Update aosp/master LLVM for rebase to r222494.

Change-Id: Ic787f5e0124df789bd26f3f24680f45e678eef2d

6 files changed, 174 insertions, 63 deletions

diff --git a/test/Analysis/CostModel/ARM/cast.ll b/test/Analysis/CostModel/ARM/cast.ll
index 662110f..18d6e84 100644
--- a/test/Analysis/CostModel/ARM/cast.ll
+++ b/test/Analysis/CostModel/ARM/cast.ll
@@ -221,35 +221,35 @@ define i32 @casts() {
   %r96 = fptoui <2 x float> undef to <2 x i32>
   ; CHECK: cost of 1 {{.*}} fptosi
   %r97 = fptosi <2 x float> undef to <2 x i32>
-  ; CHECK: cost of 28 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r98 = fptoui <2 x float> undef to <2 x i64>
-  ; CHECK: cost of 28 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r99 = fptosi <2 x float> undef to <2 x i64>
 
-  ; CHECK: cost of 8 {{.*}} fptoui
+  ; CHECK: cost of 16 {{.*}} fptoui
   %r100 = fptoui <2 x double> undef to <2 x i1>
-  ; CHECK: cost of 8 {{.*}} fptosi
+  ; CHECK: cost of 16 {{.*}} fptosi
   %r101 = fptosi <2 x double> undef to <2 x i1>
-  ; CHECK: cost of 8 {{.*}} fptoui
+  ; CHECK: cost of 16 {{.*}} fptoui
   %r102 = fptoui <2 x double> undef to <2 x i8>
-  ; CHECK: cost of 8 {{.*}} fptosi
+  ; CHECK: cost of 16 {{.*}} fptosi
   %r103 = fptosi <2 x double> undef to <2 x i8>
-  ; CHECK: cost of 8 {{.*}} fptoui
+  ; CHECK: cost of 16 {{.*}} fptoui
   %r104 = fptoui <2 x double> undef to <2 x i16>
-  ; CHECK: cost of 8 {{.*}} fptosi
+  ; CHECK: cost of 16 {{.*}} fptosi
   %r105 = fptosi <2 x double> undef to <2 x i16>
   ; CHECK: cost of 2 {{.*}} fptoui
   %r106 = fptoui <2 x double> undef to <2 x i32>
   ; CHECK: cost of 2 {{.*}} fptosi
   %r107 = fptosi <2 x double> undef to <2 x i32>
-  ; CHECK: cost of 28 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r108 = fptoui <2 x double> undef to <2 x i64>
-  ; CHECK: cost of 28 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r109 = fptosi <2 x double> undef to <2 x i64>
 
-  ; CHECK: cost of 16 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r110 = fptoui <4 x float> undef to <4 x i1>
-  ; CHECK: cost of 16 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r111 = fptosi <4 x float> undef to <4 x i1>
   ; CHECK: cost of 3 {{.*}} fptoui
   %r112 = fptoui <4 x float> undef to <4 x i8>
@@ -263,39 +263,39 @@ define i32 @casts() {
   %r116 = fptoui <4 x float> undef to <4 x i32>
   ; CHECK: cost of 1 {{.*}} fptosi
   %r117 = fptosi <4 x float> undef to <4 x i32>
-  ; CHECK: cost of 56 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r118 = fptoui <4 x float> undef to <4 x i64>
-  ; CHECK: cost of 56 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r119 = fptosi <4 x float> undef to <4 x i64>
 
-  ; CHECK: cost of 16 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r120 = fptoui <4 x double> undef to <4 x i1>
-  ; CHECK: cost of 16 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r121 = fptosi <4 x double> undef to <4 x i1>
-  ; CHECK: cost of 16 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r122 = fptoui <4 x double> undef to <4 x i8>
-  ; CHECK: cost of 16 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r123 = fptosi <4 x double> undef to <4 x i8>
-  ; CHECK: cost of 16 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r124 = fptoui <4 x double> undef to <4 x i16>
-  ; CHECK: cost of 16 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r125 = fptosi <4 x double> undef to <4 x i16>
-  ; CHECK: cost of 16 {{.*}} fptoui
+  ; CHECK: cost of 32 {{.*}} fptoui
   %r126 = fptoui <4 x double> undef to <4 x i32>
-  ; CHECK: cost of 16 {{.*}} fptosi
+  ; CHECK: cost of 32 {{.*}} fptosi
   %r127 = fptosi <4 x double> undef to <4 x i32>
-  ; CHECK: cost of 56 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r128 = fptoui <4 x double> undef to <4 x i64>
-  ; CHECK: cost of 56 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r129 = fptosi <4 x double> undef to <4 x i64>
 
