RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 19:52:54 +01:00
Code Issues Projects Releases Wiki Activity

[RISCV] Select vector shl by 1 to a vector add.

Browse Source 
A vector add may be faster than a vector shift.

Reviewed By: frasercrmck

Differential Revision: https://reviews.llvm.org/D106689
This commit is contained in:
Craig Topper 2021-07-27 09:48:18 -07:00
parent 0f86a54b34
commit 6bfc6b8665
11 changed files with 219 additions and 102 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
lib/Target/RISCV/RISCVInstrInfoVSDPatterns.td
View File

@ -429,6 +429,21 @@ defm : VPatBinarySDNode_VV_VX_VI<shl, "PseudoVSLL", uimm5>;
defm : VPatBinarySDNode_VV_VX_VI<srl, "PseudoVSRL", uimm5>;
defm : VPatBinarySDNode_VV_VX_VI<sra, "PseudoVSRA", uimm5>;
foreach vti = AllIntegerVectors in {
// Emit shift by 1 as an add since it might be faster.
def : Pat<(shl (vti.Vector vti.RegClass:$rs1),
(vti.Vector (splat_vector (XLenVT 1)))),
(!cast<Instruction>("PseudoVADD_VV_"# vti.LMul.MX)
vti.RegClass:$rs1, vti.RegClass:$rs1, vti.AVL, vti.Log2SEW)>;
}
foreach vti = [VI64M1, VI64M2, VI64M4, VI64M8] in {
def : Pat<(shl (vti.Vector vti.RegClass:$rs1),
(vti.Vector (rv32_splat_i64 (XLenVT 1)))),
(!cast<Instruction>("PseudoVADD_VV_"# vti.LMul.MX)
vti.RegClass:$rs1, vti.RegClass:$rs1, vti.AVL, vti.Log2SEW)>;
}
// 12.8. Vector Integer Comparison Instructions
defm : VPatIntegerSetCCSDNode_VV_VX_VI<SETEQ, "PseudoVMSEQ">;
defm : VPatIntegerSetCCSDNode_VV_VX_VI<SETNE, "PseudoVMSNE">;

10
lib/Target/RISCV/RISCVInstrInfoVVLPatterns.td
View File

@ -683,7 +683,15 @@ defm : VPatBinaryVL_VV_VX_VI<riscv_shl_vl, "PseudoVSLL", uimm5>;
defm : VPatBinaryVL_VV_VX_VI<riscv_srl_vl, "PseudoVSRL", uimm5>;
defm : VPatBinaryVL_VV_VX_VI<riscv_sra_vl, "PseudoVSRA", uimm5>;
foreach vti = AllIntegerVectors in {
// Emit shift by 1 as an add since it might be faster.
def : Pat<(riscv_shl_vl (vti.Vector vti.RegClass:$rs1),
(riscv_vmv_v_x_vl 1, (XLenVT srcvalue)),
(vti.Mask true_mask),
VLOpFrag),
(!cast<Instruction>("PseudoVADD_VV_"# vti.LMul.MX)
vti.RegClass:$rs1, vti.RegClass:$rs1, GPR:$vl, vti.Log2SEW)>;
}
// 12.7. Vector Narrowing Integer Right Shift Instructions
foreach vtiTowti = AllWidenableIntVectors in {

