[X86][SSE] Introduce Float/Vector WriteMove, WriteLoad and Writetore scheduler classes

As discussed on D44428 and PR36726, this patch splits off WriteFMove/WriteVecMove, WriteFLoad/WriteVecLoad and WriteFStore/WriteVecStore scheduler classes to permit vectors to be handled separately from gpr/scalar types.

I've minimised the diff here by only moving various basic SSE/AVX vector instructions across - we can fix the rest when called for. This does fix the MOVDQA vs MOVAPS/MOVAPD discrepancies mentioned on D44428.

Differential Revision: https://reviews.llvm.org/D44471

llvm-svn: 327630

7 files changed, 36 insertions, 36 deletions

diff --git a/llvm/test/CodeGen/X86/avx-schedule.ll b/llvm/test/CodeGen/X86/avx-schedule.ll
index 6d25787c190..7f583a47efd 100644
--- a/llvm/test/CodeGen/X86/avx-schedule.ll
+++ b/llvm/test/CodeGen/X86/avx-schedule.ll
@@ -2103,7 +2103,7 @@ define <2 x double> @test_maskmovpd(i8* %a0, <2 x i64> %a1, <2 x double> %a2) {
 ; ZNVER1:       # %bb.0:
 ; ZNVER1-NEXT:    vmaskmovpd (%rdi), %xmm0, %xmm2 # sched: [8:0.50]
 ; ZNVER1-NEXT:    vmaskmovpd %xmm1, %xmm0, (%rdi) # sched: [4:0.50]
-; ZNVER1-NEXT:    vmovapd %xmm2, %xmm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovapd %xmm2, %xmm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <2 x double> @llvm.x86.avx.maskload.pd(i8* %a0, <2 x i64> %a1)
   call void @llvm.x86.avx.maskstore.pd(i8* %a0, <2 x i64> %a1, <2 x double> %a2)
@@ -2166,7 +2166,7 @@ define <4 x double> @test_maskmovpd_ymm(i8* %a0, <4 x i64> %a1, <4 x double> %a2
 ; ZNVER1:       # %bb.0:
 ; ZNVER1-NEXT:    vmaskmovpd (%rdi), %ymm0, %ymm2 # sched: [8:1.00]
 ; ZNVER1-NEXT:    vmaskmovpd %ymm1, %ymm0, (%rdi) # sched: [5:1.00]
-; ZNVER1-NEXT:    vmovapd %ymm2, %ymm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovapd %ymm2, %ymm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <4 x double> @llvm.x86.avx.maskload.pd.256(i8* %a0, <4 x i64> %a1)
   call void @llvm.x86.avx.maskstore.pd.256(i8* %a0, <4 x i64> %a1, <4 x double> %a2)
@@ -2229,7 +2229,7 @@ define <4 x float> @test_maskmovps(i8* %a0, <4 x i32> %a1, <4 x float> %a2) {
 ; ZNVER1:       # %bb.0:
 ; ZNVER1-NEXT:    vmaskmovps (%rdi), %xmm0, %xmm2 # sched: [8:0.50]
 ; ZNVER1-NEXT:    vmaskmovps %xmm1, %xmm0, (%rdi) # sched: [4:0.50]
-; ZNVER1-NEXT:    vmovaps %xmm2, %xmm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovaps %xmm2, %xmm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <4 x float> @llvm.x86.avx.maskload.ps(i8* %a0, <4 x i32> %a1)
   call void @llvm.x86.avx.maskstore.ps(i8* %a0, <4 x i32> %a1, <4 x float> %a2)
@@ -2292,7 +2292,7 @@ define <8 x float> @test_maskmovps_ymm(i8* %a0, <8 x i32> %a1, <8 x float> %a2)
 ; ZNVER1:       # %bb.0:
 ; ZNVER1-NEXT:    vmaskmovps (%rdi), %ymm0, %ymm2 # sched: [8:1.00]
 ; ZNVER1-NEXT:    vmaskmovps %ymm1, %ymm0, (%rdi) # sched: [5:1.00]
-; ZNVER1-NEXT:    vmovaps %ymm2, %ymm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovaps %ymm2, %ymm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <8 x float> @llvm.x86.avx.maskload.ps.256(i8* %a0, <8 x i32> %a1)