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r130 = fptoui <8 x float> undef to <8 x i1>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r131 = fptosi <8 x float> undef to <8 x i1>
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r132 = fptoui <8 x float> undef to <8 x i8>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r133 = fptosi <8 x float> undef to <8 x i8>
   ; CHECK: cost of 4 {{.*}} fptoui
   %r134 = fptoui <8 x float> undef to <8 x i16>
@@ -305,39 +305,39 @@ define i32 @casts() {
   %r136 = fptoui <8 x float> undef to <8 x i32>
   ; CHECK: cost of 2 {{.*}} fptosi
   %r137 = fptosi <8 x float> undef to <8 x i32>
-  ; CHECK: cost of 112 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r138 = fptoui <8 x float> undef to <8 x i64>
-  ; CHECK: cost of 112 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r139 = fptosi <8 x float> undef to <8 x i64>
 
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r140 = fptoui <8 x double> undef to <8 x i1>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r141 = fptosi <8 x double> undef to <8 x i1>
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r142 = fptoui <8 x double> undef to <8 x i8>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r143 = fptosi <8 x double> undef to <8 x i8>
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r144 = fptoui <8 x double> undef to <8 x i16>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r145 = fptosi <8 x double> undef to <8 x i16>
-  ; CHECK: cost of 32 {{.*}} fptoui
+  ; CHECK: cost of 64 {{.*}} fptoui
   %r146 = fptoui <8 x double> undef to <8 x i32>
-  ; CHECK: cost of 32 {{.*}} fptosi
+  ; CHECK: cost of 64 {{.*}} fptosi
   %r147 = fptosi <8 x double> undef to <8 x i32>
-  ; CHECK: cost of 112 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r148 = fptoui <8 x double> undef to <8 x i64>
-  ; CHECK: cost of 112 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r149 = fptosi <8 x double> undef to <8 x i64>
 
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r150 = fptoui <16 x float> undef to <16 x i1>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r151 = fptosi <16 x float> undef to <16 x i1>
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r152 = fptoui <16 x float> undef to <16 x i8>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r153 = fptosi <16 x float> undef to <16 x i8>
   ; CHECK: cost of 8 {{.*}} fptoui
   %r154 = fptoui <16 x float> undef to <16 x i16>
@@ -347,30 +347,30 @@ define i32 @casts() {
   %r156 = fptoui <16 x float> undef to <16 x i32>
   ; CHECK: cost of 4 {{.*}} fptosi
   %r157 = fptosi <16 x float> undef to <16 x i32>
-  ; CHECK: cost of 224 {{.*}} fptoui
+  ; CHECK: cost of 256 {{.*}} fptoui
   %r158 = fptoui <16 x float> undef to <16 x i64>
-  ; CHECK: cost of 224 {{.*}} fptosi
+  ; CHECK: cost of 256 {{.*}} fptosi
   %r159 = fptosi <16 x float> undef to <16 x i64>
 
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r160 = fptoui <16 x double> undef to <16 x i1>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r161 = fptosi <16 x double> undef to <16 x i1>
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r162 = fptoui <16 x double> undef to <16 x i8>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r163 = fptosi <16 x double> undef to <16 x i8>
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r164 = fptoui <16 x double> undef to <16 x i16>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r165 = fptosi <16 x double> undef to <16 x i16>
-  ; CHECK: cost of 64 {{.*}} fptoui
+  ; CHECK: cost of 128 {{.*}} fptoui
   %r166 = fptoui <16 x double> undef to <16 x i32>
-  ; CHECK: cost of 64 {{.*}} fptosi
+  ; CHECK: cost of 128 {{.*}} fptosi
   %r167 = fptosi <16 x double> undef to <16 x i32>
-  ; CHECK: cost of 224 {{.*}} fptoui
+  ; CHECK: cost of 256 {{.*}} fptoui
   %r168 = fptoui <16 x double> undef to <16 x i64>
-  ; CHECK: cost of 224 {{.*}} fptosi
+  ; CHECK: cost of 256 {{.*}} fptosi
   %r169 = fptosi <16 x double> undef to <16 x i64>
 
