[llvm-commits] [llvm] r152762 - in /llvm/trunk: lib/Target/X86/X86InstrSSE.td test/CodeGen/X86/avx-vinsertf128.ll

[Chad Rosier]

Author: mcrosier
Date: Wed Mar 14 19:45:30 2012
New Revision: 152762

URL: http://llvm.org/viewvc/llvm-project?rev=152762&view=rev
Log:
[avx] Add patterns for VINSERTF128rm.

This results in things such as

	vmovaps	-96(%rbx), %xmm1
	vinsertf128	$1, %xmm1, %ymm0, %ymm0

to be combined to
         
	vinsertf128	$1, -96(%rbx), %ymm0, %ymm0

rdar://10643481

Modified:
    llvm/trunk/lib/Target/X86/X86InstrSSE.td
    llvm/trunk/test/CodeGen/X86/avx-vinsertf128.ll

Modified: llvm/trunk/lib/Target/X86/X86InstrSSE.td
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrSSE.td?rev=152762&r1=152761&r2=152762&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86InstrSSE.td (original)
+++ llvm/trunk/lib/Target/X86/X86InstrSSE.td Wed Mar 14 19:45:30 2012
@@ -7758,6 +7758,19 @@
                                    (i32 imm)),
           (VINSERTF128rr VR256:$src1, VR128:$src2,
                          (INSERT_get_vinsertf128_imm VR256:$ins))>;
+
+def : Pat<(vinsertf128_insert:$ins (v8f32 VR256:$src1), (alignedloadv4f32 addr:$src2),
+                                   (i32 imm)),
+          (VINSERTF128rm VR256:$src1, addr:$src2,
+                         (INSERT_get_vinsertf128_imm VR256:$ins))>;
+def : Pat<(vinsertf128_insert:$ins (v4f64 VR256:$src1), (alignedloadv2f64 addr:$src2),
+                                   (i32 imm)),
+          (VINSERTF128rm VR256:$src1, addr:$src2,
+                         (INSERT_get_vinsertf128_imm VR256:$ins))>;
+def : Pat<(vinsertf128_insert:$ins (v4i64 VR256:$src1), (alignedloadv2i64 addr:$src2),
+                                   (i32 imm)),
+          (VINSERTF128rm VR256:$src1, addr:$src2,
+                         (INSERT_get_vinsertf128_imm VR256:$ins))>;
 }
 
 //===----------------------------------------------------------------------===//

Modified: llvm/trunk/test/CodeGen/X86/avx-vinsertf128.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx-vinsertf128.ll?rev=152762&r1=152761&r2=152762&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx-vinsertf128.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx-vinsertf128.ll Wed Mar 14 19:45:30 2012
@@ -104,3 +104,15 @@
 }
 declare <8 x i32> @llvm.x86.avx.vinsertf128.si.256(<8 x i32>, <4 x i32>, i8) nounwind readnone
 
+; rdar://10643481
+; CHECK: vinsertf128_combine
+define <8 x float> @vinsertf128_combine(float* nocapture %f) nounwind uwtable readonly ssp {
+; CHECK-NOT: vmovaps
+; CHECK: vinsertf128
+entry:
+  %add.ptr = getelementptr inbounds float* %f, i64 4
+  %0 = bitcast float* %add.ptr to <4 x float>*
+  %1 = load <4 x float>* %0, align 16
+  %2 = tail call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float> undef, <4 x float> %1, i8 1)
+  ret <8 x float> %2
+}