Danny Taller
Danny Taller
Currently maestro supports slurm and lsf, but not PBS . PBS is desirable for some customers
Hello, I am attempting to compile scalapack with intel oneapi 2022 compiler, with gcc headers from gcc 10.2 When compiling, I keep getting errors along the lines of: `error: call...
Add an input argument for care::IntersectArrays that allows the algorithm to skip array elements in the intersection
Add a care equivalent of std::count (see https://www.geeksforgeeks.org/std-count-cpp-stl/ ) to care
get warning about pinned memory, and a test fails
Some tests in LoopFuserTest.cxx and Benchmarks.cpp have been ifdefed out for CPU-only builds to avoid failures; fix this.
Either redefine CARE_CUDA_BLOCK_SIZE to be CARE_GPU_BLOCK_SIZE, or make separate hip and cuda variables
RAJADeviceExec is equal to RAJA::cuda_exec or RAJA::hip_exec or RAJA::omp_parallel_for_exec depending on how we compile. Using RAJADeviceExec everywhere would clean things up and make things more robust for different architectures
Write a function free to replace both host free and gpuFree (cudaFree / hipFree ) A function like cudaPointerGetAttributes (and similar function for HIP) can query whether a pointer is...
Two issues 1) Most functions in array_utils.h are declared at the top of the file, but some are declared somewhere in the middle. Move all function declarations to the top...