   ; CHECK: cost of 8 {{.*}} uitofp
diff --git a/test/Analysis/CostModel/PowerPC/cmp-expanded.ll b/test/Analysis/CostModel/PowerPC/cmp-expanded.ll
new file mode 100644
index 0000000..38c8439
--- /dev/null
+++ b/test/Analysis/CostModel/PowerPC/cmp-expanded.ll
@@ -0,0 +1,14 @@
+; RUN: opt < %s -cost-model -analyze -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 -mattr=-vsx | FileCheck %s
+target datalayout = "E-m:e-i64:64-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+define void @exts() {
+
+   ; VSX is disabled, so this cost needs to include scalarization (because
+   ; <4 x double> is legalized to scalars).
+   ; CHECK: cost of 44 {{.*}} fcmp
+   %v1 = fcmp ugt <4 x double> undef, undef
+
+  ret void
+}
+
diff --git a/test/Analysis/CostModel/X86/cast.ll b/test/Analysis/CostModel/X86/cast.ll
index 7f97b17..fb16af6 100644
--- a/test/Analysis/CostModel/X86/cast.ll
+++ b/test/Analysis/CostModel/X86/cast.ll
@@ -1,3 +1,4 @@
+; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=knl | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-AVX512
 ; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=core-avx2 | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-AVX2
 ; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-AVX
 
@@ -83,6 +84,19 @@ define i32 @zext_sext(<8 x i1> %in) {
   ;CHECK-AVX: cost of 4 {{.*}} zext
   %D = zext <4 x i32> undef to <4 x i64>
 
+  ;CHECK-AVX512: cost of 3 {{.*}} %D1 = zext
+  %D1 = zext <16 x i32> undef to <16 x i64>
+
+  ;CHECK-AVX512: cost of 3 {{.*}} %D2 = sext
+  %D2 = sext <16 x i32> undef to <16 x i64>
+
+  ;CHECK-AVX512: cost of 1 {{.*}} %D3 = zext
+  %D3 = zext <16 x i16> undef to <16 x i32>
+  ;CHECK-AVX512: cost of 1 {{.*}} %D4 = zext
+  %D4 = zext <16 x i8> undef to <16 x i32>
+  ;CHECK-AVX512: cost of 2 {{.*}} %D5 = zext
+  %D5 = zext <16 x i1> undef to <16 x i32>
+
   ;CHECK-AVX2: cost of 2 {{.*}} trunc
   ;CHECK-AVX: cost of 4 {{.*}} trunc
   %E = trunc <4 x i64> undef to <4 x i32>
@@ -101,8 +115,12 @@ define i32 @zext_sext(<8 x i1> %in) {
 
   ;CHECK-AVX2: cost of 4 {{.*}} trunc
   ;CHECK-AVX: cost of 9 {{.*}} trunc
+  ;CHECK_AVX512: cost of 1 {{.*}} G = trunc
   %G = trunc <8 x i64> undef to <8 x i32>
 
+  ;CHECK-AVX512: cost of 4 {{.*}} %G1 = trunc
+  %G1 = trunc <16 x i64> undef to <16 x i32>
+
   ret i32 undef
 }
 
@@ -207,7 +225,30 @@ define void @uitofp8(<8 x i1> %a, <8 x i8> %b, <8 x i16> %c, <8 x i32> %d) {
   ; CHECK: cost of 5 {{.*}} uitofp
   %C1 = uitofp <8 x i16> %c to <8 x float>
 
