[X86][MMX] Add tests showing missed opportunities to use MMX sitofp c…

…onversions

If we are transferring MMX registers to XMM for conversion we could use the MMX equivalents (CVTPI2PD + CVTPI2PS) without affecting rounding/exceptions etc.

llvm-svn: 297481

llvm/test/CodeGen/X86/mmx-cvt.ll

@@ -3,7 +3,7 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+mmx,+sse2 | FileCheck %s --check-prefix=X64
 
 ; FIXME: If we are transferring XMM conversion results to MMX registers we could use the MMX equivalents
-; (CVTPD2PI/CVTTPD2PI + CVTPS2PI/CVTTPS2PI) with affecting rounding/expections etc.
+; (CVTPD2PI/CVTTPD2PI + CVTPS2PI/CVTTPS2PI) without affecting rounding/exceptions etc.
 
 define void @cvt_v2f64_v2i32(<2 x double>, <1 x i64>*) nounwind {
 ; X86-LABEL: cvt_v2f64_v2i32:
@@ -248,8 +248,83 @@ define void @fptosi_v2f32_v2i32(<4 x float>, <1 x i64>*) nounwind {
   ret void
 }
 
+; FIXME: If we are transferring MMX conversion results to XMM registers we could use the MMX equivalents
+; (CVTPI2PD + CVTPI2PS) without affecting rounding/exceptions etc.
+
+define <2 x double> @sitofp_v2i32_v2f64(<1 x i64>*) nounwind {
+; X86-LABEL: sitofp_v2i32_v2f64:
+; X86:       # BB#0:
+; X86-NEXT:    pushl %ebp
+; X86-NEXT:    movl %esp, %ebp
+; X86-NEXT:    andl $-8, %esp
+; X86-NEXT:    subl $8, %esp
+; X86-NEXT:    movl 8(%ebp), %eax
+; X86-NEXT:    movq (%eax), %mm0
+; X86-NEXT:    paddd %mm0, %mm0
+; X86-NEXT:    movq %mm0, (%esp)
+; X86-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-NEXT:    cvtdq2pd %xmm0, %xmm0
+; X86-NEXT:    movl %ebp, %esp
+; X86-NEXT:    popl %ebp
+; X86-NEXT:    retl
+;
+; X64-LABEL: sitofp_v2i32_v2f64:
+; X64:       # BB#0:
+; X64-NEXT:    movq (%rdi), %mm0
+; X64-NEXT:    paddd %mm0, %mm0
+; X64-NEXT:    movq2dq %mm0, %xmm0
+; X64-NEXT:    cvtdq2pd %xmm0, %xmm0
+; X64-NEXT:    retq
+  %2 = bitcast <1 x i64>* %0 to x86_mmx*
+  %3 = load x86_mmx, x86_mmx* %2, align 8
+  %4 = tail call x86_mmx @llvm.x86.mmx.padd.d(x86_mmx %3, x86_mmx %3)
+  %5 = bitcast x86_mmx %4 to i64
+  %6 = insertelement <2 x i64> undef, i64 %5, i32 0
+  %7 = bitcast <2 x i64> %6 to <4 x i32>
+  %8 = shufflevector <4 x i32> %7, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
+  %9 = sitofp <2 x i32> %8 to <2 x double>
+  ret <2 x double> %9
+}
+
+define <4 x float> @sitofp_v2i32_v2f32(<1 x i64>*) nounwind {
+; X86-LABEL: sitofp_v2i32_v2f32:
+; X86:       # BB#0:
+; X86-NEXT:    pushl %ebp
+; X86-NEXT:    movl %esp, %ebp
+; X86-NEXT:    andl $-8, %esp
+; X86-NEXT:    subl $8, %esp
+; X86-NEXT:    movl 8(%ebp), %eax
+; X86-NEXT:    movq (%eax), %mm0
+; X86-NEXT:    paddd %mm0, %mm0
+; X86-NEXT:    movq %mm0, (%esp)
+; X86-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-NEXT:    cvtdq2ps %xmm0, %xmm0
+; X86-NEXT:    movl %ebp, %esp
+; X86-NEXT:    popl %ebp
+; X86-NEXT:    retl
+;
+; X64-LABEL: sitofp_v2i32_v2f32:
+; X64:       # BB#0:
+; X64-NEXT:    movq (%rdi), %mm0
+; X64-NEXT:    paddd %mm0, %mm0
+; X64-NEXT:    movd %mm0, %rax
+; X64-NEXT:    movd %rax, %xmm0
+; X64-NEXT:    cvtdq2ps %xmm0, %xmm0
+; X64-NEXT:    retq
+  %2 = bitcast <1 x i64>* %0 to x86_mmx*
+  %3 = load x86_mmx, x86_mmx* %2, align 8
+  %4 = tail call x86_mmx @llvm.x86.mmx.padd.d(x86_mmx %3, x86_mmx %3)
+  %5 = bitcast x86_mmx %4 to i64
+  %6 = insertelement <2 x i64> undef, i64 %5, i32 0
+  %7 = insertelement <2 x i64> %6, i64 0, i32 1
+  %8 = bitcast <2 x i64> %7 to <4 x i32>
+  %9 = tail call <4 x float> @llvm.x86.sse2.cvtdq2ps(<4 x i32> %8)
+  ret <4 x float> %9
+}
+
 declare x86_mmx @llvm.x86.mmx.padd.d(x86_mmx, x86_mmx)
 declare <4 x i32> @llvm.x86.sse2.cvtpd2dq(<2 x double>)
 declare <4 x i32> @llvm.x86.sse2.cvttpd2dq(<2 x double>)
 declare <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float>)
 declare <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float>)
+declare <4 x float> @llvm.x86.sse2.cvtdq2ps(<4 x i32>)