AMD's GPUOpen HIP Project Made Progress Over The Summer
The HIP project has made good progress over the summer. HIP from AMD's GPUOpen project is part of the puzzle for converting CUDA to portable C++ code. That source code can then run on AMD GPUs while having little to no performance impact, at least according to AMD.
HIP developers have been working on stability improvements, packaging work, testing with the AMD testing suite, updates to match each ROCm release, and more. Some of the recent changes include FP16 software support, support for Hawaii discrete GPUs, support for hipArray, improved profiling, better clang-hipify utility, and support for dynamic shared memory allocations.
Developers have also made public a HIP developer-preview branch on GitHub. That developer-preview code has the latest fixes and features, including support for the CUDA Driver API as the latest addition.
You can learn more about this HIP work via GPUOpen.com or going straight to the GitHub repository.
HIP developers have been working on stability improvements, packaging work, testing with the AMD testing suite, updates to match each ROCm release, and more. Some of the recent changes include FP16 software support, support for Hawaii discrete GPUs, support for hipArray, improved profiling, better clang-hipify utility, and support for dynamic shared memory allocations.
Developers have also made public a HIP developer-preview branch on GitHub. That developer-preview code has the latest fixes and features, including support for the CUDA Driver API as the latest addition.
You can learn more about this HIP work via GPUOpen.com or going straight to the GitHub repository.
25 Comments