-  ; CHECK: cost of 9 {{.*}} uitofp
+  ; CHECK-AVX2: cost of 8 {{.*}} uitofp
+  ; CHECK-AVX512: cost of 8 {{.*}} uitofp
+  ; CHECK-AVX: cost of 9 {{.*}} uitofp
   %D1 = uitofp <8 x i32> %d to <8 x float>
   ret void
 }
+
+define void @fp_conv(<8 x float> %a, <16 x float>%b) {
+;CHECK-LABEL: for function 'fp_conv'
+  ; CHECK-AVX512: cost of 1 {{.*}} fpext
+  %A1 = fpext <8 x float> %a to <8 x double>
+
+  ; CHECK-AVX512: cost of 3 {{.*}} fpext
+  %A2 = fpext <16 x float> %b to <16 x double>
+
+  ; CHECK-AVX2:   cost of 3 {{.*}} %A3 = fpext
+  ; CHECK-AVX512: cost of 1 {{.*}} %A3 = fpext
+  %A3 = fpext <8 x float> %a to <8 x double>
+
+  ; CHECK-AVX2:   cost of 3 {{.*}} %A4 = fptrunc
+  ; CHECK-AVX512: cost of 1 {{.*}} %A4 = fptrunc
+  %A4 = fptrunc <8 x double> undef to <8 x float>
+
+  ; CHECK-AVX512: cost of 3 {{.*}} %A5 = fptrunc
+  %A5 = fptrunc <16 x double> undef to <16 x float>
+  ret void
+}
diff --git a/test/Analysis/CostModel/X86/cmp.ll b/test/Analysis/CostModel/X86/cmp.ll
index 9f2bdb3..469cd73 100644
--- a/test/Analysis/CostModel/X86/cmp.ll
+++ b/test/Analysis/CostModel/X86/cmp.ll
@@ -1,5 +1,6 @@
 ; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck -check-prefix=CHECK -check-prefix=AVX1 %s
 ; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=core-avx2 | FileCheck -check-prefix=CHECK -check-prefix=AVX2 %s
+; RUN: opt < %s  -cost-model -analyze -mtriple=x86_64-apple-macosx10.8.0 -mcpu=knl | FileCheck -check-prefix=CHECK -check-prefix=AVX512 %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 target triple = "x86_64-apple-macosx10.8.0"
@@ -22,6 +23,11 @@ define i32 @cmp(i32 %arg) {
   ;AVX2: cost of 1 {{.*}} fcmp
   %E = fcmp olt <4 x double> undef, undef
 
+  ; AVX512: cost of 1 {{.*}} %E1 = fcmp
+  %E1 = fcmp olt <16 x float> undef, undef
+  ; AVX512: cost of 2 {{.*}} %E2 = fcmp
+  %E2 = fcmp olt <16 x double> undef, undef
+
   ;  -- integers --
 
   ;AVX1: cost of 1 {{.*}} icmp
@@ -49,6 +55,11 @@ define i32 @cmp(i32 %arg) {
   ;AVX2: cost of 1 {{.*}} icmp
   %M = icmp eq <32 x i8> undef, undef
 
+  ; AVX512: cost of 1 {{.*}} %M1 = icmp
+  %M1 = icmp eq <16 x i32> undef, undef
+  ; AVX512: cost of 2 {{.*}} %M2 = icmp
+  %M2 = icmp eq <16 x i64> undef, undef
+
   ;CHECK: cost of 0 {{.*}} ret
   ret i32 undef
 }
diff --git a/test/Analysis/CostModel/X86/sitofp.ll b/test/Analysis/CostModel/X86/sitofp.ll
index 338d974..edc937e 100644
--- a/test/Analysis/CostModel/X86/sitofp.ll
+++ b/test/Analysis/CostModel/X86/sitofp.ll
@@ -1,4 +1,5 @@
 ; RUN: opt -mtriple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s
+; RUN: opt -mtriple=x86_64-apple-darwin -mcpu=knl -cost-model -analyze < %s | FileCheck --check-prefix=AVX512F %s
 
