luxcorerender / luxmark Goto Github PK

LuxMark OpenCL benchmark source repository

License: GNU General Public License v3.0

CMake 9.51% C++ 21.54% Shell 0.05% C 41.37% Ruby 1.10% CSS 9.05% JavaScript 0.10% HTML 0.35% PHP 0.80% SCSS 15.30% Assembly 0.83%

luxmark's People

Contributors

Stargazers

Watchers

Forkers

maromero3d bacilladoro stlvnub lyx634449800 umadevimcw illwieckz

luxmark's Issues

Does LuxMark support CPU with ARM arch?

The binary of LuxMark only support X86 arch CPU.
I tried to compile LuxMark with source code 3.1 but failed. I have two problems:

Can the source code be supported to compile on CPU with ARM arch?
I complied failed because of lack of LuxRay. I search the website and find several packages of LuxRay. I'm not sure which one to install. Is there any introduction to install LuxMark and the way to install dependent packages?

LuxMark crashes

I try to follow https://blenderartists.org/forum/showthread.php?448186-LuxCoreRender-2-0&p=3303419&viewfull=1#post3303419 to solve ImportError: DLL load failed: The specified module could not be found which I stated in https://blenderartists.org/forum/showthread.php?448186-LuxCoreRender-2-0&p=3303040&viewfull=1#post3303040

I use Geforce 740M on win7 64bits, and update the Nvidia driver to the newest 391.35-notebook-win8-win7-64bit-international-whql

I just run LuxMark, as we can see it choose both embeded Intel HD Graphics 4000 and Geforce GT 740M. Some messages are shown as above, then LuxMark crashes

does this mean my Geforce GT 740M is too old and can not run luxmark/luxrender?

