[llvm] r268872 - [X86] Add patterns for 256-bit non-temporal stores when only AVX1 is supported. While there, add a predicate to the SSE2 patterns to avoid an ordering dependency.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Sun May 8 00:10:51 PDT 2016
Author: ctopper
Date: Sun May 8 02:10:50 2016
New Revision: 268872
URL: http://llvm.org/viewvc/llvm-project?rev=268872&view=rev
Log:
[X86] Add patterns for 256-bit non-temporal stores when only AVX1 is supported. While there, add a predicate to the SSE2 patterns to avoid an ordering dependency.
Modified:
llvm/trunk/lib/Target/X86/X86InstrSSE.td
llvm/trunk/test/CodeGen/X86/nontemporal-2.ll
Modified: llvm/trunk/lib/Target/X86/X86InstrSSE.td
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrSSE.td?rev=268872&r1=268871&r2=268872&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86InstrSSE.td (original)
+++ llvm/trunk/lib/Target/X86/X86InstrSSE.td Sun May 8 02:10:50 2016
@@ -3648,6 +3648,18 @@ let Predicates = [HasAVX2, NoVLX] in {
(VMOVNTDQYmr addr:$dst, VR256:$src)>;
}
+// If integer type isn't available, use the floating point type.
+let Predicates = [HasAVX1Only] in {
+ def : Pat<(alignednontemporalstore (v4i64 VR256:$src), addr:$dst),
+ (VMOVNTPSYmr addr:$dst, VR256:$src)>;
+ def : Pat<(alignednontemporalstore (v8i32 VR256:$src), addr:$dst),
+ (VMOVNTPSYmr addr:$dst, VR256:$src)>;
+ def : Pat<(alignednontemporalstore (v16i16 VR256:$src), addr:$dst),
+ (VMOVNTPSYmr addr:$dst, VR256:$src)>;
+ def : Pat<(alignednontemporalstore (v32i8 VR256:$src), addr:$dst),
+ (VMOVNTPSYmr addr:$dst, VR256:$src)>;
+}
+
let Predicates = [HasAVX, NoVLX] in {
def : Pat<(alignednontemporalstore (v4i32 VR128:$src), addr:$dst),
(VMOVNTDQmr addr:$dst, VR128:$src)>;
@@ -3657,12 +3669,14 @@ let Predicates = [HasAVX, NoVLX] in {
(VMOVNTDQmr addr:$dst, VR128:$src)>;
}
-def : Pat<(alignednontemporalstore (v4i32 VR128:$src), addr:$dst),
- (MOVNTDQmr addr:$dst, VR128:$src)>;
-def : Pat<(alignednontemporalstore (v8i16 VR128:$src), addr:$dst),
- (MOVNTDQmr addr:$dst, VR128:$src)>;
-def : Pat<(alignednontemporalstore (v16i8 VR128:$src), addr:$dst),
- (MOVNTDQmr addr:$dst, VR128:$src)>;
+let Predicates = [UseSSE2] in {
+ def : Pat<(alignednontemporalstore (v4i32 VR128:$src), addr:$dst),
+ (MOVNTDQmr addr:$dst, VR128:$src)>;
+ def : Pat<(alignednontemporalstore (v8i16 VR128:$src), addr:$dst),
+ (MOVNTDQmr addr:$dst, VR128:$src)>;
+ def : Pat<(alignednontemporalstore (v16i8 VR128:$src), addr:$dst),
+ (MOVNTDQmr addr:$dst, VR128:$src)>;
+}
} // AddedComplexity
Modified: llvm/trunk/test/CodeGen/X86/nontemporal-2.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/nontemporal-2.ll?rev=268872&r1=268871&r2=268872&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/nontemporal-2.ll (original)
+++ llvm/trunk/test/CodeGen/X86/nontemporal-2.ll Sun May 8 02:10:50 2016
@@ -9,6 +9,7 @@ define void @test_zero_v4f32(<4 x float>
; CHECK-LABEL: test_zero_v4f32:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <4 x float> zeroinitializer, <4 x float>* %dst, align 16, !nontemporal !1
ret void
}
