From: Simon Pilgrim Date: Sun, 3 Jan 2016 19:17:37 +0000 (+0000) Subject: [X86][MMX] Regenerated vector insertion test. X-Git-Url: http://plrg.eecs.uci.edu/git/?p=oota-llvm.git;a=commitdiff_plain;h=5da99c7bd4ef5caf08fd31570d34ed8f819353ad;ds=sidebyside [X86][MMX] Regenerated vector insertion test. Shows the true horror of what is going on.... git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256713 91177308-0d34-0410-b5e6-96231b3b80d8 --- diff --git a/test/CodeGen/X86/vec_insert-7.ll b/test/CodeGen/X86/vec_insert-7.ll index 6d4f8287cab..4f72c66ecba 100644 --- a/test/CodeGen/X86/vec_insert-7.ll +++ b/test/CodeGen/X86/vec_insert-7.ll @@ -1,15 +1,29 @@ +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -march=x86 -mattr=+mmx,+sse4.2 -mtriple=i686-apple-darwin9 | FileCheck %s + ; MMX insertelement is not available; these are promoted to XMM. ; (Without SSE they are split to two ints, and the code is much better.) -define x86_mmx @mmx_movzl(x86_mmx %x) nounwind { -entry: -; CHECK: mmx_movzl -; CHECK: pinsrd -; CHECK: pinsrd - %tmp = bitcast x86_mmx %x to <2 x i32> - %tmp3 = insertelement <2 x i32> %tmp, i32 32, i32 0 ; <<2 x i32>> [#uses=1] - %tmp8 = insertelement <2 x i32> %tmp3, i32 0, i32 1 ; <<2 x i32>> [#uses=1] - %tmp9 = bitcast <2 x i32> %tmp8 to x86_mmx - ret x86_mmx %tmp9 +define x86_mmx @mmx_movzl(x86_mmx %x) nounwind { +; CHECK-LABEL: mmx_movzl: +; CHECK: ## BB#0: +; CHECK-NEXT: subl $20, %esp +; CHECK-NEXT: movq %mm0, {{[0-9]+}}(%esp) +; CHECK-NEXT: pmovzxdq {{.*#+}} xmm0 = mem[0],zero,mem[1],zero +; CHECK-NEXT: movl $32, %eax +; CHECK-NEXT: pinsrd $0, %eax, %xmm0 +; CHECK-NEXT: xorl %eax, %eax +; CHECK-NEXT: pinsrd $1, %eax, %xmm0 +; CHECK-NEXT: pinsrd $2, %eax, %xmm0 +; CHECK-NEXT: pinsrd $3, %eax, %xmm0 +; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3] +; CHECK-NEXT: movq %xmm0, (%esp) +; CHECK-NEXT: movq (%esp), %mm0 +; CHECK-NEXT: addl $20, %esp +; CHECK-NEXT: retl + %tmp = bitcast x86_mmx %x to <2 x i32> + %tmp3 = insertelement <2 x i32> %tmp, i32 32, i32 0 ; <<2 x i32>> [#uses=1] + %tmp8 = insertelement <2 x i32> %tmp3, i32 0, i32 1 ; <<2 x i32>> [#uses=1] + %tmp9 = bitcast <2 x i32> %tmp8 to x86_mmx + ret x86_mmx %tmp9 }