60
test/CodeGen/RISCV/rvv/fixed-vectors-bitreverse.ll
View File

@ -33,7 +33,7 @@ define void @bitreverse_v8i16(<8 x i16>* %x, <8 x i16>* %y) {
; LMULMAX2-RV32-NEXT: lui a1, 5
; LMULMAX2-RV32-NEXT: addi a1, a1, 1365
; LMULMAX2-RV32-NEXT: vand.vx v26, v25, a1
; LMULMAX2-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV32-NEXT: lui a1, 11
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vand.vx v25, v25, a1
@ -70,7 +70,7 @@ define void @bitreverse_v8i16(<8 x i16>* %x, <8 x i16>* %y) {
; LMULMAX2-RV64-NEXT: lui a1, 5
; LMULMAX2-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX2-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV64-NEXT: lui a1, 11
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1366
; LMULMAX2-RV64-NEXT: vand.vx v25, v25, a1
@ -107,7 +107,7 @@ define void @bitreverse_v8i16(<8 x i16>* %x, <8 x i16>* %y) {
; LMULMAX1-RV32-NEXT: lui a1, 5
; LMULMAX1-RV32-NEXT: addi a1, a1, 1365
; LMULMAX1-RV32-NEXT: vand.vx v26, v25, a1
; LMULMAX1-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV32-NEXT: lui a1, 11
; LMULMAX1-RV32-NEXT: addi a1, a1, -1366
; LMULMAX1-RV32-NEXT: vand.vx v25, v25, a1
@ -144,7 +144,7 @@ define void @bitreverse_v8i16(<8 x i16>* %x, <8 x i16>* %y) {
; LMULMAX1-RV64-NEXT: lui a1, 5
; LMULMAX1-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX1-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX1-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV64-NEXT: lui a1, 11
; LMULMAX1-RV64-NEXT: addiw a1, a1, -1366
; LMULMAX1-RV64-NEXT: vand.vx v25, v25, a1
@ -198,7 +198,7 @@ define void @bitreverse_v4i32(<4 x i32>* %x, <4 x i32>* %y) {
; LMULMAX2-RV32-NEXT: lui a1, 349525
; LMULMAX2-RV32-NEXT: addi a1, a1, 1365
; LMULMAX2-RV32-NEXT: vand.vx v26, v25, a1
; LMULMAX2-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV32-NEXT: lui a1, 699051
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vand.vx v25, v25, a1
@ -248,7 +248,7 @@ define void @bitreverse_v4i32(<4 x i32>* %x, <4 x i32>* %y) {
; LMULMAX2-RV64-NEXT: lui a1, 349525
; LMULMAX2-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX2-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV64-NEXT: lui a1, 171
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
@ -296,7 +296,7 @@ define void @bitreverse_v4i32(<4 x i32>* %x, <4 x i32>* %y) {
; LMULMAX1-RV32-NEXT: lui a1, 349525
; LMULMAX1-RV32-NEXT: addi a1, a1, 1365
; LMULMAX1-RV32-NEXT: vand.vx v26, v25, a1
; LMULMAX1-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV32-NEXT: lui a1, 699051
; LMULMAX1-RV32-NEXT: addi a1, a1, -1366
; LMULMAX1-RV32-NEXT: vand.vx v25, v25, a1
@ -346,7 +346,7 @@ define void @bitreverse_v4i32(<4 x i32>* %x, <4 x i32>* %y) {
; LMULMAX1-RV64-NEXT: lui a1, 349525
; LMULMAX1-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX1-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX1-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV64-NEXT: lui a1, 171
; LMULMAX1-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX1-RV64-NEXT: slli a1, a1, 12
@ -452,7 +452,7 @@ define void @bitreverse_v2i64(<2 x i64>* %x, <2 x i64>* %y) {
; LMULMAX2-RV32-NEXT: vmv.v.x v26, a1
; LMULMAX2-RV32-NEXT: vsetivli zero, 2, e64, m1, ta, mu
; LMULMAX2-RV32-NEXT: vand.vv v26, v25, v26
; LMULMAX2-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV32-NEXT: lui a1, 699051
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vsetivli zero, 4, e32, m1, ta, mu
@ -550,7 +550,7 @@ define void @bitreverse_v2i64(<2 x i64>* %x, <2 x i64>* %y) {
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
; LMULMAX2-RV64-NEXT: addi a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX2-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX2-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX2-RV64-NEXT: lui a1, 1026731
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
@ -652,7 +652,7 @@ define void @bitreverse_v2i64(<2 x i64>* %x, <2 x i64>* %y) {
; LMULMAX1-RV32-NEXT: vmv.v.x v26, a1
; LMULMAX1-RV32-NEXT: vsetivli zero, 2, e64, m1, ta, mu
; LMULMAX1-RV32-NEXT: vand.vv v26, v25, v26
; LMULMAX1-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV32-NEXT: lui a1, 699051
; LMULMAX1-RV32-NEXT: addi a1, a1, -1366
; LMULMAX1-RV32-NEXT: vsetivli zero, 4, e32, m1, ta, mu
@ -750,7 +750,7 @@ define void @bitreverse_v2i64(<2 x i64>* %x, <2 x i64>* %y) {
; LMULMAX1-RV64-NEXT: slli a1, a1, 12
; LMULMAX1-RV64-NEXT: addi a1, a1, 1365
; LMULMAX1-RV64-NEXT: vand.vx v26, v25, a1
; LMULMAX1-RV64-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV64-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV64-NEXT: lui a1, 1026731
; LMULMAX1-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX1-RV64-NEXT: slli a1, a1, 12
@ -801,7 +801,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX2-RV32-NEXT: lui a1, 5
; LMULMAX2-RV32-NEXT: addi a1, a1, 1365
; LMULMAX2-RV32-NEXT: vand.vx v28, v26, a1
; LMULMAX2-RV32-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV32-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV32-NEXT: lui a1, 11
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vand.vx v26, v26, a1
@ -838,7 +838,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX2-RV64-NEXT: lui a1, 5
; LMULMAX2-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v28, v26, a1
; LMULMAX2-RV64-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV64-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV64-NEXT: lui a1, 11
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1366
; LMULMAX2-RV64-NEXT: vand.vx v26, v26, a1
@ -877,7 +877,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX1-RV32-NEXT: lui a1, 5
; LMULMAX1-RV32-NEXT: addi a1, a1, 1365
; LMULMAX1-RV32-NEXT: vand.vx v27, v25, a1
; LMULMAX1-RV32-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV32-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV32-NEXT: lui a2, 11
; LMULMAX1-RV32-NEXT: addi a2, a2, -1366
; LMULMAX1-RV32-NEXT: vand.vx v25, v25, a2