@@ -17,6 +18,8 @@ define void @test_zero_v4i32(<4 x i32>*
; CHECK-LABEL: test_zero_v4i32:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
+ store <4 x i32> zeroinitializer, <4 x i32>* %dst, align 16, !nontemporal !1
store <4 x i32> zeroinitializer, <4 x i32>* %dst, align 16, !nontemporal !1
ret void
}
@@ -25,6 +28,7 @@ define void @test_zero_v2f64(<2 x double
; CHECK-LABEL: test_zero_v2f64:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <2 x double> zeroinitializer, <2 x double>* %dst, align 16, !nontemporal !1
ret void
}
@@ -33,6 +37,7 @@ define void @test_zero_v2i64(<2 x i64>*
; CHECK-LABEL: test_zero_v2i64:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <2 x i64> zeroinitializer, <2 x i64>* %dst, align 16, !nontemporal !1
ret void
}
@@ -41,6 +46,7 @@ define void @test_zero_v8i16(<8 x i16>*
; CHECK-LABEL: test_zero_v8i16:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <8 x i16> zeroinitializer, <8 x i16>* %dst, align 16, !nontemporal !1
ret void
}
@@ -49,6 +55,7 @@ define void @test_zero_v16i8(<16 x i8>*
; CHECK-LABEL: test_zero_v16i8:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <16 x i8> zeroinitializer, <16 x i8>* %dst, align 16, !nontemporal !1
ret void
}
@@ -58,12 +65,14 @@ define void @test_zero_v16i8(<16 x i8>*
define void @test_zero_v8f32(<8 x float>* %dst) {
; CHECK-LABEL: test_zero_v8f32:
; AVX: vmovntps %ymm
+; AVX2: vmovntps %ymm
store <8 x float> zeroinitializer, <8 x float>* %dst, align 32, !nontemporal !1
ret void
}
define void @test_zero_v8i32(<8 x i32>* %dst) {
; CHECK-LABEL: test_zero_v8i32:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <8 x i32> zeroinitializer, <8 x i32>* %dst, align 32, !nontemporal !1
ret void
@@ -72,12 +81,14 @@ define void @test_zero_v8i32(<8 x i32>*
define void @test_zero_v4f64(<4 x double>* %dst) {
; CHECK-LABEL: test_zero_v4f64:
; AVX: vmovntps %ymm
+; AVX2: vmovntps %ymm
store <4 x double> zeroinitializer, <4 x double>* %dst, align 32, !nontemporal !1
ret void
}
define void @test_zero_v4i64(<4 x i64>* %dst) {
; CHECK-LABEL: test_zero_v4i64:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <4 x i64> zeroinitializer, <4 x i64>* %dst, align 32, !nontemporal !1
ret void
@@ -85,6 +96,7 @@ define void @test_zero_v4i64(<4 x i64>*
define void @test_zero_v16i16(<16 x i16>* %dst) {
; CHECK-LABEL: test_zero_v16i16:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <16 x i16> zeroinitializer, <16 x i16>* %dst, align 32, !nontemporal !1
ret void
@@ -92,6 +104,7 @@ define void @test_zero_v16i16(<16 x i16>
define void @test_zero_v32i8(<32 x i8>* %dst) {
; CHECK-LABEL: test_zero_v32i8:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <32 x i8> zeroinitializer, <32 x i8>* %dst, align 32, !nontemporal !1
ret void
@@ -104,6 +117,7 @@ define void @test_arg_v4f32(<4 x float>
; CHECK-LABEL: test_arg_v4f32:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <4 x float> %arg, <4 x float>* %dst, align 16, !nontemporal !1
ret void
}
@@ -112,6 +126,7 @@ define void @test_arg_v4i32(<4 x i32> %a
