xref: /llvm-project/llvm/test/CodeGen/X86/insertps-from-constantpool.ll (revision a82ca1cd1b57671ce5ddbed63c6418f6841fe71d)
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --no_x86_scrub_mem_shuffle --version 4
2; RUN: llc -mtriple=i686-unknown-unknown -mattr=+sse4.1 < %s | FileCheck %s --check-prefix=X86
3; RUN: llc -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 < %s | FileCheck %s --check-prefix=X64
4
5; Test for case where insertps folds the load of an insertion element from a constant pool.
6
7define <4 x float> @fold_from_constantpool(<4 x float> %a) {
8; X86-LABEL: fold_from_constantpool:
9; X86:       # %bb.0:
10; X86-NEXT:    insertps $0, {{\.?LCPI[0-9]+_[0-9]+}}+4, %xmm0 # xmm0 = mem[0],xmm0[1,2,3]
11; X86-NEXT:    retl
12;
13; X64-LABEL: fold_from_constantpool:
14; X64:       # %bb.0:
15; X64-NEXT:    insertps $0, {{\.?LCPI[0-9]+_[0-9]+}}+4(%rip), %xmm0 # xmm0 = mem[0],xmm0[1,2,3]
16; X64-NEXT:    retq
17  %1 = call <4 x float> @llvm.x86.sse41.insertps(<4 x float> %a, <4 x float> <float 0.0, float 1.0, float 0.0, float 0.0>, i8 64)
18  ret <4 x float> %1
19}
20
21declare <4 x float> @llvm.x86.sse41.insertps(<4 x float>, <4 x float>, i8) nounwind readnone
22