@ -897,7 +897,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX1-RV32-NEXT: vsrl.vi v26, v26, 2
; LMULMAX1-RV32-NEXT: vor.vv v26, v26, v27
; LMULMAX1-RV32-NEXT: vand.vx v27, v26, a1
; LMULMAX1-RV32-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV32-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV32-NEXT: vand.vx v26, v26, a2
; LMULMAX1-RV32-NEXT: vsrl.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vor.vv v26, v26, v27
@ -935,7 +935,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX1-RV64-NEXT: lui a1, 5
; LMULMAX1-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX1-RV64-NEXT: vand.vx v27, v25, a1
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: lui a2, 11
; LMULMAX1-RV64-NEXT: addiw a2, a2, -1366
; LMULMAX1-RV64-NEXT: vand.vx v25, v25, a2
@ -955,7 +955,7 @@ define void @bitreverse_v16i16(<16 x i16>* %x, <16 x i16>* %y) {
; LMULMAX1-RV64-NEXT: vsrl.vi v26, v26, 2
; LMULMAX1-RV64-NEXT: vor.vv v26, v26, v27
; LMULMAX1-RV64-NEXT: vand.vx v27, v26, a1
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: vand.vx v26, v26, a2
; LMULMAX1-RV64-NEXT: vsrl.vi v26, v26, 1
; LMULMAX1-RV64-NEXT: vor.vv v26, v26, v27
@ -1008,7 +1008,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX2-RV32-NEXT: lui a1, 349525
; LMULMAX2-RV32-NEXT: addi a1, a1, 1365
; LMULMAX2-RV32-NEXT: vand.vx v28, v26, a1
; LMULMAX2-RV32-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV32-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV32-NEXT: lui a1, 699051
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vand.vx v26, v26, a1
@ -1058,7 +1058,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX2-RV64-NEXT: lui a1, 349525
; LMULMAX2-RV64-NEXT: addiw a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v28, v26, a1
; LMULMAX2-RV64-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV64-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV64-NEXT: lui a1, 171
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
@ -1108,7 +1108,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX1-RV32-NEXT: lui a2, 349525
; LMULMAX1-RV32-NEXT: addi a2, a2, 1365
; LMULMAX1-RV32-NEXT: vand.vx v27, v25, a2
; LMULMAX1-RV32-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV32-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV32-NEXT: lui a4, 699051
; LMULMAX1-RV32-NEXT: addi a4, a4, -1366
; LMULMAX1-RV32-NEXT: vand.vx v25, v25, a4
@ -1134,7 +1134,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX1-RV32-NEXT: vsrl.vi v26, v26, 2
; LMULMAX1-RV32-NEXT: vor.vv v26, v26, v27
; LMULMAX1-RV32-NEXT: vand.vx v27, v26, a2
; LMULMAX1-RV32-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV32-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV32-NEXT: vand.vx v26, v26, a4
; LMULMAX1-RV32-NEXT: vsrl.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vor.vv v26, v26, v27
@ -1185,7 +1185,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX1-RV64-NEXT: lui a2, 349525
; LMULMAX1-RV64-NEXT: addiw a2, a2, 1365
; LMULMAX1-RV64-NEXT: vand.vx v27, v25, a2
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: lui a4, 171
; LMULMAX1-RV64-NEXT: addiw a4, a4, -1365
; LMULMAX1-RV64-NEXT: slli a4, a4, 12
@ -1213,7 +1213,7 @@ define void @bitreverse_v8i32(<8 x i32>* %x, <8 x i32>* %y) {
; LMULMAX1-RV64-NEXT: vsrl.vi v26, v26, 2
; LMULMAX1-RV64-NEXT: vor.vv v26, v26, v27
; LMULMAX1-RV64-NEXT: vand.vx v27, v26, a2
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: vand.vx v26, v26, a4
; LMULMAX1-RV64-NEXT: vsrl.vi v26, v26, 1
; LMULMAX1-RV64-NEXT: vor.vv v26, v26, v27
@ -1316,7 +1316,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX2-RV32-NEXT: vmv.v.x v28, a1
; LMULMAX2-RV32-NEXT: vsetivli zero, 4, e64, m2, ta, mu
; LMULMAX2-RV32-NEXT: vand.vv v28, v26, v28
; LMULMAX2-RV32-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV32-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV32-NEXT: lui a1, 699051
; LMULMAX2-RV32-NEXT: addi a1, a1, -1366
; LMULMAX2-RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
@ -1414,7 +1414,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
; LMULMAX2-RV64-NEXT: addi a1, a1, 1365
; LMULMAX2-RV64-NEXT: vand.vx v28, v26, a1
; LMULMAX2-RV64-NEXT: vsll.vi v28, v28, 1
; LMULMAX2-RV64-NEXT: vadd.vv v28, v28, v28
; LMULMAX2-RV64-NEXT: lui a1, 1026731
; LMULMAX2-RV64-NEXT: addiw a1, a1, -1365
; LMULMAX2-RV64-NEXT: slli a1, a1, 12
@ -1518,7 +1518,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX1-RV32-NEXT: vmv.v.x v10, a5
; LMULMAX1-RV32-NEXT: vsetivli zero, 2, e64, m1, ta, mu
; LMULMAX1-RV32-NEXT: vand.vv v12, v31, v10
; LMULMAX1-RV32-NEXT: vsll.vi v12, v12, 1
; LMULMAX1-RV32-NEXT: vadd.vv v12, v12, v12
; LMULMAX1-RV32-NEXT: lui a5, 699051
; LMULMAX1-RV32-NEXT: addi a5, a5, -1366
; LMULMAX1-RV32-NEXT: vsetivli zero, 4, e32, m1, ta, mu
@ -1559,7 +1559,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX1-RV32-NEXT: vsrl.vi v25, v25, 2
; LMULMAX1-RV32-NEXT: vor.vv v25, v25, v26
; LMULMAX1-RV32-NEXT: vand.vv v26, v25, v10
; LMULMAX1-RV32-NEXT: vsll.vi v26, v26, 1
; LMULMAX1-RV32-NEXT: vadd.vv v26, v26, v26
; LMULMAX1-RV32-NEXT: vand.vv v25, v25, v13
; LMULMAX1-RV32-NEXT: vsrl.vi v25, v25, 1
; LMULMAX1-RV32-NEXT: vor.vv v25, v25, v26
@ -1659,7 +1659,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX1-RV64-NEXT: slli a1, a1, 12
; LMULMAX1-RV64-NEXT: addi a1, a1, 1365
; LMULMAX1-RV64-NEXT: vand.vx v27, v26, a1
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: lui s0, 1026731
; LMULMAX1-RV64-NEXT: addiw s0, s0, -1365
; LMULMAX1-RV64-NEXT: slli s0, s0, 12
@ -1703,7 +1703,7 @@ define void @bitreverse_v4i64(<4 x i64>* %x, <4 x i64>* %y) {
; LMULMAX1-RV64-NEXT: vsrl.vi v25, v25, 2
; LMULMAX1-RV64-NEXT: vor.vv v25, v25, v27
; LMULMAX1-RV64-NEXT: vand.vx v27, v25, a1
; LMULMAX1-RV64-NEXT: vsll.vi v27, v27, 1
; LMULMAX1-RV64-NEXT: vadd.vv v27, v27, v27
; LMULMAX1-RV64-NEXT: vand.vx v25, v25, s0
; LMULMAX1-RV64-NEXT: vsrl.vi v25, v25, 1
; LMULMAX1-RV64-NEXT: vor.vv v25, v25, v27

