MAGMA Forum

Hi, I read the paper "Performance, Design, and Autotuning of Batched GEMM for GPUs", several concept such as Search Space Generation and Pruning
and Autotuning output anaylsis. Are these kind of process offline? and the aim is to get the best kernel for specific case?

Hi,
The current autotunning process is performed offline since it is done once per GPU type.
Thus we generate all acceptable kernel configuration and run them, and analyze the performance and choose the best for the target architecture.
Azzam