used reinterpret_cast<uint32_t*> for numBlocks, as expected by hipOccupancyMaxActiveBlocksPerMultiprocessor() api. [ROCm/hip commit: 4d472b777f]
4d472b777f
Simple test for hipLaunchCooperativeKernelMultiDevice API. [ROCm/hip commit: 48d1f766d5]
48d1f766d5