LLVM's default threshold is 150, too small. In general, the loop unroll in GEN can improve the perfomance, so enlarge it to 1024.
Signed-off-by: Yang Rong <[email protected]> --- backend/src/llvm/llvm_to_gen.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backend/src/llvm/llvm_to_gen.cpp b/backend/src/llvm/llvm_to_gen.cpp index e0f6bf2..891f2a1 100644 --- a/backend/src/llvm/llvm_to_gen.cpp +++ b/backend/src/llvm/llvm_to_gen.cpp @@ -160,7 +160,7 @@ namespace gbe MPM.add(createIndVarSimplifyPass()); // Canonicalize indvars MPM.add(createLoopIdiomPass()); // Recognize idioms like memset. MPM.add(createLoopDeletionPass()); // Delete dead loops - MPM.add(createLoopUnrollPass()); //1024, 32, 1024, 512)); //Unroll loops + MPM.add(createLoopUnrollPass(1024)); //1024, 32, 1024, 512)); //Unroll loops if(optLevel > 0) { MPM.add(createSROAPass(/*RequiresDomTree*/ false)); MPM.add(createGVNPass()); // Remove redundancies -- 1.8.3.2 _______________________________________________ Beignet mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/beignet