 define <2 x double> @sitofpv2i8v2double(<2 x i8> %a) {
   ; SSE2: sitofpv2i8v2double
@@ -279,3 +280,47 @@ define <32 x float> @sitofpv32i64v32float(<32 x i64> %a) {
   %1 = sitofp <32 x i64> %a to <32 x float>
   ret <32 x float> %1
 }
+
+; AVX512F-LABEL: sitofp_16i8_float
+; AVX512F: cost of 2 {{.*}} sitofp
+define <16 x float> @sitofp_16i8_float(<16 x i8> %a) {
+  %1 = sitofp <16 x i8> %a to <16 x float>
+  ret <16 x float> %1
+}
+
+define <16 x float> @sitofp_16i16_float(<16 x i16> %a) {
+  ; AVX512F-LABEL: sitofp_16i16_float
+  ; AVX512F: cost of 2 {{.*}} sitofp
+  %1 = sitofp <16 x i16> %a to <16 x float>
+  ret <16 x float> %1
+}
+
+; AVX512F-LABEL: sitofp_8i8_double
+; AVX512F: cost of 2 {{.*}} sitofp
+define <8 x double> @sitofp_8i8_double(<8 x i8> %a) {
+  %1 = sitofp <8 x i8> %a to <8 x double>
+  ret <8 x double> %1
+}
+
+; AVX512F-LABEL: sitofp_8i16_double
+; AVX512F: cost of 2 {{.*}} sitofp
+define <8 x double> @sitofp_8i16_double(<8 x i16> %a) {
+  %1 = sitofp <8 x i16> %a to <8 x double>
+  ret <8 x double> %1
+}
+
+; AVX512F-LABEL: sitofp_8i1_double
+; AVX512F: cost of 4 {{.*}} sitofp
+define <8 x double> @sitofp_8i1_double(<8 x double> %a) {
+  %cmpres = fcmp ogt <8 x double> %a, zeroinitializer
+  %1 = sitofp <8 x i1> %cmpres to <8 x double>
+  ret <8 x double> %1
+}
+
+; AVX512F-LABEL: sitofp_16i1_float
+; AVX512F: cost of 3 {{.*}} sitofp
+define <16 x float> @sitofp_16i1_float(<16 x float> %a) {
+  %cmpres = fcmp ogt <16 x float> %a, zeroinitializer
+  %1 = sitofp <16 x i1> %cmpres to <16 x float>
+  ret <16 x float> %1
+}
diff --git a/test/Analysis/CostModel/X86/uitofp.ll b/test/Analysis/CostModel/X86/uitofp.ll
index a41a04d..27ec268 100644
--- a/test/Analysis/CostModel/X86/uitofp.ll
+++ b/test/Analysis/CostModel/X86/uitofp.ll
@@ -235,7 +235,7 @@ define <2 x float> @uitofpv2i8v2float(<2 x i8> %a) {
 
 define <4 x float> @uitofpv4i8v4float(<4 x i8> %a) {
   ; SSE2: uitofpv4i8v4float
-  ; SSE2: cost of 15 {{.*}} uitofp
+  ; SSE2: cost of 8 {{.*}} uitofp
   %1 = uitofp <4 x i8> %a to <4 x float>
   ret <4 x float> %1
 }
@@ -270,7 +270,7 @@ define <2 x float> @uitofpv2i16v2float(<2 x i16> %a) {
 
 define <4 x float> @uitofpv4i16v4float(<4 x i16> %a) {
   ; SSE2: uitofpv4i16v4float
-  ; SSE2: cost of 15 {{.*}} uitofp
+  ; SSE2: cost of 8 {{.*}} uitofp
   %1 = uitofp <4 x i16> %a to <4 x float>
   ret <4 x float> %1
 }
@@ -305,28 +305,28 @@ define <2 x float> @uitofpv2i32v2float(<2 x i32> %a) {
 
 define <4 x float> @uitofpv4i32v4float(<4 x i32> %a) {
   ; SSE2: uitofpv4i32v4float
-  ; SSE2: cost of 15 {{.*}} uitofp
+  ; SSE2: cost of 8 {{.*}} uitofp
   %1 = uitofp <4 x i32> %a to <4 x float>
   ret <4 x float> %1
 }
 
 define <8 x float> @uitofpv8i32v8float(<8 x i32> %a) {
   ; SSE2: uitofpv8i32v8float
-  ; SSE2: cost of 30 {{.*}} uitofp
+  ; SSE2: cost of 16 {{.*}} uitofp
   %1 = uitofp <8 x i32> %a to <8 x float>
   ret <8 x float> %1
 }
 
 define <16 x float> @uitofpv16i32v16float(<16 x i32> %a) {
   ; SSE2: uitofpv16i32v16float
-  ; SSE2: cost of 60 {{.*}} uitofp
+  ; SSE2: cost of 32 {{.*}} uitofp
   %1 = uitofp <16 x i32> %a to <16 x float>
   ret <16 x float> %1
 }
 
 define <32 x float> @uitofpv32i32v32float(<32 x i32> %a) {
   ; SSE2: uitofpv32i32v32float
-  ; SSE2: cost of 120 {{.*}} uitofp
+  ; SSE2: cost of 64 {{.*}} uitofp
   %1 = uitofp <32 x i32> %a to <32 x float>
   ret <32 x float> %1
 }