2018-04-21 21:58:52 - LuxMark v3.1
2018-04-21 21:58:52 - Based on LuxCore v1.5
2018-04-21 21:58:52 - [LuxRays] OpenCL Platform 0: Intel(R) Corporation
2018-04-21 21:58:52 - [LuxRays] OpenCL Platform 1: NVIDIA Corporation
2018-04-21 21:58:52 - [LuxRays] Device 0 name: NativeThread
2018-04-21 21:58:52 - [LuxRays] Device 0 type: NATIVE_THREAD
2018-04-21 21:58:52 - [LuxRays] Device 0 compute units: 1
2018-04-21 21:58:52 - [LuxRays] Device 0 preferred float vector width: 4
2018-04-21 21:58:52 - [LuxRays] Device 0 max allocable memory: 0MBytes
2018-04-21 21:58:52 - [LuxRays] Device 0 max allocable memory block size: 0MBytes
2018-04-21 21:58:52 - [LuxRays] Device 1 name: Intel(R) Core(TM) i5-3337U CPU @ 1.80GHz
2018-04-21 21:58:52 - [LuxRays] Device 1 type: OPENCL_CPU
2018-04-21 21:58:52 - [LuxRays] Device 1 compute units: 4
2018-04-21 21:58:52 - [LuxRays] Device 1 preferred float vector width: 4
2018-04-21 21:58:52 - [LuxRays] Device 1 max allocable memory: 16276MBytes
2018-04-21 21:58:52 - [LuxRays] Device 1 max allocable memory block size: 4069MBytes
2018-04-21 21:58:52 - [LuxRays] Device 2 name: Intel(R) HD Graphics 4000
2018-04-21 21:58:52 - [LuxRays] Device 2 type: OPENCL_GPU
2018-04-21 21:58:52 - [LuxRays] Device 2 compute units: 16
2018-04-21 21:58:52 - [LuxRays] Device 2 preferred float vector width: 1
2018-04-21 21:58:52 - [LuxRays] Device 2 max allocable memory: 1624MBytes
2018-04-21 21:58:52 - [LuxRays] Device 2 max allocable memory block size: 406MBytes
2018-04-21 21:58:52 - [LuxRays] Device 3 name: GeForce GT 740M
2018-04-21 21:58:52 - [LuxRays] Device 3 type: OPENCL_GPU
2018-04-21 21:58:52 - [LuxRays] Device 3 compute units: 2
2018-04-21 21:58:52 - [LuxRays] Device 3 preferred float vector width: 1
2018-04-21 21:58:52 - [LuxRays] Device 3 max allocable memory: 2048MBytes
2018-04-21 21:58:52 - [LuxRays] Device 3 max allocable memory block size: 512MBytes
2018-04-21 21:58:52 - [LuxCore] [0.265] Configuration: 
2018-04-21 21:58:52 - [LuxCore] [0.265] renderengine.type = "PATHOCL"
2018-04-21 21:58:52 - [LuxCore] [0.265] opencl.platform.index = "-1"
2018-04-21 21:58:52 - [LuxCore] [0.265] opencl.cpu.use = 0
2018-04-21 21:58:52 - [LuxCore] [0.265] opencl.gpu.use = 1
2018-04-21 21:58:52 - [LuxCore] [0.265] opencl.gpu.workgroup.size = "64"
2018-04-21 21:58:52 - [LuxCore] [0.265] scene.epsilon.min = "9.99999971718068536574719e-10"
2018-04-21 21:58:52 - [LuxCore] [0.265] scene.epsilon.max = "0.100000001490116119384766"
2018-04-21 21:58:52 - [LuxCore] [0.265] accelerator.instances.enable = "0"
2018-04-21 21:58:52 - [LuxCore] [0.265] path.maxdepth = "12"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.width = "800"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.height = "800"
2018-04-21 21:58:52 - [LuxCore] [0.265] sampler.type = "RANDOM"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.filter.type = "GAUSSIAN"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.filter.xwidth = "1"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.filter.ywidth = "1"
2018-04-21 21:58:52 - [LuxCore] [0.265] scene.file = "scenes/luxball/scene.scn"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.imagepipeline.0.type = "TONEMAP_LINEAR"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.imagepipeline.1.type = "GAMMA_CORRECTION"
2018-04-21 21:58:52 - [LuxCore] [0.265] film.imagepipeline.1.value = "2.2"
2018-04-21 21:58:52 - [LuxCore] [0.265] screen.refresh.interval = 2000
2018-04-21 21:58:52 - [LuxCore] [0.265] opencl.kernel.options = "-cl-fast-relaxed-math -cl-mad-enable -cl-no-signed-zeros"
2018-04-21 21:58:52 - [SDL] [0.265] Reading scene: scenes/luxball/scene.scn
2018-04-21 21:58:52 - [SDL] [0.265] Camera type: perspective
2018-04-21 21:58:52 - [SDL] [0.265] Camera position: Point[0.649247, 0.279196, 0.148695]
2018-04-21 21:58:52 - [SDL] [0.265] Camera target: Point[0.10751, 1.04879, -0.189305]
2018-04-21 21:58:52 - [SDL] [0.265] Camera clipping plane disabled
2018-04-21 21:58:52 - [SDL] [0.265] Texture definition: texture-0x7f4e0b479080
2018-04-21 21:58:52 - [SDL] [0.265] Reading texture map: scenes/luxball/imagemap-00001.exr
2018-04-21 21:58:52 - [SDL] [0.343] Texture definition: texture-0x7f4e0b47ab80
2018-04-21 21:58:52 - [SDL] [0.343] Material definition: material-0x7f4e0b4790b0
2018-04-21 21:58:52 - [SDL] [0.343] Material definition: material-0x7f4e0b479170
2018-04-21 21:58:52 - [SDL] [0.343] Material definition: material-0x7f4e0b4a7a50
2018-04-21 21:58:52 - [SDL] [0.343] Material definition: material-0x7f4e0b47a5e0
2018-04-21 21:58:52 - [SDL] [0.437] Scene objects count: 7
2018-04-21 21:58:52 - [SDL] [0.437] Light definition: infinitelight
2018-04-21 21:58:52 - [SDL] [0.437] Reading texture map: scenes/luxball/imagemap-00000.exr
2018-04-21 21:58:52 - [LuxCore] [0.640] Film resolution: 800x800
2018-04-21 21:58:52 - [LuxRays] [0.640] OpenCL Platform 0: Intel(R) Corporation
2018-04-21 21:58:52 - [LuxRays] [0.640] OpenCL Platform 1: NVIDIA Corporation