   call void @llvm.x86.avx.maskstore.ps.256(i8* %a0, <8 x i32> %a1, <8 x float> %a2)
diff --git a/llvm/test/CodeGen/X86/avx2-schedule.ll b/llvm/test/CodeGen/X86/avx2-schedule.ll
index d06b6a78a8c..e7152f867e8 100644
--- a/llvm/test/CodeGen/X86/avx2-schedule.ll
+++ b/llvm/test/CodeGen/X86/avx2-schedule.ll
@@ -573,7 +573,7 @@ define <8 x i32> @test_inserti128(<8 x i32> %a0, <4 x i32> %a1, <4 x i32> *%a2)
 define <4 x i64> @test_movntdqa(i8* %a0) {
 ; GENERIC-LABEL: test_movntdqa:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vmovntdqa (%rdi), %ymm0 # sched: [4:0.50]
+; GENERIC-NEXT:    vmovntdqa (%rdi), %ymm0 # sched: [6:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
 ;
 ; HASWELL-LABEL: test_movntdqa:
@@ -3380,7 +3380,7 @@ declare <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16>, <16 x i16>) nounwind readn
 define <4 x i32> @test_pmaskmovd(i8* %a0, <4 x i32> %a1, <4 x i32> %a2) {
 ; GENERIC-LABEL: test_pmaskmovd:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vpmaskmovd (%rdi), %xmm0, %xmm2 # sched: [4:0.50]
+; GENERIC-NEXT:    vpmaskmovd (%rdi), %xmm0, %xmm2 # sched: [6:0.50]
 ; GENERIC-NEXT:    vpmaskmovd %xmm1, %xmm0, (%rdi) # sched: [1:1.00]
 ; GENERIC-NEXT:    vmovdqa %xmm2, %xmm0 # sched: [1:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
@@ -3429,7 +3429,7 @@ declare void @llvm.x86.avx2.maskstore.d(i8*, <4 x i32>, <4 x i32>) nounwind
 define <8 x i32> @test_pmaskmovd_ymm(i8* %a0, <8 x i32> %a1, <8 x i32> %a2) {
 ; GENERIC-LABEL: test_pmaskmovd_ymm:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vpmaskmovd (%rdi), %ymm0, %ymm2 # sched: [4:0.50]
+; GENERIC-NEXT:    vpmaskmovd (%rdi), %ymm0, %ymm2 # sched: [6:0.50]
 ; GENERIC-NEXT:    vpmaskmovd %ymm1, %ymm0, (%rdi) # sched: [1:1.00]
 ; GENERIC-NEXT:    vmovdqa %ymm2, %ymm0 # sched: [1:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
@@ -3478,7 +3478,7 @@ declare void @llvm.x86.avx2.maskstore.d.256(i8*, <8 x i32>, <8 x i32>) nounwind
 define <2 x i64> @test_pmaskmovq(i8* %a0, <2 x i64> %a1, <2 x i64> %a2) {
 ; GENERIC-LABEL: test_pmaskmovq:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vpmaskmovq (%rdi), %xmm0, %xmm2 # sched: [4:0.50]
+; GENERIC-NEXT:    vpmaskmovq (%rdi), %xmm0, %xmm2 # sched: [6:0.50]
 ; GENERIC-NEXT:    vpmaskmovq %xmm1, %xmm0, (%rdi) # sched: [1:1.00]
 ; GENERIC-NEXT:    vmovdqa %xmm2, %xmm0 # sched: [1:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
@@ -3527,7 +3527,7 @@ declare void @llvm.x86.avx2.maskstore.q(i8*, <2 x i64>, <2 x i64>) nounwind
 define <4 x i64> @test_pmaskmovq_ymm(i8* %a0, <4 x i64> %a1, <4 x i64> %a2) {
 ; GENERIC-LABEL: test_pmaskmovq_ymm:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vpmaskmovq (%rdi), %ymm0, %ymm2 # sched: [4:0.50]
+; GENERIC-NEXT:    vpmaskmovq (%rdi), %ymm0, %ymm2 # sched: [6:0.50]
 ; GENERIC-NEXT:    vpmaskmovq %ymm1, %ymm0, (%rdi) # sched: [1:1.00]
