Henry Linjamäki
Henry Linjamäki
`clGetKernelInfo()` incorrectly reports that the `saxpy` kernel, as shown in the following SPIR-V listing, has eight kernel arguments instead of four - the original amount. ``` ; A test case...
Link to OpenCL reproducer: https://github.com/pocl/pocl/blob/master/tests/spirv/printf-kernels.cl#L55. Instead of printing the expected output: ``` hello % hello % world %s ``` It print something like this instead (output may vary): ``` hello...
bitonic-sort-sycl is a lot slower than the HIP version on chipStar targeting the same device through OpenCL. I guess, it's because sycl::queue creation initiates the SYCL runtime initialization while chipStar...
Make a lot of benchmarks to exit with error code 1 when the verification fails and make autohecbench.py skip result recording for the failing benchmarks. (This commit is extended version...
This draft is an updated version of https://github.com/KhronosGroup/OpenCL-Docs/pull/867. Notable changes: * Changed defined built-in kernels to use the new `cl_tensor` data type for data arguments. The data type is drafted...
This draft introduces an extension for tensor data type (`cl_tensor`) for storing N-dimensional data in implementation-defined memory layout. This extension is designed for the [defined built-in kernel extension](https://github.com/KhronosGroup/OpenCL-Docs/pull/1007) for starters....
A issue discovered while running HIP programs on OpenCL-BE->rusticl (with #830). Linking of __chip_atomic_add_f* symbols failed because the caller's and callee's function signature differed by their pointer parameters (pointee type...
Use [cl_ext_buffer_device_address](https://github.com/pocl/pocl/blob/1f80005557222a35f44a7d0d076aa4ae86499bdc/include/CL/cl_ext_pocl.h#L69) extension (experimental at the time of this commit) for managing HIP device allocations on devices where neither USM nor SVM is available. The latest main of POCL and,...
HIP kernels with bool parameters fail to run on rusticl. This is due to the produced SPIR-V being non-conformant respect to the OpenCL SPIR-V environment specification. The SPIR-V kernels have...
This patch introduces a new CUDA compiler - `cucc`. This is meant to replace `cuspvc` and aims to be a drop-in replacement for nvcc. It is still a work in...