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 name: NativeThread
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 type: NATIVE_THREAD
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 compute units: 1
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 preferred float vector width: 4
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 max allocable memory: 0MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 0 max allocable memory block size: 0MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 name: Intel(R) Core(TM) i5-3337U CPU @ 1.80GHz
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 type: OPENCL_CPU
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 compute units: 4
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 preferred float vector width: 4
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 max allocable memory: 16276MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 1 max allocable memory block size: 4069MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 name: Intel(R) HD Graphics 4000
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 type: OPENCL_GPU
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 compute units: 16
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 preferred float vector width: 1
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 max allocable memory: 1624MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 2 max allocable memory block size: 406MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 name: GeForce GT 740M
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 type: OPENCL_GPU
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 compute units: 2
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 preferred float vector width: 1
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 max allocable memory: 2048MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Device 3 max allocable memory block size: 512MBytes
2018-04-21 21:58:52 - [LuxRays] [0.640] Preprocessing DataSet
2018-04-21 21:58:52 - [LuxRays] [0.640] Total vertex count: 124324
2018-04-21 21:58:52 - [LuxRays] [0.640] Total triangle count: 217544
2018-04-21 21:58:52 - [LuxRays] [0.655] Preprocessing DataSet done
2018-04-21 21:58:52 - [LuxRays] [0.687] Creating 2 intersection device(s)
2018-04-21 21:58:52 - [LuxRays] [0.687] Allocating intersection device 0: Intel(R) HD Graphics 4000 (Type = OPENCL_GPU)
2018-04-21 21:58:52 - [LuxRays] [0.687] Allocating intersection device 1: GeForce GT 740M (Type = OPENCL_GPU)
2018-04-21 21:58:52 - [LuxCore] [0.687] OpenCL Devices used:
2018-04-21 21:58:52 - [LuxCore] [0.687] [Intel(R) HD Graphics 4000Intersect]
2018-04-21 21:58:52 - [LuxCore] [0.687] Device OpenCL version: OpenCL 1.1 
2018-04-21 21:58:52 - [LuxCore] [0.687] [GeForce GT 740MIntersect]
2018-04-21 21:58:52 - [LuxCore] [0.687] Device OpenCL version: OpenCL 1.2 CUDA
2018-04-21 21:58:52 - [LuxRays] [0.687] Adding DataSet accelerator: BVH
2018-04-21 21:58:52 - [LuxRays] [0.687] Total vertex count: 124324
2018-04-21 21:58:52 - [LuxRays] [0.687] Total triangle count: 217544
2018-04-21 21:58:52 - [LuxRays] [0.702] BVH Dataset preprocessing time: 15ms
2018-04-21 21:58:52 - [LuxRays] [0.702] Building BVH, primitives: 217544
2018-04-21 21:58:52 - [LuxRays] [0.843] BVH build hierarchy time: 140ms
2018-04-21 21:58:52 - [LuxRays] [0.843] Pre-processing BVH, total nodes: 324654
2018-04-21 21:58:52 - [LuxRays] [0.874] BVH build array time: 30ms
2018-04-21 21:58:52 - [LuxRays] [0.874] Total BVH memory usage: 10145Kbytes
2018-04-21 21:58:52 - [LuxRays] [0.874] BVH build time: 186ms
2018-04-21 21:58:52 - [LuxCore] [0.874] Configuring 2 CPU render threads
2018-04-21 21:58:52 - [LuxRays] [0.889] [OpenCL device::Intel(R) HD Graphics 4000Intersect] Vertices buffer size (Page 0, 124324 vertices): 1456Kbytes
2018-04-21 21:58:52 - [LuxRays] [0.905] [OpenCL device::Intel(R) HD Graphics 4000Intersect] BVH buffer size (Page 0, 324654 nodes): 10145Kbytes
2018-04-21 21:58:53 - [LuxRays] [2.013] [OpenCL device::GeForce GT 740MIntersect] Vertices buffer size (Page 0, 124324 vertices): 1456Kbytes
2018-04-21 21:58:53 - [LuxRays] [2.028] [OpenCL device::GeForce GT 740MIntersect] BVH buffer size (Page 0, 324654 nodes): 10145Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.044] [PathOCLRenderEngine] OpenCL task count: 262144
2018-04-21 21:58:53 - [LuxCore] [2.044] [PathOCLBaseRenderEngine] OpenCL max. page memory size: 415744Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.044] Compile Geometry
2018-04-21 21:58:53 - [LuxCore] [2.059] Scene geometry compilation time: 15ms
2018-04-21 21:58:53 - [LuxCore] [2.059] Compile 4 Textures
2018-04-21 21:58:53 - [LuxCore] [2.059] Textures compilation time: 0ms
2018-04-21 21:58:53 - [LuxCore] [2.059] Compile 4 Materials
2018-04-21 21:58:53 - [LuxCore] [2.059] Material compilation time: 0ms
2018-04-21 21:58:53 - [LuxCore] [2.059] Compile Lights
2018-04-21 21:58:53 - [LuxCore] [2.153] Lights compilation time: 93ms