 ; GENERIC-NEXT:    vmovdqa %ymm2, %ymm0 # sched: [1:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
diff --git a/llvm/test/CodeGen/X86/sha-schedule.ll b/llvm/test/CodeGen/X86/sha-schedule.ll
index 138ff888b92..cecdc1c41b6 100644
--- a/llvm/test/CodeGen/X86/sha-schedule.ll
+++ b/llvm/test/CodeGen/X86/sha-schedule.ll
@@ -210,11 +210,11 @@ define <4 x i32> @test_sha256rnds2(<4 x i32> %a0, <4 x i32> %a1, <4 x i32> %a2,
 ;
 ; GOLDMONT-LABEL: test_sha256rnds2:
 ; GOLDMONT:       # %bb.0:
-; GOLDMONT-NEXT:    movaps %xmm0, %xmm3 # sched: [1:1.00]
-; GOLDMONT-NEXT:    movaps %xmm2, %xmm0 # sched: [1:1.00]
+; GOLDMONT-NEXT:    movaps %xmm0, %xmm3 # sched: [1:0.50]
+; GOLDMONT-NEXT:    movaps %xmm2, %xmm0 # sched: [1:0.50]
 ; GOLDMONT-NEXT:    sha256rnds2 %xmm0, %xmm1, %xmm3 # sched: [4:1.00]
 ; GOLDMONT-NEXT:    sha256rnds2 %xmm0, (%rdi), %xmm3 # sched: [7:1.00]
-; GOLDMONT-NEXT:    movaps %xmm3, %xmm0 # sched: [1:1.00]
+; GOLDMONT-NEXT:    movaps %xmm3, %xmm0 # sched: [1:0.50]
 ; GOLDMONT-NEXT:    retq # sched: [4:1.00]
 ;
 ; CANNONLAKE-LABEL: test_sha256rnds2:
@@ -228,11 +228,11 @@ define <4 x i32> @test_sha256rnds2(<4 x i32> %a0, <4 x i32> %a1, <4 x i32> %a2,
 ;
 ; ZNVER1-LABEL: test_sha256rnds2:
 ; ZNVER1:       # %bb.0:
-; ZNVER1-NEXT:    vmovaps %xmm0, %xmm3 # sched: [1:0.50]
-; ZNVER1-NEXT:    vmovaps %xmm2, %xmm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovaps %xmm0, %xmm3 # sched: [1:0.25]
+; ZNVER1-NEXT:    vmovaps %xmm2, %xmm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    sha256rnds2 %xmm0, %xmm1, %xmm3 # sched: [4:1.00]
 ; ZNVER1-NEXT:    sha256rnds2 %xmm0, (%rdi), %xmm3 # sched: [11:1.00]
-; ZNVER1-NEXT:    vmovaps %xmm3, %xmm0 # sched: [1:0.50]
+; ZNVER1-NEXT:    vmovaps %xmm3, %xmm0 # sched: [1:0.25]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = load <4 x i32>, <4 x i32>* %a3
   %2 = tail call <4 x i32> @llvm.x86.sha256rnds2(<4 x i32> %a0, <4 x i32> %a1, <4 x i32> %a2)
diff --git a/llvm/test/CodeGen/X86/sse-schedule.ll b/llvm/test/CodeGen/X86/sse-schedule.ll
index 8ebea98274d..3a123df699f 100644
--- a/llvm/test/CodeGen/X86/sse-schedule.ll
+++ b/llvm/test/CodeGen/X86/sse-schedule.ll
@@ -2557,7 +2557,7 @@ define <4 x float> @test_rcpps(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    rcpps (%rdi), %xmm1 # sched: [8:1.00]
 ; SLM-NEXT:    rcpps %xmm0, %xmm0 # sched: [5:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_rcpps:
@@ -2729,7 +2729,7 @@ define <4 x float> @test_rsqrtps(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    rsqrtps (%rdi), %xmm1 # sched: [8:1.00]
 ; SLM-NEXT:    rsqrtps %xmm0, %xmm0 # sched: [5:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_rsqrtps:
@@ -3038,7 +3038,7 @@ define <4 x float> @test_sqrtps(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    sqrtps (%rdi), %xmm1 # sched: [18:1.00]
 ; SLM-NEXT:    sqrtps %xmm0, %xmm0 # sched: [15:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_sqrtps:
