[llvm] 92593f9 - [SVE] Lower fixed length vXi32/vXi64 SDIV to scalable vectors.
Cameron McInally via llvm-commits
llvm-commits at lists.llvm.org
Fri Aug 14 16:47:32 PDT 2020
Author: Cameron McInally
Date: 2020-08-14T18:47:22-05:00
New Revision: 92593f9e77c387cee1ab5573b2f4b5c42162e8b8
URL: https://github.com/llvm/llvm-project/commit/92593f9e77c387cee1ab5573b2f4b5c42162e8b8
DIFF: https://github.com/llvm/llvm-project/commit/92593f9e77c387cee1ab5573b2f4b5c42162e8b8.diff
LOG: [SVE] Lower fixed length vXi32/vXi64 SDIV to scalable vectors.
Differential Revision: https://reviews.llvm.org/D85982
Added:
Modified:
llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
llvm/test/CodeGen/AArch64/sve-fixed-length-int-arith.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index 50da2bf1b713..eb87f6be73e9 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -997,6 +997,10 @@ AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
// These operations are not supported on NEON but SVE can do them.
setOperationAction(ISD::MUL, MVT::v1i64, Custom);
setOperationAction(ISD::MUL, MVT::v2i64, Custom);
+ setOperationAction(ISD::SDIV, MVT::v2i32, Custom);
+ setOperationAction(ISD::SDIV, MVT::v4i32, Custom);
+ setOperationAction(ISD::SDIV, MVT::v1i64, Custom);
+ setOperationAction(ISD::SDIV, MVT::v2i64, Custom);
setOperationAction(ISD::SMAX, MVT::v1i64, Custom);
setOperationAction(ISD::SMAX, MVT::v2i64, Custom);
setOperationAction(ISD::SMIN, MVT::v1i64, Custom);
@@ -1129,6 +1133,10 @@ void AArch64TargetLowering::addTypeForFixedLengthSVE(MVT VT) {
setOperationAction(ISD::UMIN, VT, Custom);
setOperationAction(ISD::XOR, VT, Custom);
setOperationAction(ISD::ZERO_EXTEND, VT, Custom);
+
+ if (VT.getVectorElementType() == MVT::i32 ||
+ VT.getVectorElementType() == MVT::i64)
+ setOperationAction(ISD::SDIV, VT, Custom);
}
void AArch64TargetLowering::addDRTypeForNEON(MVT VT) {
@@ -8925,6 +8933,12 @@ SDValue AArch64TargetLowering::LowerDIV(SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
bool Signed = Op.getOpcode() == ISD::SDIV;
unsigned PredOpcode = Signed ? AArch64ISD::SDIV_PRED : AArch64ISD::UDIV_PRED;
+
+ if (useSVEForFixedLengthVectorVT(Op.getValueType(), /*OverrideNEON=*/true) &&
+ (VT.getVectorElementType() == MVT::i32 ||
+ VT.getVectorElementType() == MVT::i64))
+ return LowerToPredicatedOp(Op, DAG, PredOpcode, /*OverrideNEON=*/true);
+
if (VT == MVT::nxv4i32 || VT == MVT::nxv2i64)
return LowerToPredicatedOp(Op, DAG, PredOpcode);
diff --git a/llvm/test/CodeGen/AArch64/sve-fixed-length-int-arith.ll b/llvm/test/CodeGen/AArch64/sve-fixed-length-int-arith.ll
index a42f89d90e21..e4a0ced83e26 100644
--- a/llvm/test/CodeGen/AArch64/sve-fixed-length-int-arith.ll
+++ b/llvm/test/CodeGen/AArch64/sve-fixed-length-int-arith.ll
@@ -419,6 +419,170 @@ define void @add_v32i64(<32 x i64>* %a, <32 x i64>* %b) #0 {
; VBYTES because the add tests already validate the legalisation code paths.
;
+;
+; SDIV
+;
+
+; Vector v2i32 sdiv are not legal for NEON so use SVE when available.