10
test/CodeGen/RISCV/rvv/fixed-vectors-int-buildvec.ll
View File

@ -68,7 +68,7 @@ define void @buildvec_vid_step2_add0_v4i8(<4 x i8>* %z0, <4 x i8>* %z1, <4 x i8>
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, mu
; CHECK-NEXT: vid.v v25
; CHECK-NEXT: vsll.vi v25, v25, 1
; CHECK-NEXT: vadd.vv v25, v25, v25
; CHECK-NEXT: vse8.v v25, (a0)
; CHECK-NEXT: vse8.v v25, (a1)
; CHECK-NEXT: vse8.v v25, (a2)
@ -86,7 +86,7 @@ define void @buildvec_vid_step2_add1_v4i8(<4 x i8>* %z0, <4 x i8>* %z1, <4 x i8>
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, mu
; CHECK-NEXT: vid.v v25
; CHECK-NEXT: vsll.vi v25, v25, 1
; CHECK-NEXT: vadd.vv v25, v25, v25
; CHECK-NEXT: vadd.vi v25, v25, 1
; CHECK-NEXT: vse8.v v25, (a0)
; CHECK-NEXT: vse8.v v25, (a1)
@ -127,7 +127,7 @@ define void @buildvec_vid_stepn2_add0_v4i8(<4 x i8>* %z0, <4 x i8>* %z1, <4 x i8
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, mu
; CHECK-NEXT: vid.v v25
; CHECK-NEXT: vsll.vi v25, v25, 1
; CHECK-NEXT: vadd.vv v25, v25, v25
; CHECK-NEXT: vrsub.vi v25, v25, 0
; CHECK-NEXT: vse8.v v25, (a0)
; CHECK-NEXT: vse8.v v25, (a1)
@ -146,7 +146,7 @@ define void @buildvec_vid_stepn2_add3_v4i8(<4 x i8>* %z0, <4 x i8>* %z1, <4 x i8
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, mu
; CHECK-NEXT: vid.v v25
; CHECK-NEXT: vsll.vi v25, v25, 1
; CHECK-NEXT: vadd.vv v25, v25, v25
; CHECK-NEXT: vrsub.vi v25, v25, 3
; CHECK-NEXT: vse8.v v25, (a0)
; CHECK-NEXT: ret
@ -232,7 +232,7 @@ define <4 x i64> @buildvec_vid_step2_add0_v4i64() {
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 2, e64, m1, ta, mu
; RV64-NEXT: vid.v v25
; RV64-NEXT: vsll.vi v8, v25, 1
; RV64-NEXT: vadd.vv v8, v25, v25
; RV64-NEXT: vadd.vi v9, v8, 4
; RV64-NEXT: ret
ret <4 x i64> <i64 0, i64 2, i64 4, i64 6>

