-
- Downloads
sgemm shared memory execution time reduced by running unroll and scalarrepl pass
on the kernel. This promotes the private memory allocation in each thread to registers.
Showing
- llvm/test/VISC/parboil/benchmarks/sgemm/src/visc_sh/Makefile 1 addition, 0 deletionsllvm/test/VISC/parboil/benchmarks/sgemm/src/visc_sh/Makefile
- llvm/test/VISC/parboil/benchmarks/sgemm/src/visc_sh/main.cc 133 additions, 178 deletionsllvm/test/VISC/parboil/benchmarks/sgemm/src/visc_sh/main.cc
- llvm/test/VISC/parboil/common/mk/visc.mk 5 additions, 1 deletionllvm/test/VISC/parboil/common/mk/visc.mk
Loading
Please register or sign in to comment