+define <2 x i32> @sdiv_v2i32(<2 x i32> %op1, <2 x i32> %op2) #0 {
+; CHECK-LABEL: sdiv_v2i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),2)]]
+; CHECK: sdiv z0.s, [[PG]]/m, z0.s, z1.s
+; CHECK: ret
+ %res = sdiv <2 x i32> %op1, %op2
+ ret <2 x i32> %res
+}
+
+; Vector v4i32 sdiv are not legal for NEON so use SVE when available.
+define <4 x i32> @sdiv_v4i32(<4 x i32> %op1, <4 x i32> %op2) #0 {
+; CHECK-LABEL: sdiv_v4i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),4)]]
+; CHECK: sdiv z0.s, [[PG]]/m, z0.s, z1.s
+; CHECK: ret
+ %res = sdiv <4 x i32> %op1, %op2
+ ret <4 x i32> %res
+}
+
+define void @sdiv_v8i32(<8 x i32>* %a, <8 x i32>* %b) #0 {
+; CHECK-LABEL: sdiv_v8i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),8)]]
+; CHECK-DAG: ld1w { [[OP1:z[0-9]+]].s }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1w { [[OP2:z[0-9]+]].s }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].s, [[PG]]/m, [[OP1]].s, [[OP2]].s
+; CHECK: st1w { [[RES]].s }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <8 x i32>, <8 x i32>* %a
+ %op2 = load <8 x i32>, <8 x i32>* %b
+ %res = sdiv <8 x i32> %op1, %op2
+ store <8 x i32> %res, <8 x i32>* %a
+ ret void
+}
+
+define void @sdiv_v16i32(<16 x i32>* %a, <16 x i32>* %b) #0 {
+; CHECK-LABEL: sdiv_v16i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),16)]]
+; CHECK-DAG: ld1w { [[OP1:z[0-9]+]].s }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1w { [[OP2:z[0-9]+]].s }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].s, [[PG]]/m, [[OP1]].s, [[OP2]].s
+; CHECK: st1w { [[RES]].s }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <16 x i32>, <16 x i32>* %a
+ %op2 = load <16 x i32>, <16 x i32>* %b
+ %res = sdiv <16 x i32> %op1, %op2
+ store <16 x i32> %res, <16 x i32>* %a
+ ret void
+}
+
+define void @sdiv_v32i32(<32 x i32>* %a, <32 x i32>* %b) #0 {
+; CHECK-LABEL: sdiv_v32i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),32)]]
+; CHECK-DAG: ld1w { [[OP1:z[0-9]+]].s }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1w { [[OP2:z[0-9]+]].s }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].s, [[PG]]/m, [[OP1]].s, [[OP2]].s
+; CHECK: st1w { [[RES]].s }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <32 x i32>, <32 x i32>* %a
+ %op2 = load <32 x i32>, <32 x i32>* %b
+ %res = sdiv <32 x i32> %op1, %op2
+ store <32 x i32> %res, <32 x i32>* %a
+ ret void
+}
+
+define void @sdiv_v64i32(<64 x i32>* %a, <64 x i32>* %b) #0 {
+; CHECK-LABEL: sdiv_v64i32:
+; CHECK: ptrue [[PG:p[0-9]+]].s, vl[[#min(div(VBYTES,4),64)]]
+; CHECK-DAG: ld1w { [[OP1:z[0-9]+]].s }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1w { [[OP2:z[0-9]+]].s }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].s, [[PG]]/m, [[OP1]].s, [[OP2]].s
+; CHECK: st1w { [[RES]].s }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <64 x i32>, <64 x i32>* %a
+ %op2 = load <64 x i32>, <64 x i32>* %b
+ %res = sdiv <64 x i32> %op1, %op2
+ store <64 x i32> %res, <64 x i32>* %a
+ ret void
+}
+
+; Vector i64 sdiv are not legal for NEON so use SVE when available.
