 2080 Ti CUDAIntersect] Memory used for hardware image pipeline: 111561Kbytes
[LuxRays][18.616] [Optix][4][DISK CACHE] Closed database: "/var/tmp/OptixCache_xxxx/cache7.db"
[LuxRays][18.616] [Optix][4][DISK CACHE] Cache data size: "38.3 MiB"

Here is the Output from the Console:

You can use this quote features to make line of code easy to read

The problem is here:

johannes.wilde wrote: ↑Tue Jan 12, 2021 1:38 pm [LuxRays][18.592] [MergeSampleBuffersOCL] Compiling kernels
[LuxRays][18.606] [MergeSampleBuffersOCL] CUDA program compilation error:
nvrtc: error: failed to open libnvrtc-builtins.so.
Make sure that libnvrtc-builtins.so is installed correctly.
Traceback (most recent call last):
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/engine/base.py", line 124, in view_draw
viewport.view_draw(self, context, depsgraph)
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/engine/viewport.py", line 207, in view_draw
framebuffer.update(engine.session, scene)
File "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/draw/viewport.py", line 229, in update
luxcore_session.GetFilm().GetOutputFloat(self._output_type, self.buffer)
RuntimeError: MergeSampleBuffersOCL CUDA program compilation error

Check your "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/bin" directory, you should have there a "libnvrtc.so" and "libnvrtc-builtins.so" files.

Can you post here the list of files in your "/home/xxxx/.config/blender/2.92/scripts/addons/BlendLuxCore/bin" directory ?

Hi Dade!
Thanks for your help!
The files seem to in place. Is it relevant, that i don't use a installed version of Blender but a daily build?

johannes.wilde wrote: ↑Wed Jan 13, 2021 9:33 am The files seem to in place.

It is hard to say. One reason could be you have another "libnvrtc.so" installed on your PC and it is loaded instead of the one in the bin directory. Another possibility is that .so are not loaded because the fail of some other dependecy.

Can you try to go in bin directory and run the following commands:

ldd libnvrtc.so

and post here the output. It will show how/where the .so dependencies are resolved.

johannes.wilde wrote: ↑Wed Jan 13, 2021 9:33 am Is it relevant, that i don't use a installed version of Blender but a daily build?

Have you tried with an official release (like 2.83) ? Do you have the same problem ?

Here is the output from
ldd libnvrtc.so in ./bin
On Blender 2.83 Optix isn't working at all. It fails in Cycles as well.
On Blender 2.91 Cycles runs with Optix, but in Luxcore i get the same Error.

johannes.wilde wrote: ↑Wed Jan 13, 2021 12:00 pm On Blender 2.83 Optix isn't working at all. It fails in Cycles as well.
On Blender 2.91 Cycles runs with Optix, but in Luxcore i get the same Error.

This is quite strange and I can not think a reason why it is happening. You could try to install the GPU drivers from NVIDIA if you are using the one included in your Linux distribution but it is very hard to say if it can be of any help.

Hi,
i now completely removed the systems Nvidia Driver and installed the latest Driver from the Nvidia homepage (455.46.04)
But still i get the same error message

[LuxRays][28.786] [MergeSampleBuffersOCL] CUDA program compilation error:
nvrtc: error: failed to open libnvrtc-builtins.so.
Make sure that libnvrtc-builtins.so is installed correctly.