Level-3 BLAS on a GPU: Picking the low hanging fruit

The arrival of hardware accelerators has created a new gold rush to be the first to deliver their promise of high performance for numerical applications. Despite the recent advances in programmability, it is still hard to develop tuned programs that extract all the potential performance promised by the manufacturers. In this paper we remind the community that while this development effort is a noble endeavor, there is a lot of low hanging fruit that can be harvested easily. Picking this low hanging fruit benefits the scientific computing community immediately and prototypes the approach that further optimizations may follow. We demonstrate this by focusing on a widely used set of operations, the level-3 BLAS, targeting the NVIDIA GPUs.