; CHECK-LABEL: test_arg_v4i32:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <4 x i32> %arg, <4 x i32>* %dst, align 16, !nontemporal !1
ret void
}
@@ -120,6 +135,7 @@ define void @test_arg_v2f64(<2 x double>
; CHECK-LABEL: test_arg_v2f64:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <2 x double> %arg, <2 x double>* %dst, align 16, !nontemporal !1
ret void
}
@@ -128,6 +144,7 @@ define void @test_arg_v2i64(<2 x i64> %a
; CHECK-LABEL: test_arg_v2i64:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <2 x i64> %arg, <2 x i64>* %dst, align 16, !nontemporal !1
ret void
}
@@ -136,6 +153,7 @@ define void @test_arg_v8i16(<8 x i16> %a
; CHECK-LABEL: test_arg_v8i16:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <8 x i16> %arg, <8 x i16>* %dst, align 16, !nontemporal !1
ret void
}
@@ -144,6 +162,7 @@ define void @test_arg_v16i8(<16 x i8> %a
; CHECK-LABEL: test_arg_v16i8:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
store <16 x i8> %arg, <16 x i8>* %dst, align 16, !nontemporal !1
ret void
}
@@ -153,12 +172,14 @@ define void @test_arg_v16i8(<16 x i8> %a
define void @test_arg_v8f32(<8 x float> %arg, <8 x float>* %dst) {
; CHECK-LABEL: test_arg_v8f32:
; AVX: vmovntps %ymm
+; AVX2: vmovntps %ymm
store <8 x float> %arg, <8 x float>* %dst, align 32, !nontemporal !1
ret void
}
define void @test_arg_v8i32(<8 x i32> %arg, <8 x i32>* %dst) {
; CHECK-LABEL: test_arg_v8i32:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <8 x i32> %arg, <8 x i32>* %dst, align 32, !nontemporal !1
ret void
@@ -167,12 +188,14 @@ define void @test_arg_v8i32(<8 x i32> %a
define void @test_arg_v4f64(<4 x double> %arg, <4 x double>* %dst) {
; CHECK-LABEL: test_arg_v4f64:
; AVX: vmovntps %ymm
+; AVX2: vmovntps %ymm
store <4 x double> %arg, <4 x double>* %dst, align 32, !nontemporal !1
ret void
}
define void @test_arg_v4i64(<4 x i64> %arg, <4 x i64>* %dst) {
; CHECK-LABEL: test_arg_v4i64:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <4 x i64> %arg, <4 x i64>* %dst, align 32, !nontemporal !1
ret void
@@ -180,6 +203,7 @@ define void @test_arg_v4i64(<4 x i64> %a
define void @test_arg_v16i16(<16 x i16> %arg, <16 x i16>* %dst) {
; CHECK-LABEL: test_arg_v16i16:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <16 x i16> %arg, <16 x i16>* %dst, align 32, !nontemporal !1
ret void
@@ -187,6 +211,7 @@ define void @test_arg_v16i16(<16 x i16>
define void @test_arg_v32i8(<32 x i8> %arg, <32 x i8>* %dst) {
; CHECK-LABEL: test_arg_v32i8:
+; AVX: vmovntps %ymm
; AVX2: vmovntps %ymm
store <32 x i8> %arg, <32 x i8>* %dst, align 32, !nontemporal !1
ret void
@@ -200,6 +225,7 @@ define void @test_op_v4f32(<4 x float> %
; CHECK-LABEL: test_op_v4f32:
; SSE: movntps
; AVX: vmovntps
+; AVX2: vmovntps
%r = fadd <4 x float> %a, %b
store <4 x float> %r, <4 x float>* %dst, align 16, !nontemporal !1
ret void
@@ -209,6 +235,7 @@ define void @test_op_v4i32(<4 x i32> %a,
; CHECK-LABEL: test_op_v4i32:
; SSE: movntdq
; AVX: vmovntdq
+; AVX2: vmovntdq
%r = add <4 x i32> %a, %b
store <4 x i32> %r, <4 x i32>* %dst, align 16, !nontemporal !1
ret void
@@ -218,6 +245,7 @@ define void @test_op_v2f64(<2 x double>