2018-04-21 21:58:53 - [LuxCore] [2.153] Compile ImageMaps
2018-04-21 21:58:53 - [LuxCore] [2.169] Image maps page(s) count: 1
2018-04-21 21:58:53 - [LuxCore] [2.169] RGB channel page 0 size: 35788Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.169] Texture maps compilation time: 16ms
2018-04-21 21:58:53 - [LuxCore] [2.169] Starting 2 PathOCL render threads
2018-04-21 21:58:53 - [LuxCore] [2.184] [PathOCLBaseRenderThread::0] RADIANCE_PER_PIXEL_NORMALIZEDs[0] buffer size: 10000Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] RADIANCE_PER_PIXEL_NORMALIZEDs[0] buffer size: 10000Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] Camera buffer size: 5444bytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] Normals buffer size: 1456Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] UVs buffer size: 832Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] Vertices buffer size: 1456Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.200] [PathOCLBaseRenderThread::0] Triangles buffer size: 2549Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.215] [PathOCLBaseRenderThread::0] Mesh description buffer size: 1064bytes
2018-04-21 21:58:53 - [LuxCore] [2.215] [PathOCLBaseRenderThread::0] ImageMap descriptions buffer size: 48bytes
2018-04-21 21:58:53 - [LuxCore] [2.215] [PathOCLBaseRenderThread::0] ImageMaps buffer size: 35788Kbytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Textures buffer size: 1088bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Materials buffer size: 480bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Mesh material index buffer size: 28bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Lights buffer size: 292bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Env. light indices buffer size: 4bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] Light offsets buffer size: 28bytes
2018-04-21 21:58:53 - [LuxCore] [2.231] [PathOCLBaseRenderThread::0] InfiniteLight distributions buffer size: 65600Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLBaseRenderThread::0] LightsDistribution buffer size: 16bytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLBaseRenderThread::0] Ray buffer size: 12288Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLBaseRenderThread::0] RayHit buffer size: 5120Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLRenderThread::0] Size of a GPUTask: 12bytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLBaseRenderThread::0] GPUTask buffer size: 3072Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLRenderThread::0] Size of a GPUTask DirectLight: 64bytes
2018-04-21 21:58:54 - [LuxCore] [2.247] [PathOCLBaseRenderThread::0] GPUTaskDirectLight buffer size: 16384Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.262] [PathOCLRenderThread::0] Size of a GPUTask State: 116bytes
2018-04-21 21:58:54 - [LuxCore] [2.262] [PathOCLBaseRenderThread::0] GPUTaskState buffer size: 29696Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.262] [PathOCLBaseRenderThread::0] GPUTask Stats buffer size: 1024Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.262] [PathOCLRenderThread::0] Size of a Sample: 32bytes
2018-04-21 21:58:54 - [LuxCore] [2.262] [PathOCLBaseRenderThread::0] Sample buffer size: 8192Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.278] [PathOCLRenderThread::0] Sample dimensions: 89
2018-04-21 21:58:54 - [LuxCore] [2.278] [PathOCLRenderThread::0] Size of a SampleData: 8bytes
2018-04-21 21:58:54 - [LuxCore] [2.278] [PathOCLBaseRenderThread::0] SampleData buffer size: 2048Kbytes
2018-04-21 21:58:54 - [LuxCore] [2.278] [PathOCLBaseRenderThread::0] Defined symbols: -D LUXRAYS_OPENCL_KERNEL -D SLG_OPENCL_KERNEL -D RENDER_ENGINE_PATHOCL -D PARAM_RAY_EPSILON_MIN=1.000000e-009f -D PARAM_RAY_EPSILON_MAX=1.000000e-001f -D PARAM_LIGHT_WORLD_RADIUS_SCALE=1.000000e+001f -D PARAM_ACCEL_BVH -D PARAM_FILM_RADIANCE_GROUP_0 -D PARAM_FILM_RADIANCE_GROUP_COUNT=1 -D PARAM_HAS_NORMALS_BUFFER -D PARAM_HAS_UVS_BUFFER -D PARAM_ENABLE_TEX_CONST_FLOAT3 -D PARAM_ENABLE_TEX_IMAGEMAP -D PARAM_ENABLE_MAT_MATTE -D PARAM_ENABLE_MAT_MIRROR -D PARAM_CAMERA_TYPE=0 -D PARAM_CAMERA_HAS_DOF -D PARAM_HAS_INFINITELIGHT -D PARAM_TRIANGLE_LIGHT_COUNT=0 -D PARAM_LIGHT_COUNT=1 -D PARAM_HAS_INFINITELIGHTS -D PARAM_HAS_ENVLIGHTS -D PARAM_HAS_IMAGEMAPS -D PARAM_IMAGEMAPS_PAGE_0 -D PARAM_IMAGEMAPS_COUNT=1 -D PARAM_DEVICE_INDEX=0 -D PARAM_DEVICE_COUNT=2 -cl-fast-relaxed-math -cl-mad-enable -cl-no-signed-zeros -D PARAM_TASK_COUNT=262144 -D PARAM_MAX_PATH_DEPTH=12 -D PARAM_RR_DEPTH=3 -D PARAM_RR_CAP=5.000000e-001f -D PARAM_RADIANCE_CLAMP_MAXVALUE=0.000000e+000f -D PARAM_PDF_CLAMP_VALUE=0.000000e+000f -D PARAM_IMAGE_FILTER_TYPE=2 -D PARAM_IMAGE_FILTER_WIDTH_X=1.000000e+000f -D PARAM_IMAGE_FILTER_WIDTH_Y=1.000000e+000f -D PARAM_IMAGE_FILTER_PIXEL_WIDTH_X=1 -D PARAM_IMAGE_FILTER_PIXEL_WIDTH_Y=1 -D PARAM_IMAGE_FILTER_GAUSSIAN_ALPHA=2.000000e+000f -D PARAM_SAMPLER_TYPE=0
2018-04-21 21:58:54 - [LuxCore] [2.278] [PathOCLBaseRenderThread::0] Compiling kernels

