Ömercan Yazici
Ömercan Yazici
Fixed compilation error when using AVX512F with MSVC, GCC and Clang. Added Intel SVML replacement (the original code) for users. The option SIMDPP_USE_SVML could be further used or dropped. Just...
JIT code generation triggers a lot of errors and produces (for some targets) artifacts while rendering, some targets (nvvm) do not even run. The bug appeared after updating AnyDSL (from...
GPU kernels get polluted by the `rv_all` instruction. The instruction will be filtered out only if Thorin is compiled with RV. This might not happen, as we should not expect...
The cmake-based setup is still a good workaround to get AnyDSL working on Windows. However, some configs are old. The following patch updates a single file such that it is...
Simple and small PR to ensure the tests can be configured via CMake and produce the intended behavior. Also added some missing test cases to the list. The list of...
Compiling AnyDSL on Windows with LLVM 16 and RV 16.x branch will fail to configure as `target_link_libraries` can not be applied to the **utility** target `RVPLUG`. https://github.com/cdl-saarland/rv/blob/bb21dd7b1e38c22495e5baa9268836c88e8b00ab/plugin/CMakeLists.txt#L12 Commenting out the...
This PR adds event support to AnyDSL runtime. Currently, only CPU & Cuda devices are supported. Support can be queried via `device_check_feature_support` and "event". Might add HSA support in the...
Add asynchronous copy operation `anydsl_copy_async`. The "async" is only a hint and only works on CUDA and OpenCL. Did not find a suitable method for HSA. CPU could have async,...
Get rid of RV for CPU backend and make it an experimental CPU backend besides the default one. This will simplify deployment, build and many more. The current performance of...