Tom Scogland
Tom Scogland
These are grouped together because I found the cmake bug in spack due to the rocprim dependency of raja. It seems rocprim needs cmake 3.16, while I have 3.14 in...
build the world test for #31739
In discussing updated sycl testing in RAJA, it came up that we currently don't have a way to specify what the target architectures are when enabling portable targets like SYCL...
For example, commenting out the 3 reorderings and switches of loop/seq in the SequentialKernelLocExecPols list in test-kernel-reduceloc.cpp.in reduces compile time for the Min2DView-Sequential object file from ~37 seconds to ~7...
This is a draft of new range ergonomics. I did it for now as new overloads of `make_range` because we already have that, but I'm tempted to do something else...
The full summary is above. - Adds macros to identify the cuda compiler currently in use (see `RAJA_CUDA_COMPILER_*` - Adds a workaround for the unsigned long long __shfl(_sync)? intrinsics that...
``` :1:1: warning: umbrella header for module 'RAJA' does not include header '/util/SoAArray.hpp' [-Wincomplete-umbrella] :1:1: warning: umbrella header for module 'RAJA' does not include header '/policy/cuda/scan.hpp' [-Wincomplete-umbrella] :1:1: warning: umbrella...
Everything looks good overall. This may be for a future PR, but it seems like the LaunchContext can encapsulate Teams() and Threads(). The LaunchContext gets passed all the way into...
In the OpenMP target backend we should be mapping the lambdas, but as in PR #912 we currently first privatize because of incorrect handling of the mapping of captured pointers....
We have had some support for using clang style modules support for some time, but that support has atrophied to the point that it no longer builds a working RAJA....