Hi,
seems the latest prelease doesn't have prebuilt binaries in contrast to 4.0alpha0.. any hope of getting it on releases page.. mainly was wanting to see advancements/changes over the older alpha0 that was shipped over a year before than alpha1..

Luxmark 4.0 alpha Mac support (&binaries)?

Luxmark 2.0 how to download / compile

Hello ,

ia am looking for this program with "sala" scene to download / compile for MacOSX.

Thank you

Rendering issues with AMD GPUs due to unsafe default settings

The current default settings for OpenCL cause rendering issues on AMD GPUs. These issues occur for multiple driver versions since at least 19.12.2 to current with several different GPUs (RX 480, Vega56, RX 5700XT and RX 6800.) The issues stem from the default setting -cl-fast-relaxed-math. Source

This setup explicitly allows the compiler to perform unsafe and not standard-compliant math optimizations, and as such this should not be the default setting.
Disabling the setting fixes the rendering issues.

Default:

-cl-fast-relaxed-math disabled:

LuxMark v3.1hangs with render scene missing when chosing Native C++ on Linux

Hello

Whenever I choose Native C++ in my Linux installation, the render scene never shows up, despite my CPU going to 100%. The CPU stays at 100% even after the 2 minutes for a benchmark. I tried all 3 scenes (hotel lobby, Neumann and LuxBall, same result. Tried with Fedora 38 with Wayland and openSUSE Tumbleweed 2023-07-18 with X server. Selecting BENCHMARK_NATIVE on the command line shows the same issue. The text output in LuxMark window shows no error and stops at:

2023-07-23 21:05:31 - [LuxRays] [0.772] Total vertex count: 124324
2023-07-23 21:05:31 - [LuxRays] [0.772] Total triangle count: 217544

I believe this used to work some time ago, and now despite I use the same LuxMark v3.1, it is not working any more ... I used a fresh Linux installation, native (no VM), and the render scene never shows up.

It works as expected if I choose an OpenCL device, but not for Native C++. I have an AMD Ryzen Threadripper 1920X CPU.

LuxMark 3.1 not print result to stdout with --single-run option

windows 10 x64 OS, LuxMark release v3.1
open cmd windows, run command line:

luxmark.exe --scene=LUXBALL_HDR --mode=BENCHMARK_OCL_GPU --single-run --ext-info
I found no score result is printed after benchmark finish and luxmark window closed.

LuxMark 4.0 alpha on Linux amd64 crashing

OpenCL on GPU only mode.