32
test/CodeGen/RISCV/rvv/fixed-vectors-masked-gather.ll
View File

@ -470,7 +470,7 @@ define <8 x i16> @mgather_baseidx_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8 x i1
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -480,7 +480,7 @@ define <8 x i16> @mgather_baseidx_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8 x i1
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -495,7 +495,7 @@ define <8 x i16> @mgather_baseidx_sext_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -505,7 +505,7 @@ define <8 x i16> @mgather_baseidx_sext_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -521,7 +521,7 @@ define <8 x i16> @mgather_baseidx_zext_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vzext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -531,7 +531,7 @@ define <8 x i16> @mgather_baseidx_zext_v8i8_v8i16(i16* %base, <8 x i8> %idxs, <8
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vzext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -547,7 +547,7 @@ define <8 x i16> @mgather_baseidx_v8i16(i16* %base, <8 x i16> %idxs, <8 x i1> %m
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf2 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -557,7 +557,7 @@ define <8 x i16> @mgather_baseidx_v8i16(i16* %base, <8 x i16> %idxs, <8 x i1> %m
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf4 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -1385,7 +1385,7 @@ define <8 x half> @mgather_baseidx_v8i8_v8f16(half* %base, <8 x i8> %idxs, <8 x
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -1395,7 +1395,7 @@ define <8 x half> @mgather_baseidx_v8i8_v8f16(half* %base, <8 x i8> %idxs, <8 x
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -1410,7 +1410,7 @@ define <8 x half> @mgather_baseidx_sext_v8i8_v8f16(half* %base, <8 x i8> %idxs,
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -1420,7 +1420,7 @@ define <8 x half> @mgather_baseidx_sext_v8i8_v8f16(half* %base, <8 x i8> %idxs,
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -1436,7 +1436,7 @@ define <8 x half> @mgather_baseidx_zext_v8i8_v8f16(half* %base, <8 x i8> %idxs,
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vzext.vf4 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -1446,7 +1446,7 @@ define <8 x half> @mgather_baseidx_zext_v8i8_v8f16(half* %base, <8 x i8> %idxs,
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vzext.vf8 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9
@ -1462,7 +1462,7 @@ define <8 x half> @mgather_baseidx_v8f16(half* %base, <8 x i16> %idxs, <8 x i1>
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf2 v26, v8
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV32-NEXT: vluxei32.v v9, (a0), v26, v0.t
; RV32-NEXT: vmv1r.v v8, v9
@ -1472,7 +1472,7 @@ define <8 x half> @mgather_baseidx_v8f16(half* %base, <8 x i16> %idxs, <8 x i1>
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf4 v28, v8
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, tu, mu
; RV64-NEXT: vluxei64.v v9, (a0), v28, v0.t
; RV64-NEXT: vmv1r.v v8, v9

