restored intrinsics and added tuning options to ptxgen

2013-11-15 15:04:04 +01:00
parent f9d2ede83c
commit c93e71698e
2 changed files with 8 additions and 6 deletions
--- a/examples_cuda/aobench/Makefile_gpu
+++ b/examples_cuda/aobench/Makefile_gpu
@@ -13,6 +13,8 @@ ISPCFLAGS=-O3 --math-lib=default --target=nvptx64 --opt=fast-math
 LLVM32 = $(HOME)/usr/local/llvm/bin-3.2
 LLVM   = $(HOME)/usr/local/llvm/bin-3.3
 PTXGEN = $(HOME)/ptxgen
+PTXGEN += -opt=3
+PTXGEN += -ftz=1 -prec-div=0 -prec-sqrt=0 -fma=1

 LLVM32DIS=$(LLVM32)/bin/llvm-dis