After starting, showing up the window and detecting devices it works a bit, I can see scene loaded and BVH being build, but it crashes after few seconds:

user@debian:~/lux/luxmark$ ROOT=/home/user/lux/luxmark
user@debian:~/lux/luxmark$ export LD_PRELOAD="$ROOT/lib/libembree3.so.3:$ROOT/lib/libOpenImageDenoise.so.0:$ROOT/lib/libtbb.so.2:$ROOT/lib/libtbbmalloc.so.2"
user@debian:~/lux/luxmark$ gdb --args ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_GPU 
GNU gdb (Debian 9.2-1) 9.2
Type "apropos word" to search for commands related to "word"...
Reading symbols from ./luxmark.bin...
(No debugging symbols found in ./luxmark.bin)
(gdb) r
Starting program: /home/user/lux/luxmark/luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_GPU
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffec80a700 (LWP 2174629)]
...
[Thread 0x7ffe4cff9700 (LWP 2174774) exited]
--Type <RET> for more, q to quit, c to continue without paging--

Thread 1 "luxmark.bin" received signal SIGSEGV, Segmentation fault.
0x0000000000000000 in ?? ()
(gdb) bt
#0  0x0000000000000000 in  ()
#1  0x0000555555eaceed in luxrays::OpenCLIntersectionDevice::Stop() ()
#2  0x0000555555eacf55 in luxrays::OpenCLIntersectionDevice::~OpenCLIntersectionDevice() ()
#3  0x0000555555eacf69 in luxrays::OpenCLIntersectionDevice::~OpenCLIntersectionDevice() ()
#4  0x0000555555e589ba in luxrays::Context::~Context() ()
#5  0x00005555559908c2 in slg::RenderEngine::~RenderEngine() ()
#6  0x0000555555b351ff in slg::PathOCLBaseRenderEngine::~PathOCLBaseRenderEngine() ()
#7  0x00005555559b97e9 in slg::PathOCLRenderEngine::~PathOCLRenderEngine() ()
#8  0x0000555555a998e5 in slg::RenderSession::~RenderSession() ()
#9  0x000055555593a665 in luxcore::detail::RenderSessionImpl::~RenderSessionImpl() ()
#10 0x000055555593a6a9 in luxcore::detail::RenderSessionImpl::~RenderSessionImpl() ()
#11 0x000055555591eb13 in LuxCoreRenderSession::Stop() ()
#12 0x000055555591eba5 in LuxCoreRenderSession::~LuxCoreRenderSession() ()
#13 0x00005555558efb4a in LuxMarkApp::Stop() ()
#14 0x00005555558efc66 in LuxMarkApp::InitRendering(LuxMarkAppMode, char const*) ()
#15 0x0000555555900802 in MainWindow::event(QEvent*) ()
#16 0x00007ffff270fc32 in QApplicationPrivate::notify_helper(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Widgets.so.5
#17 0x00007ffff2719190 in QApplication::notify(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Widgets.so.5
#18 0x00007ffff1ac3a52 in QCoreApplication::notifyInternal2(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#19 0x00007ffff1ac6648 in QCoreApplicationPrivate::sendPostedEvents(QObject*, int, QThreadData*) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#20 0x00007ffff1b1a183 in  () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#21 0x00007ffff0aa260d in g_main_context_dispatch () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#22 0x00007ffff0aa2890 in  () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#23 0x00007ffff0aa291f in g_main_context_iteration () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#24 0x00007ffff1b197c1 in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#25 0x00007ffff1ac26db in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#26 0x00007ffff1aca182 in QCoreApplication::exec() () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#27 0x000055555588026e in main ()
(gdb)

I am using AMD Radeon R9 Fury X with ROCm 3.5 installed and in my OpenCL OCL icd.

The LuxMark however does selects both Mesa Clover OpenCL implementation (which as of version Mesa 20.0.0 is actually working extremely well in other apps!) and also AMD APP opencl implementation.

I do not have pocl (CPU based OpenCL implementation) installed so that is not a problem. (And when trying to use it with --mode=BENCHMARK_OCL_CPU the luxmark report back an error that it can't find any CPU based OpenCL devices).