32
test/CodeGen/RISCV/rvv/fixed-vectors-masked-scatter.ll
View File

@ -345,7 +345,7 @@ define void @mscatter_baseidx_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i8> %i
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -354,7 +354,7 @@ define void @mscatter_baseidx_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i8> %i
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -368,7 +368,7 @@ define void @mscatter_baseidx_sext_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -377,7 +377,7 @@ define void @mscatter_baseidx_sext_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -392,7 +392,7 @@ define void @mscatter_baseidx_zext_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vzext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -401,7 +401,7 @@ define void @mscatter_baseidx_zext_v8i8_v8i16(<8 x i16> %val, i16* %base, <8 x i
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vzext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -416,7 +416,7 @@ define void @mscatter_baseidx_v8i16(<8 x i16> %val, i16* %base, <8 x i16> %idxs,
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf2 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -425,7 +425,7 @@ define void @mscatter_baseidx_v8i16(<8 x i16> %val, i16* %base, <8 x i16> %idxs,
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf4 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -1137,7 +1137,7 @@ define void @mscatter_baseidx_v8i8_v8f16(<8 x half> %val, half* %base, <8 x i8>
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -1146,7 +1146,7 @@ define void @mscatter_baseidx_v8i8_v8f16(<8 x half> %val, half* %base, <8 x i8>
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -1160,7 +1160,7 @@ define void @mscatter_baseidx_sext_v8i8_v8f16(<8 x half> %val, half* %base, <8 x
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -1169,7 +1169,7 @@ define void @mscatter_baseidx_sext_v8i8_v8f16(<8 x half> %val, half* %base, <8 x
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -1184,7 +1184,7 @@ define void @mscatter_baseidx_zext_v8i8_v8f16(<8 x half> %val, half* %base, <8 x
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vzext.vf4 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -1193,7 +1193,7 @@ define void @mscatter_baseidx_zext_v8i8_v8f16(<8 x half> %val, half* %base, <8 x
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vzext.vf8 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret
@ -1208,7 +1208,7 @@ define void @mscatter_baseidx_v8f16(<8 x half> %val, half* %base, <8 x i16> %idx
; RV32: # %bb.0:
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, mu
; RV32-NEXT: vsext.vf2 v26, v9
; RV32-NEXT: vsll.vi v26, v26, 1
; RV32-NEXT: vadd.vv v26, v26, v26
; RV32-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v26, v0.t
; RV32-NEXT: ret
@ -1217,7 +1217,7 @@ define void @mscatter_baseidx_v8f16(<8 x half> %val, half* %base, <8 x i16> %idx
; RV64: # %bb.0:
; RV64-NEXT: vsetivli zero, 8, e64, m4, ta, mu
; RV64-NEXT: vsext.vf4 v28, v9
; RV64-NEXT: vsll.vi v28, v28, 1
; RV64-NEXT: vadd.vv v28, v28, v28
; RV64-NEXT: vsetvli zero, zero, e16, m1, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v28, v0.t
; RV64-NEXT: ret

32
test/CodeGen/RISCV/rvv/mgather-sdnode.ll
View File

@ -474,7 +474,7 @@ define <vscale x 8 x i16> @mgather_baseidx_nxv8i8_nxv8i16(i16* %base, <vscale x
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -484,7 +484,7 @@ define <vscale x 8 x i16> @mgather_baseidx_nxv8i8_nxv8i16(i16* %base, <vscale x
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -499,7 +499,7 @@ define <vscale x 8 x i16> @mgather_baseidx_sext_nxv8i8_nxv8i16(i16* %base, <vsca
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -509,7 +509,7 @@ define <vscale x 8 x i16> @mgather_baseidx_sext_nxv8i8_nxv8i16(i16* %base, <vsca
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -525,7 +525,7 @@ define <vscale x 8 x i16> @mgather_baseidx_zext_nxv8i8_nxv8i16(i16* %base, <vsca
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vzext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -535,7 +535,7 @@ define <vscale x 8 x i16> @mgather_baseidx_zext_nxv8i8_nxv8i16(i16* %base, <vsca
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vzext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -551,7 +551,7 @@ define <vscale x 8 x i16> @mgather_baseidx_nxv8i16(i16* %base, <vscale x 8 x i16
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf2 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -561,7 +561,7 @@ define <vscale x 8 x i16> @mgather_baseidx_nxv8i16(i16* %base, <vscale x 8 x i16
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf4 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -1457,7 +1457,7 @@ define <vscale x 8 x half> @mgather_baseidx_nxv8i8_nxv8f16(half* %base, <vscale
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -1467,7 +1467,7 @@ define <vscale x 8 x half> @mgather_baseidx_nxv8i8_nxv8f16(half* %base, <vscale
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -1482,7 +1482,7 @@ define <vscale x 8 x half> @mgather_baseidx_sext_nxv8i8_nxv8f16(half* %base, <vs
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -1492,7 +1492,7 @@ define <vscale x 8 x half> @mgather_baseidx_sext_nxv8i8_nxv8f16(half* %base, <vs
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -1508,7 +1508,7 @@ define <vscale x 8 x half> @mgather_baseidx_zext_nxv8i8_nxv8f16(half* %base, <vs
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vzext.vf4 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -1518,7 +1518,7 @@ define <vscale x 8 x half> @mgather_baseidx_zext_nxv8i8_nxv8f16(half* %base, <vs
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vzext.vf8 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10
@ -1534,7 +1534,7 @@ define <vscale x 8 x half> @mgather_baseidx_nxv8f16(half* %base, <vscale x 8 x i
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf2 v28, v8
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV32-NEXT: vluxei32.v v10, (a0), v28, v0.t
; RV32-NEXT: vmv2r.v v8, v10
@ -1544,7 +1544,7 @@ define <vscale x 8 x half> @mgather_baseidx_nxv8f16(half* %base, <vscale x 8 x i
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf4 v16, v8
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, tu, mu
; RV64-NEXT: vluxei64.v v10, (a0), v16, v0.t
; RV64-NEXT: vmv2r.v v8, v10

