search for: i94

...ub float %tmp71.i, %tmp74.i > store float %tmp75.i, float addrspace(1)* %arrayidx68.i, align 4 > %call.i87.i = tail call float @fabs(float %tmp11.i.i) nounwind > %tmp5.i92.i = fmul float %call.i87.i, 0x3FCDA67120000000 > %tmp6.i93.i = fadd float %tmp5.i92.i, 1.000000e+00 > %tmp7.i94.i = fdiv float 1.000000e+00, %tmp6.i93.i > %call16.i103.i = tail call float @llvm.exp.f32(float %tmp15.i.i) nounwind > %tmp17.i104.i = fmul float %call16.i103.i, 0x3FD9884540000000 > %tmp19.i106.i = fmul float %tmp17.i104.i, %tmp7.i94.i > %tmp29.i116.i = fmul float %tmp7.i94.i, 0x3F...

Hello everybody, I have stumbled upon a test case (the attached module is a slightly reduced version) that shows extremely reduced performance on linux compared to windows when executed using LLVM's JIT. We narrowed the problem down to the actual code being generated, the source IR on both systems is the same. Try compiling the attached module: llc -O3 -filetype=asm -o BAD.s BAD.ll Under

In 3.1, the backend is very slow to legalize types. Following is the code snippet which may be the culprit: %Result.i.i.i97 = alloca i33, align 8 %Result.i.i.i96= alloca i33, align 8 %Result.i.i.i95 = alloca i33, align 8 %Result.i.i.i94 = alloca i33, align 8 %Result.i.i.i93 = alloca i33, align 8 %Result.i.i.i92= alloca i33, align 8 %Result.i.i.i91 = alloca i33, align 8 %Result.i.i.i90 = alloca i33, align 8 %Result.i.i.i89 = alloca i33, align 8 The compilation time improve significant if running on LLVM 3.9. Unfortunate...

...6* %incdec.ptr.6.i914.i, align 2, !tbaa !5 store i16 %327, i16* %incdec.ptr1.6.i.i, align 2, !tbaa !5 %328 = load i16* %incdec.ptr.7.i916.i, align 2, !tbaa !5 store i16 %328, i16* %incdec.ptr1.7.i.i, align 2, !tbaa !5 %329 = load i16* %incdec.ptr.8.i918.i, align 2, !tbaa !5 %incdec.ptr1.9.i944.i = getelementptr inbounds [13 x i16]* %y.i, i32 0, i32 10 store i16 %329, i16* %incdec.ptr1.8.i.i, align 2, !tbaa !5 %330 = load i16* %incdec.ptr.9.i920.i, align 2, !tbaa !5 %incdec.ptr1.10.i946.i = getelementptr inbounds [13 x i16]* %y.i, i32 0, i32 11 store i16 %330, i16* %incdec.ptr1.9...