llvm-project/llvm/test/CodeGen/X86/combine-srem.ll

68 lines
2.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
; fold (srem undef, x) -> 0
define <4 x i32> @combine_vec_srem_undef0(<4 x i32> %x) {
; SSE-LABEL: combine_vec_srem_undef0:
; SSE: # BB#0:
; SSE-NEXT: retq
;
; AVX-LABEL: combine_vec_srem_undef0:
; AVX: # BB#0:
; AVX-NEXT: retq
%1 = srem <4 x i32> undef, %x
ret <4 x i32> %1
}
; fold (srem x, undef) -> undef
define <4 x i32> @combine_vec_srem_undef1(<4 x i32> %x) {
; SSE-LABEL: combine_vec_srem_undef1:
; SSE: # BB#0:
; SSE-NEXT: retq
;
; AVX-LABEL: combine_vec_srem_undef1:
; AVX: # BB#0:
; AVX-NEXT: retq
%1 = srem <4 x i32> %x, undef
ret <4 x i32> %1
}
; fold (srem x, y) -> (urem x, y) iff x and y are positive
define <4 x i32> @combine_vec_srem_by_pos0(<4 x i32> %x) {
; SSE-LABEL: combine_vec_srem_by_pos0:
; SSE: # BB#0:
; SSE-NEXT: andps {{.*}}(%rip), %xmm0
; SSE-NEXT: retq
;
; AVX1-LABEL: combine_vec_srem_by_pos0:
; AVX1: # BB#0:
; AVX1-NEXT: vandps {{.*}}(%rip), %xmm0, %xmm0
; AVX1-NEXT: retq
;
; AVX2-LABEL: combine_vec_srem_by_pos0:
; AVX2: # BB#0:
; AVX2-NEXT: vbroadcastss {{.*}}(%rip), %xmm1
; AVX2-NEXT: vandps %xmm1, %xmm0, %xmm0
; AVX2-NEXT: retq
%1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255>
%2 = srem <4 x i32> %1, <i32 4, i32 4, i32 4, i32 4>
ret <4 x i32> %2
}
define <4 x i32> @combine_vec_srem_by_pos1(<4 x i32> %x) {
; SSE-LABEL: combine_vec_srem_by_pos1:
; SSE: # BB#0:
; SSE-NEXT: andps {{.*}}(%rip), %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_vec_srem_by_pos1:
; AVX: # BB#0:
; AVX-NEXT: vandps {{.*}}(%rip), %xmm0, %xmm0
; AVX-NEXT: retq
%1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255>
%2 = srem <4 x i32> %1, <i32 1, i32 4, i32 8, i32 16>
ret <4 x i32> %2
}