Machine learning and AI kernels implemented in various languages and frameworks.
General information and links for each framework are located in the vector_add subdirectories.
- vector_add - The simplest kernel that adds two vectors
- dot - Dot product of two vectors
- gemv - Matrix-vector multiplication
- gemm - Matrix-matrix multiplication
- latency - Microbenchmark to measure memory latency
There are links to additional technical information about GPU topics in the info directory.