32
test/CodeGen/RISCV/rvv/mscatter-sdnode.ll
View File

@ -345,7 +345,7 @@ define void @mscatter_baseidx_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16* %base
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -354,7 +354,7 @@ define void @mscatter_baseidx_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16* %base
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -368,7 +368,7 @@ define void @mscatter_baseidx_sext_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16*
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -377,7 +377,7 @@ define void @mscatter_baseidx_sext_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16*
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -392,7 +392,7 @@ define void @mscatter_baseidx_zext_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16*
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vzext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -401,7 +401,7 @@ define void @mscatter_baseidx_zext_nxv8i8_nxv8i16(<vscale x 8 x i16> %val, i16*
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vzext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -416,7 +416,7 @@ define void @mscatter_baseidx_nxv8i16(<vscale x 8 x i16> %val, i16* %base, <vsca
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf2 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -425,7 +425,7 @@ define void @mscatter_baseidx_nxv8i16(<vscale x 8 x i16> %val, i16* %base, <vsca
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf4 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -1137,7 +1137,7 @@ define void @mscatter_baseidx_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half* %ba
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -1146,7 +1146,7 @@ define void @mscatter_baseidx_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half* %ba
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -1160,7 +1160,7 @@ define void @mscatter_baseidx_sext_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -1169,7 +1169,7 @@ define void @mscatter_baseidx_sext_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -1184,7 +1184,7 @@ define void @mscatter_baseidx_zext_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vzext.vf4 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -1193,7 +1193,7 @@ define void @mscatter_baseidx_zext_nxv8i8_nxv8f16(<vscale x 8 x half> %val, half
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vzext.vf8 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret
@ -1208,7 +1208,7 @@ define void @mscatter_baseidx_nxv8f16(<vscale x 8 x half> %val, half* %base, <vs
; RV32: # %bb.0:
; RV32-NEXT: vsetvli a1, zero, e32, m4, ta, mu
; RV32-NEXT: vsext.vf2 v28, v10
; RV32-NEXT: vsll.vi v28, v28, 1
; RV32-NEXT: vadd.vv v28, v28, v28
; RV32-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV32-NEXT: vsoxei32.v v8, (a0), v28, v0.t
; RV32-NEXT: ret
@ -1217,7 +1217,7 @@ define void @mscatter_baseidx_nxv8f16(<vscale x 8 x half> %val, half* %base, <vs
; RV64: # %bb.0:
; RV64-NEXT: vsetvli a1, zero, e64, m8, ta, mu
; RV64-NEXT: vsext.vf4 v16, v10
; RV64-NEXT: vsll.vi v16, v16, 1
; RV64-NEXT: vadd.vv v16, v16, v16
; RV64-NEXT: vsetvli zero, zero, e16, m2, ta, mu
; RV64-NEXT: vsoxei64.v v8, (a0), v16, v0.t
; RV64-NEXT: ret

47
test/CodeGen/RISCV/rvv/vshl-sdnode-rv32.ll
View File

@ -470,6 +470,18 @@ define <vscale x 1 x i64> @vshl_vx_nxv1i64_1(<vscale x 1 x i64> %va) {
ret <vscale x 1 x i64> %vc
}
define <vscale x 1 x i64> @vshl_vx_nxv1i64_2(<vscale x 1 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv1i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 1 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 1 x i64> %head, <vscale x 1 x i64> undef, <vscale x 1 x i32> zeroinitializer
%vc = shl <vscale x 1 x i64> %va, %splat
ret <vscale x 1 x i64> %vc
}
define <vscale x 2 x i64> @vshl_vx_nxv2i64(<vscale x 2 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv2i64:
; CHECK: # %bb.0:
@ -507,6 +519,18 @@ define <vscale x 2 x i64> @vshl_vx_nxv2i64_1(<vscale x 2 x i64> %va) {
ret <vscale x 2 x i64> %vc
}
define <vscale x 2 x i64> @vshl_vx_nxv2i64_2(<vscale x 2 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv2i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 2 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 2 x i64> %head, <vscale x 2 x i64> undef, <vscale x 2 x i32> zeroinitializer
%vc = shl <vscale x 2 x i64> %va, %splat
ret <vscale x 2 x i64> %vc
}
define <vscale x 4 x i64> @vshl_vx_nxv4i64(<vscale x 4 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv4i64:
; CHECK: # %bb.0:
@ -544,6 +568,18 @@ define <vscale x 4 x i64> @vshl_vx_nxv4i64_1(<vscale x 4 x i64> %va) {
ret <vscale x 4 x i64> %vc
}
define <vscale x 4 x i64> @vshl_vx_nxv4i64_2(<vscale x 4 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv4i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 4 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 4 x i64> %head, <vscale x 4 x i64> undef, <vscale x 4 x i32> zeroinitializer
%vc = shl <vscale x 4 x i64> %va, %splat
ret <vscale x 4 x i64> %vc
}
define <vscale x 8 x i64> @vshl_vx_nxv8i64(<vscale x 8 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv8i64:
; CHECK: # %bb.0:
@ -581,3 +617,14 @@ define <vscale x 8 x i64> @vshl_vx_nxv8i64_1(<vscale x 8 x i64> %va) {
ret <vscale x 8 x i64> %vc
}
define <vscale x 8 x i64> @vshl_vx_nxv8i64_2(<vscale x 8 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv8i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 8 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 8 x i64> %head, <vscale x 8 x i64> undef, <vscale x 8 x i32> zeroinitializer
%vc = shl <vscale x 8 x i64> %va, %splat
ret <vscale x 8 x i64> %vc
}