+define <1 x i64> @sdiv_v1i64(<1 x i64> %op1, <1 x i64> %op2) #0 {
+; CHECK-LABEL: sdiv_v1i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl1
+; CHECK: sdiv z0.d, [[PG]]/m, z0.d, z1.d
+; CHECK: ret
+ %res = sdiv <1 x i64> %op1, %op2
+ ret <1 x i64> %res
+}
+
+; Vector i64 sdiv are not legal for NEON so use SVE when available.
+define <2 x i64> @sdiv_v2i64(<2 x i64> %op1, <2 x i64> %op2) #0 {
+; CHECK-LABEL: sdiv_v2i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl2
+; CHECK: sdiv z0.d, [[PG]]/m, z0.d, z1.d
+; CHECK: ret
+ %res = sdiv <2 x i64> %op1, %op2
+ ret <2 x i64> %res
+}
+
+define void @sdiv_v4i64(<4 x i64>* %a, <4 x i64>* %b) #0 {
+; CHECK-LABEL: sdiv_v4i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl[[#min(div(VBYTES,8),4)]]
+; CHECK-DAG: ld1d { [[OP1:z[0-9]+]].d }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1d { [[OP2:z[0-9]+]].d }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].d, [[PG]]/m, [[OP1]].d, [[OP2]].d
+; CHECK: st1d { [[RES]].d }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <4 x i64>, <4 x i64>* %a
+ %op2 = load <4 x i64>, <4 x i64>* %b
+ %res = sdiv <4 x i64> %op1, %op2
+ store <4 x i64> %res, <4 x i64>* %a
+ ret void
+}
+
+define void @sdiv_v8i64(<8 x i64>* %a, <8 x i64>* %b) #0 {
+; CHECK-LABEL: sdiv_v8i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl[[#min(div(VBYTES,8),8)]]
+; CHECK-DAG: ld1d { [[OP1:z[0-9]+]].d }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1d { [[OP2:z[0-9]+]].d }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].d, [[PG]]/m, [[OP1]].d, [[OP2]].d
+; CHECK: st1d { [[RES]].d }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <8 x i64>, <8 x i64>* %a
+ %op2 = load <8 x i64>, <8 x i64>* %b
+ %res = sdiv <8 x i64> %op1, %op2
+ store <8 x i64> %res, <8 x i64>* %a
+ ret void
+}
+
+define void @sdiv_v16i64(<16 x i64>* %a, <16 x i64>* %b) #0 {
+; CHECK-LABEL: sdiv_v16i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl[[#min(div(VBYTES,8),16)]]
+; CHECK-DAG: ld1d { [[OP1:z[0-9]+]].d }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1d { [[OP2:z[0-9]+]].d }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].d, [[PG]]/m, [[OP1]].d, [[OP2]].d
+; CHECK: st1d { [[RES]].d }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <16 x i64>, <16 x i64>* %a
+ %op2 = load <16 x i64>, <16 x i64>* %b
+ %res = sdiv <16 x i64> %op1, %op2
+ store <16 x i64> %res, <16 x i64>* %a
+ ret void
+}
+
+define void @sdiv_v32i64(<32 x i64>* %a, <32 x i64>* %b) #0 {
+; CHECK-LABEL: sdiv_v32i64:
+; CHECK: ptrue [[PG:p[0-9]+]].d, vl[[#min(div(VBYTES,8),32)]]
+; CHECK-DAG: ld1d { [[OP1:z[0-9]+]].d }, [[PG]]/z, [x0]
+; CHECK-DAG: ld1d { [[OP2:z[0-9]+]].d }, [[PG]]/z, [x1]
+; CHECK: sdiv [[RES:z[0-9]+]].d, [[PG]]/m, [[OP1]].d, [[OP2]].d
+; CHECK: st1d { [[RES]].d }, [[PG]], [x0]
+; CHECK: ret
+ %op1 = load <32 x i64>, <32 x i64>* %a
+ %op2 = load <32 x i64>, <32 x i64>* %b
+ %res = sdiv <32 x i64> %op1, %op2
+ store <32 x i64> %res, <32 x i64>* %a
+ ret void
+}
+
;
; MUL
;
More information about the llvm-commits
mailing list