clMathLibraries / clSPARSE

a software library containing Sparse functions written in OpenCL
Apache License 2.0
173 stars 60 forks source link

"Kernel not found in cache" for SPMV sample #198

Open gpu opened 7 years ago

gpu commented 7 years ago

Again, apologies if this is issue is related to the build system being messed up on my side (although I doubt that this is the case)

After compiling the samples, trying to run the sample_spmv results in a message

kernel not found in cache

and a painful crash. The output when running the debug build of the sample:

C:\Develop\clMathLibraries\clSPARSE.build\clSPARSE-samples-build\staging\Debug>sample-spmv.exe C:\Develop\clMathLibraries\matrices\pde225.mtx
Executing sample clSPARSE SpMV (y = A*x) C++
Matrix will be read from: C:\Develop\clMathLibraries\matrices\pde225.mtx
Platform ID 0 : AMD Accelerated Parallel Processing
Platform ID 1 : NVIDIA CUDA

Getting devices from platform 0
Device ID 0 : Spectre
key: [control/control] -cl-kernel-arg-info -cl-std=CL1.2 -DWG_SIZE=256 hash = 174490369
kernel not found in cache: 174490369
Matrix: C:\Develop\clMathLibraries\matrices\pde225.mtx [nRow: 225] [nCol: 225] [nNZ: 1065]
[validateMemObject] Buffer size: 4,260 bytes. Required size: 4,260 bytes.
[validateMemObject] Buffer size: 4,260 bytes. Required size: 4,260 bytes.
[validateMemObject] Buffer size: 904 bytes. Required size: 904 bytes.
key: [csrmv_adaptive/csrmv_adaptive] -cl-kernel-arg-info -cl-std=CL1.2 -DROWBITS=32 -DWGBITS=24 -DVALUE_TYPE=float -DWG_SIZE=256 -DBLOCKSIZE=1024 -DBLOCK_MULTIPLIER=3 -DROWS_FOR_VECTOR=1 -DINDEX_TYPE=uint hash
= 4124998496
kernel not found in cache: 4124998496

Environment:

blueberry commented 7 years ago

I am also interested in this issue, since I would really like to try clSparse, but the building process is really messy EDIT: challenging. Is there a way to have a clear, step-by-step guide with the most common pitfalls?