diff --git a/llvm/test/CodeGen/X86/sse2-schedule.ll b/llvm/test/CodeGen/X86/sse2-schedule.ll
index a7c1c75226e..cb554e7c931 100644
--- a/llvm/test/CodeGen/X86/sse2-schedule.ll
+++ b/llvm/test/CodeGen/X86/sse2-schedule.ll
@@ -3564,7 +3564,7 @@ define <2 x double> @test_movsd_reg(<2 x double> %a0, <2 x double> %a1) {
 ; SLM-LABEL: test_movsd_reg:
 ; SLM:       # %bb.0:
 ; SLM-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] sched: [1:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_movsd_reg:
@@ -8756,7 +8756,7 @@ define <2 x double> @test_sqrtpd(<2 x double> %a0, <2 x double> *%a1) {
 ; SLM-NEXT:    sqrtpd (%rdi), %xmm1 # sched: [18:1.00]
 ; SLM-NEXT:    sqrtpd %xmm0, %xmm0 # sched: [15:1.00]
 ; SLM-NEXT:    addpd %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_sqrtpd:
@@ -9284,10 +9284,10 @@ define <2 x double> @test_unpcklpd(<2 x double> %a0, <2 x double> %a1, <2 x doub
 ; SLM-LABEL: test_unpcklpd:
 ; SLM:       # %bb.0:
 ; SLM-NEXT:    unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0] sched: [1:1.00]
-; SLM-NEXT:    movapd %xmm0, %xmm1 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm0, %xmm1 # sched: [1:0.50]
 ; SLM-NEXT:    unpcklpd {{.*#+}} xmm1 = xmm1[0],mem[0] sched: [4:1.00]
 ; SLM-NEXT:    addpd %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_unpcklpd:
diff --git a/llvm/test/CodeGen/X86/sse3-schedule.ll b/llvm/test/CodeGen/X86/sse3-schedule.ll
index bb7694c1e4a..1b1e765ab27 100644
--- a/llvm/test/CodeGen/X86/sse3-schedule.ll
+++ b/llvm/test/CodeGen/X86/sse3-schedule.ll
@@ -566,7 +566,7 @@ define <2 x double> @test_movddup(<2 x double> %a0, <2 x double> *%a1) {
 ; SLM-NEXT:    movddup {{.*#+}} xmm1 = mem[0,0] sched: [4:1.00]
 ; SLM-NEXT:    movddup {{.*#+}} xmm0 = xmm0[0,0] sched: [1:1.00]
 ; SLM-NEXT:    subpd %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_movddup:
@@ -645,7 +645,7 @@ define <4 x float> @test_movshdup(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    movshdup {{.*#+}} xmm1 = mem[1,1,3,3] sched: [4:1.00]
 ; SLM-NEXT:    movshdup {{.*#+}} xmm0 = xmm0[1,1,3,3] sched: [1:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_movshdup:
@@ -724,7 +724,7 @@ define <4 x float> @test_movsldup(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    movsldup {{.*#+}} xmm1 = mem[0,0,2,2] sched: [4:1.00]
 ; SLM-NEXT:    movsldup {{.*#+}} xmm0 = xmm0[0,0,2,2] sched: [1:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_movsldup:
diff --git a/llvm/test/CodeGen/X86/sse41-schedule.ll b/llvm/test/CodeGen/X86/sse41-schedule.ll
index 60fc0e8e514..17188884983 100644
--- a/llvm/test/CodeGen/X86/sse41-schedule.ll
+++ b/llvm/test/CodeGen/X86/sse41-schedule.ll
@@ -163,11 +163,11 @@ define <2 x double> @test_blendvpd(<2 x double> %a0, <2 x double> %a1, <2 x doub
 ;
 ; SLM-LABEL: test_blendvpd:
 ; SLM:       # %bb.0:
-; SLM-NEXT:    movapd %xmm0, %xmm3 # sched: [1:1.00]
-; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm0, %xmm3 # sched: [1:0.50]
+; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    blendvpd %xmm0, %xmm1, %xmm3 # sched: [1:1.00]
 ; SLM-NEXT:    blendvpd %xmm0, (%rdi), %xmm3 # sched: [4:1.00]
-; SLM-NEXT:    movapd %xmm3, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm3, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_blendvpd:
@@ -230,11 +230,11 @@ define <4 x float> @test_blendvps(<4 x float> %a0, <4 x float> %a1, <4 x float>
 ;
 ; SLM-LABEL: test_blendvps:
 ; SLM:       # %bb.0:
-; SLM-NEXT:    movaps %xmm0, %xmm3 # sched: [1:1.00]
-; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm0, %xmm3 # sched: [1:0.50]
+; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    blendvps %xmm0, %xmm1, %xmm3 # sched: [1:1.00]
 ; SLM-NEXT:    blendvps %xmm0, (%rdi), %xmm3 # sched: [4:1.00]
-; SLM-NEXT:    movaps %xmm3, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm3, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_blendvps:
@@ -717,7 +717,7 @@ define <16 x i8> @test_pblendvb(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> %a2, <16
 ; SLM-LABEL: test_pblendvb:
 ; SLM:       # %bb.0:
 ; SLM-NEXT:    movdqa %xmm0, %xmm3 # sched: [1:0.50]
-; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm2, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    pblendvb %xmm0, %xmm1, %xmm3 # sched: [1:1.00]
 ; SLM-NEXT:    pblendvb %xmm0, (%rdi), %xmm3 # sched: [4:1.00]
 ; SLM-NEXT:    movdqa %xmm3, %xmm0 # sched: [1:0.50]
@@ -2991,7 +2991,7 @@ define <2 x double> @test_roundpd(<2 x double> %a0, <2 x double> *%a1) {
 ; SLM-NEXT:    roundpd $7, (%rdi), %xmm1 # sched: [6:1.00]
 ; SLM-NEXT:    roundpd $7, %xmm0, %xmm0 # sched: [3:1.00]
 ; SLM-NEXT:    addpd %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_roundpd:
@@ -3063,7 +3063,7 @@ define <4 x float> @test_roundps(<4 x float> %a0, <4 x float> *%a1) {
 ; SLM-NEXT:    roundps $7, (%rdi), %xmm1 # sched: [6:1.00]
 ; SLM-NEXT:    roundps $7, %xmm0, %xmm0 # sched: [3:1.00]
 ; SLM-NEXT:    addps %xmm0, %xmm1 # sched: [3:1.00]
-; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm1, %xmm0 # sched: [1:0.50]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-LABEL: test_roundps:
@@ -3133,7 +3133,7 @@ define <2 x double> @test_roundsd(<2 x double> %a0, <2 x double> %a1, <2 x doubl
 ;
 ; SLM-LABEL: test_roundsd:
 ; SLM:       # %bb.0:
-; SLM-NEXT:    movapd %xmm0, %xmm2 # sched: [1:1.00]
+; SLM-NEXT:    movapd %xmm0, %xmm2 # sched: [1:0.50]
 ; SLM-NEXT:    roundsd $7, (%rdi), %xmm0 # sched: [6:1.00]
 ; SLM-NEXT:    roundsd $7, %xmm1, %xmm2 # sched: [3:1.00]
 ; SLM-NEXT:    addpd %xmm2, %xmm0 # sched: [3:1.00]
@@ -3206,7 +3206,7 @@ define <4 x float> @test_roundss(<4 x float> %a0, <4 x float> %a1, <4 x float> *
 ;
 ; SLM-LABEL: test_roundss:
 ; SLM:       # %bb.0:
-; SLM-NEXT:    movaps %xmm0, %xmm2 # sched: [1:1.00]
+; SLM-NEXT:    movaps %xmm0, %xmm2 # sched: [1:0.50]
 ; SLM-NEXT:    roundss $7, (%rdi), %xmm0 # sched: [6:1.00]
 ; SLM-NEXT:    roundss $7, %xmm1, %xmm2 # sched: [3:1.00]
 ; SLM-NEXT:    addps %xmm2, %xmm0 # sched: [3:1.00]