What is rocBLAS?

rocBLAS is the AMD library for Basic Linear Algebra Subprograms (BLAS) on the ROCm platform. It is implemented in the HIP C++ and optimized for AMD GPUs. For more detailed component information, see rocBLAS design notes.

rocBLAS aims to provide the following:

• Functionality similar to legacy BLAS, adapted to run on AMD GPUs

• A high-performance, robust implementation

rocBLAS is written in HIP C++ (using C++17 features) and it uses the HIP runtime.

The rocBLAS API is a thin C99 API that uses the hourglass pattern. It contains:

• rocBLAS Level-1 functions, rocBLAS Level-2 functions, and rocBLAS Level-3 functions with batched and strided_batched versions

• Extensions to legacy BLAS, including functions for mixed precision

• Auxiliary functions

• Device memory functions

Note

• The official rocBLAS API is the C99 API defined in rocblas.h. Therefore, the use of any other public symbols is discouraged. Other C/C++ interfaces might not follow a deprecation model and could change without warning from one release to the next.

• The rocBLAS array storage format is column major and one-based. This is to maintain compatibility with the legacy BLAS Fortran code.

• rocBLAS calls the AMD Tensile and hipBLASLt libraries for Level-3 GEMMs (matrix matrix multiplication).