= A fast GEMM implementation on a Cypress GPU =
by N.Nakasato (University of Aizu), submitted September 7, 2010.

We will present out results on this paper at 1st International Workshop on 
Performance Modeling, Benchmarking and Simulation of High Performance Computing Systems (PMBS 10)
held as part of SC10, New Orleans, November 13-19, 2010.

== abstract ==
We present benchmark results of optimized dense matrix multiplication
kernels for a Cypress GPU. We write general matrix multiply (GEMM) kernels
for single (SP), double (DP) and double-double (DDP) precision. 
Our SGEMM and DGEMM kernels show 73% and 87% of 
the theoretical performance of the GPU, respectively.
Currently, our SGEMM and DGEMM kernels are fastest 
with one GPU chip to our knowledge.
Furthermore, the performance of our matrix multiply kernel in DDP is 31 Gflop/s.
It is more than 200 times faster than the performance 
results on single core of a recent CPU (with mpack version 0.6.5).
We describe our GEMM kernels with main focus on the SGEMM implementation
since all GEMM kernels share common programming and optimization techniques.
While a conventional wisdom of GPU programming recommends us 
to heavily use shared memory on GPUs,  
we show that texture cache is very effective on the Cypress architecture. 

== preliminary results ==
 * [wiki:"GEMM_Performance_Cypress"]
 * [wiki:"MatrixMultiply"]

== preprint ==
Posted later.

== Sample program for DGEMM ==