Jerome Kieffer
Jerome Kieffer
Many things to be done regarding this format: * [ ] Implement write of uncompressed and TY5 compressed images * [ ] Move TY5 in compression module * [ ]...
only first frame got converted.
if possible ...
**Describe the bug** Very large (host) memory consumption has been observed when running OpenCL application in profiling mode. Example: Processing 10000 4Mpix images (int32) with ~6 kernels per image on...
This is not really an issue, it is just that I do not understand this warning generated by PoCL on device `pthread-haswell-Intel(R) Xeon(R) CPU E5-1650 v4 @ 3.60GHz`: ``` AVX...
Hi, I spotted a numerical precision issue with PoCL + CUDA. The tests are part of a non regression test of `pyFAI`, but the code is from the `silx` library...
Hi, I tried to compiled PoCL on power9 (ubuntu 20.04, LLVM 10 and 12) and it fails (compiling) while the same succeeds on x86_64 (same OS+LLVM) and in addition, all...
When the version of `numpy` installed contains the git tag (like in the https://anaconda.org/scientific-python-nightly-wheels), silx fails to build due to the version label which cannot be parsed.
I stumbled on this bug when setting up an environment for debugging/profiling OpenCL code (Portable CL + Cuda JIT compilation to profile on nvidia GPU). The error message printed before...
Changelog: LZ4 compressor on GPU * [ ] OpenCL code targeted to run on GPU * [X] Find position of interest * [X] Create segment descriptor (position in input, length...