; CHECK-LABEL: test_op_v2f64:
; SSE: movntpd
; AVX: vmovntpd
+; AVX2: vmovntpd
%r = fadd <2 x double> %a, %b
store <2 x double> %r, <2 x double>* %dst, align 16, !nontemporal !1
ret void
@@ -227,6 +255,7 @@ define void @test_op_v2i64(<2 x i64> %a,
; CHECK-LABEL: test_op_v2i64:
; SSE: movntdq
; AVX: vmovntdq
+; AVX2: vmovntdq
%r = add <2 x i64> %a, %b
store <2 x i64> %r, <2 x i64>* %dst, align 16, !nontemporal !1
ret void
@@ -236,6 +265,7 @@ define void @test_op_v8i16(<8 x i16> %a,
; CHECK-LABEL: test_op_v8i16:
; SSE: movntdq
; AVX: vmovntdq
+; AVX2: vmovntdq
%r = add <8 x i16> %a, %b
store <8 x i16> %r, <8 x i16>* %dst, align 16, !nontemporal !1
ret void
@@ -245,6 +275,7 @@ define void @test_op_v16i8(<16 x i8> %a,
; CHECK-LABEL: test_op_v16i8:
; SSE: movntdq
; AVX: vmovntdq
+; AVX2: vmovntdq
%r = add <16 x i8> %a, %b
store <16 x i8> %r, <16 x i8>* %dst, align 16, !nontemporal !1
ret void
@@ -255,6 +286,7 @@ define void @test_op_v16i8(<16 x i8> %a,
define void @test_op_v8f32(<8 x float> %a, <8 x float> %b, <8 x float>* %dst) {
; CHECK-LABEL: test_op_v8f32:
; AVX: vmovntps %ymm
+; AVX2: vmovntps %ymm
%r = fadd <8 x float> %a, %b
store <8 x float> %r, <8 x float>* %dst, align 32, !nontemporal !1
ret void
@@ -262,6 +294,7 @@ define void @test_op_v8f32(<8 x float> %
define void @test_op_v8i32(<8 x i32> %a, <8 x i32> %b, <8 x i32>* %dst) {
; CHECK-LABEL: test_op_v8i32:
+; AVX: vmovntps %ymm
; AVX2: vmovntdq %ymm
%r = add <8 x i32> %a, %b
store <8 x i32> %r, <8 x i32>* %dst, align 32, !nontemporal !1
@@ -271,6 +304,7 @@ define void @test_op_v8i32(<8 x i32> %a,
define void @test_op_v4f64(<4 x double> %a, <4 x double> %b, <4 x double>* %dst) {
; CHECK-LABEL: test_op_v4f64:
; AVX: vmovntpd %ymm
+; AVX2: vmovntpd %ymm
%r = fadd <4 x double> %a, %b
store <4 x double> %r, <4 x double>* %dst, align 32, !nontemporal !1
ret void
@@ -278,6 +312,7 @@ define void @test_op_v4f64(<4 x double>
define void @test_op_v4i64(<4 x i64> %a, <4 x i64> %b, <4 x i64>* %dst) {
; CHECK-LABEL: test_op_v4i64:
+; AVX: vmovntps %ymm
; AVX2: vmovntdq %ymm
%r = add <4 x i64> %a, %b
store <4 x i64> %r, <4 x i64>* %dst, align 32, !nontemporal !1
@@ -286,6 +321,7 @@ define void @test_op_v4i64(<4 x i64> %a,
define void @test_op_v16i16(<16 x i16> %a, <16 x i16> %b, <16 x i16>* %dst) {
; CHECK-LABEL: test_op_v16i16:
+; AVX: vmovntps %ymm
; AVX2: vmovntdq %ymm
%r = add <16 x i16> %a, %b
store <16 x i16> %r, <16 x i16>* %dst, align 32, !nontemporal !1
@@ -294,6 +330,7 @@ define void @test_op_v16i16(<16 x i16> %
define void @test_op_v32i8(<32 x i8> %a, <32 x i8> %b, <32 x i8>* %dst) {
; CHECK-LABEL: test_op_v32i8:
+; AVX: vmovntps %ymm
; AVX2: vmovntdq %ymm
%r = add <32 x i8> %a, %b
store <32 x i8> %r, <32 x i8>* %dst, align 32, !nontemporal !1
@@ -310,6 +347,8 @@ define void @test_unaligned_v8f32(<8 x f
; SSE: movntps %xmm
; AVX-NOT: movnt
; AVX: vmovups %ymm
+; AVX2-NOT: movnt
+; AVX2: vmovups %ymm
%r = fadd <8 x float> %a, %b
store <8 x float> %r, <8 x float>* %dst, align 16, !nontemporal !1
ret void
More information about the llvm-commits
mailing list