I tried also disabling devices, and it doesn't help:

user@debian:~/lux/luxmark$ ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_CUSTOM --devices=1
Segmentation fault
user@debian:~/lux/luxmark$ ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_CUSTOM --devices=00
Segmentation fault
user@debian:~/lux/luxmark$ ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_CUSTOM --devices=01
Segmentation fault
user@debian:~/lux/luxmark$ ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_CUSTOM --devices=10
Segmentation fault
user@debian:~/lux/luxmark$ ./luxmark.bin --scene=FOOD --mode=BENCHMARK_OCL_CUSTOM --devices=11
Segmentation fault

I can see (before the crash) proper devices being selected on the right side, but it still crashes.

My clinfo in the attachment.
clinfo.txt

LuxMark v3.1 crash for OpenCL platform with 0 devices

Hello

LuxMark v3.1 always had a crash on start when I have an OpenCL platform with no devices. Lately this can happen for example with the new rusticl platform enabled by Mesa. But I have other platforms with no devices as well, like mesa when they do not support my GPU.

This is annoying users for a long time, can it be fixed please ?

SIGABRT during Luxmark 3.1 execution

I am copying KhronosGroup/SPIRV-LLVM-Translator#404 to LuxMark repo.
Luxmark crashed on std::stringstream destructor https://github.com/KhronosGroup/SPIRV-LLVM-Translator/blob/llvm_release_110/lib/SPIRV/OCLToSPIRV.cpp#LL321C7-L323C24). The std::stringstream usage is right, it failed with "vload3" string process but passed with "vload2". Dumped LLVM IR could be translated using llvm-spirv and opencl kernel compiled with ocloc. We suspect problems are with Luxmark itself and maybe Luxmark binary overwrites some functions from libstdc++.

Project	Branch	Revision
opencl-clang	ocl-open-110	ee31812ea8b89d08c2918f045d11a19bd33525c5
SPIRV-LLVM-Translator	llvm_release_110	d739c01d65ec00dee64dedd40deed805216a7193
LLVM project	release/11.x	1fdec59bffc11ae37eb51a1b9869f0696bfd5312

Doesn't provide useful message when clError is thrown

LuxMark v4.0 alpha0 on Linux amd64.

I installed pocl just for kicks, and tried running luxmark.

It does build the scene, BVH and compiles all the kernels fine, but then aborts:

terminate called after throwing an instance of 'cl::Error'
  what():  clCreateBuffer
--Type <RET> for more, q to quit, c to continue without paging--

Thread 1 "luxmark.bin" received signal SIGABRT, Aborted.
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
50	../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bt
#0  __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#1  0x00007ffff150855b in __GI_abort () at abort.c:79
#2  0x0000555556a76265 in __gnu_cxx::__verbose_terminate_handler() ()
#3  0x00005555569d8dd6 in __cxxabiv1::__terminate(void (*)()) ()
#4  0x00005555569d8e21 in  ()
#5  0x00005555569da344 in  ()
#6  0x00005555558796f4 in cl::detail::errHandler(int, char const*) [clone .part.34] ()
#7  0x0000555555eac028 in luxrays::OpenCLIntersectionDevice::OpenCLDeviceQueue::OpenCLDeviceQueueElem::OpenCLDeviceQueueElem(luxrays::OpenCLIntersectionDevice*, cl::CommandQueue*, unsigned int, unsigned long) ()
#8  0x0000555555eb0e92 in luxrays::OpenCLIntersectionDevice::OpenCLDeviceQueue::OpenCLDeviceQueue(luxrays::OpenCLIntersectionDevice*, unsigned int, unsigned long) ()
#9  0x0000555555eb11d6 in luxrays::OpenCLIntersectionDevice::Start() ()
#10 0x0000555555e57296 in luxrays::Context::Start() ()
#11 0x0000555555d446ab in slg::Film::CreateOCLContext() ()
#12 0x0000555555d3d7d0 in slg::Film::ExecuteImagePipelineImpl(unsigned int) ()
#13 0x0000555555d3a8cd in float const* slg::Film::GetChannel<float>(slg::Film::FilmChannelType, unsigned int, bool) ()
#14 0x00005555559398c4 in luxcore::detail::FilmImpl::GetChannelFloat(luxcore::Film::FilmChannelType, unsigned int, bool) ()
#15 0x000055555591ee97 in LuxCoreRenderSession::UpdateFrameBuffer(unsigned int) ()
#16 0x00005555558f16b7 in LuxMarkApp::RenderRefreshTimeout() ()
#17 0x00007ffff1aee388 in QMetaObject::activate(QObject*, int, int, void**) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#18 0x00007ffff1afb10a in QTimer::timeout(QTimer::QPrivateSignal) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#19 0x00007ffff1aeed45 in QObject::event(QEvent*) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#20 0x00007ffff270fc32 in QApplicationPrivate::notify_helper(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Widgets.so.5
#21 0x00007ffff2719190 in QApplication::notify(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Widgets.so.5
#22 0x00007ffff1ac3a52 in QCoreApplication::notifyInternal2(QObject*, QEvent*) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#23 0x00007ffff1b18b90 in QTimerInfoList::activateTimers() () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#24 0x00007ffff1b1944c in  () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#25 0x00007ffff0aa260d in g_main_context_dispatch () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#26 0x00007ffff0aa2890 in  () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#27 0x00007ffff0aa291f in g_main_context_iteration () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#28 0x00007ffff1b197c1 in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#29 0x00007ffff1ac26db in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#30 0x00007ffff1aca182 in QCoreApplication::exec() () at /lib/x86_64-linux-gnu/libQt5Core.so.5
#31 0x000055555588026e in main ()
(gdb) q