47
test/CodeGen/RISCV/rvv/vshl-sdnode-rv64.ll
View File

@ -470,6 +470,18 @@ define <vscale x 1 x i64> @vshl_vx_nxv1i64_1(<vscale x 1 x i64> %va) {
ret <vscale x 1 x i64> %vc
}
define <vscale x 1 x i64> @vshl_vx_nxv1i64_2(<vscale x 1 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv1i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 1 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 1 x i64> %head, <vscale x 1 x i64> undef, <vscale x 1 x i32> zeroinitializer
%vc = shl <vscale x 1 x i64> %va, %splat
ret <vscale x 1 x i64> %vc
}
define <vscale x 2 x i64> @vshl_vx_nxv2i64(<vscale x 2 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv2i64:
; CHECK: # %bb.0:
@ -507,6 +519,18 @@ define <vscale x 2 x i64> @vshl_vx_nxv2i64_1(<vscale x 2 x i64> %va) {
ret <vscale x 2 x i64> %vc
}
define <vscale x 2 x i64> @vshl_vx_nxv2i64_2(<vscale x 2 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv2i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 2 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 2 x i64> %head, <vscale x 2 x i64> undef, <vscale x 2 x i32> zeroinitializer
%vc = shl <vscale x 2 x i64> %va, %splat
ret <vscale x 2 x i64> %vc
}
define <vscale x 4 x i64> @vshl_vx_nxv4i64(<vscale x 4 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv4i64:
; CHECK: # %bb.0:
@ -544,6 +568,18 @@ define <vscale x 4 x i64> @vshl_vx_nxv4i64_1(<vscale x 4 x i64> %va) {
ret <vscale x 4 x i64> %vc
}
define <vscale x 4 x i64> @vshl_vx_nxv4i64_2(<vscale x 4 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv4i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 4 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 4 x i64> %head, <vscale x 4 x i64> undef, <vscale x 4 x i32> zeroinitializer
%vc = shl <vscale x 4 x i64> %va, %splat
ret <vscale x 4 x i64> %vc
}
define <vscale x 8 x i64> @vshl_vx_nxv8i64(<vscale x 8 x i64> %va, i64 %b) {
; CHECK-LABEL: vshl_vx_nxv8i64:
; CHECK: # %bb.0:
@ -581,3 +617,14 @@ define <vscale x 8 x i64> @vshl_vx_nxv8i64_1(<vscale x 8 x i64> %va) {
ret <vscale x 8 x i64> %vc
}
define <vscale x 8 x i64> @vshl_vx_nxv8i64_2(<vscale x 8 x i64> %va) {
; CHECK-LABEL: vshl_vx_nxv8i64_2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, mu
; CHECK-NEXT: vadd.vv v8, v8, v8
; CHECK-NEXT: ret
%head = insertelement <vscale x 8 x i64> undef, i64 1, i32 0
%splat = shufflevector <vscale x 8 x i64> %head, <vscale x 8 x i64> undef, <vscale x 8 x i32> zeroinitializer
%vc = shl <vscale x 8 x i64> %va, %splat
ret <vscale x 8 x i64> %vc
}

4
test/CodeGen/RISCV/urem-seteq-illegal-types.ll
View File

@ -575,7 +575,7 @@ define void @test_urem_vec(<3 x i11>* %X) nounwind {
; RV32MV-NEXT: vid.v v27
; RV32MV-NEXT: vsub.vv v25, v25, v27
; RV32MV-NEXT: vmul.vv v25, v25, v26
; RV32MV-NEXT: vsll.vi v26, v25, 1
; RV32MV-NEXT: vadd.vv v26, v25, v25
; RV32MV-NEXT: addi a1, zero, 9
; RV32MV-NEXT: vmv.v.i v27, 10
; RV32MV-NEXT: vsetvli zero, zero, e16, mf2, tu, mu
@ -640,7 +640,7 @@ define void @test_urem_vec(<3 x i11>* %X) nounwind {
; RV64MV-NEXT: vid.v v27
; RV64MV-NEXT: vsub.vv v25, v25, v27
; RV64MV-NEXT: vmul.vv v25, v25, v26
; RV64MV-NEXT: vsll.vi v26, v25, 1
; RV64MV-NEXT: vadd.vv v26, v25, v25
; RV64MV-NEXT: addi a1, zero, 9
; RV64MV-NEXT: vmv.v.i v27, 10
; RV64MV-NEXT: vsetvli zero, zero, e16, mf2, tu, mu