Would be nice if the int and char const* part of the error were displayed back to stderr, so I can investigate deeper, before abort is called.

Due to lack of debug symbols I couldn't debug it too far with precompiled binaries. errno had a value 11 (EAGAIN), but not sure if this is related.

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

Luxmark does not check for errors returned by driver when it fails to allocate memory. This causes the app to crash. Typically seen with limited GPU memory ( ~1 GB) in VM setup.

When FB is restricted to make available free video memory as ~160Mib, luxmark app is crashing on win7, win8.1 and win10.
In all the cases nvidia driver is returning proper error (not able to allocate memory). Following is the debugging info
from eventviewer and windbg logs when the application crashes.

Config Used:
Quadro GV100 + cuda_a-190502-26255178.

On Win7 and win8.1,
Event viewer and windbg shows faulting module as luxmark.exe.
BUGCHECK_STR INVALID_POINTER_READ.
ExceptionCode: c0000005 (Access violation).
Full crashdump is available at \devrel\share\ngubba\luxmark_crash\luxmark.exe.5272.

On Win10 RS4
Event viewer and windbg shows faulting module as MSVCR120.dll.
BUGCHECK_STR: FAIL_FAST_FATAL_APP_EXIT.
ExceptionCode: c0000409 (Security check failure or stack buffer overrun).
Full crashdump is available at \devrel\share\ngubba\luxmark_crash\luxmark.exe.12656.

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

division by zero with default compilation options produced artefacts

We recently found a bug in Luxmark3.1

The kernel code has the following
min(1.f, lastPdfW / PARAM_PDF_CLAMP_VALUE);

While by default the variable PARAM_PDF_CLAMP_VALUE is set to 0 during OpenCL kernel compilation time, see below the constants passed
....-D PARAM_RADIANCE_CLAMP_MAXVALUE=0.000000e+000f -D PARAM_PDF_CLAMP_VALUE=0.000000e+000f -D PARAM_IMAGE_FILTER_TYPE=2...

When -cl-fast-relaxed-math is enabled it can produce undefined behaviour as this compiler option triggers the -finite-math-only and the compiler is not supposed to handle division by zero (inf cases) properly

https://www.khronos.org/registry/OpenCL/specs/2.2/html/OpenCL_API.html#compiler-options
-cl-finite-math-only
Allow optimizations for floating-point arithmetic that assume that arguments and results are not NaNs, +Inf, -Inf. This option may violate the OpenCL numerical compliance requirements for single precision and double precision floating-point, as well as edge case behavior. The original and modified values are defined in the SPIR-V OpenCL environment specification

-cl-fast-relaxed-math
Sets the optimization options -cl-finite-math-only and -cl-unsafe-math-optimizations. This option causes the preprocessor macro FAST_RELAXED_MATH to be defined in the OpenCL program.

These options are actually the default for luxmark3.1

Luxmark 3.1x cmd line doesn't print any logs on Windows

As per documentation, Luxmark app should print stdout and stderr to console and exit. Currently, its not printing anything to the console.

--single-run (run the benchmark, print the result to the stdout and exit)

cmd used: start /WAIT /max luxmark.exe --scene="LUXBALL_HDR" --mode="BENCHMARK_OCL_CPUGPU" --single-run --ext-info