Skip to content

[PowerPC] hoist xxspltib out of loop body #127121

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Mar 3, 2025
Merged

[PowerPC] hoist xxspltib out of loop body #127121

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 0 additions & 2 deletions llvm/lib/Target/PowerPC/PPCInstrVSX.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1659,8 +1659,6 @@ let Predicates = [HasVSX, HasP9Vector] in {
RegConstraint<"$XTi = $XT">, NoEncode<"$XTi">;

// Vector Splat Immediate Byte
// FIXME: Setting the hasSideEffects flag here to match current behaviour.
let hasSideEffects = 1 in
def XXSPLTIB : X_RD6_IMM8<60, 360, (outs vsrc:$XT), (ins u8imm:$IMM8),
"xxspltib $XT, $IMM8", IIC_VecPerm, []>;

Expand Down
62 changes: 31 additions & 31 deletions llvm/test/CodeGen/PowerPC/licm-xxsplti.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,6 +82,7 @@ for.body:
; AIX32-NEXT: li 6, 0
; AIX32-NEXT: beq 0, L..BB0_4
; AIX32-NEXT: # %bb.2: # %for.body.preheader.new
; AIX32-NEXT: xxspltib 0, 6
; AIX32-NEXT: addi 12, 4, -8
; AIX32-NEXT: addi 9, 3, -8
; AIX32-NEXT: rlwinm 7, 5, 0, 1, 30
Expand All @@ -91,20 +92,19 @@ for.body:
; AIX32-NEXT: .align 4
; AIX32-NEXT: L..BB0_3: # %for.body
; AIX32-NEXT: # =>This Inner Loop Header: Depth=1
; AIX32-NEXT: lxvwsx 0, 12, 10
; AIX32-NEXT: xxspltib 1, 6
; AIX32-NEXT: lxvwsx 2, 12, 11
; AIX32-NEXT: lxvwsx 1, 12, 10
; AIX32-NEXT: addic 6, 6, 2
; AIX32-NEXT: addi 12, 12, 8
; AIX32-NEXT: addze 8, 8
; AIX32-NEXT: xor 0, 6, 7
; AIX32-NEXT: or. 0, 0, 8
; AIX32-NEXT: xxland 0, 0, 1
; AIX32-NEXT: xxland 1, 2, 1
; AIX32-NEXT: xscvspdpn 0, 0
; AIX32-NEXT: stfsu 0, 8(9)
; AIX32-NEXT: xscvspdpn 0, 1
; AIX32-NEXT: stfs 0, 4(9)
; AIX32-NEXT: xxland 1, 1, 0
; AIX32-NEXT: xscvspdpn 1, 1
; AIX32-NEXT: stfsu 1, 8(9)
; AIX32-NEXT: lxvwsx 1, 12, 11
; AIX32-NEXT: addi 12, 12, 8
; AIX32-NEXT: xxland 1, 1, 0
; AIX32-NEXT: xscvspdpn 1, 1
; AIX32-NEXT: stfs 1, 4(9)
; AIX32-NEXT: bne 0, L..BB0_3

; AIX64: ._Z3fooPfS_Pi:
Expand All @@ -118,10 +118,11 @@ for.body:
; AIX64-NEXT: beq 0, L..BB0_4
; AIX64-NEXT: # %bb.2: # %for.body.preheader.new
; AIX64-NEXT: rlwinm 6, 5, 0, 1, 30
; AIX64-NEXT: addi 10, 4, -8
; AIX64-NEXT: xxspltib 0, 6
; AIX64-NEXT: addi 9, 4, -8
; AIX64-NEXT: addi 7, 3, -8
; AIX64-NEXT: li 8, 8
; AIX64-NEXT: li 9, 12
; AIX64-NEXT: li 10, 12
; AIX64-NEXT: li 11, 4
; AIX64-NEXT: addi 6, 6, -2
; AIX64-NEXT: rldicl 6, 6, 63, 1
Expand All @@ -131,17 +132,16 @@ for.body:
; AIX64-NEXT: .align 4
; AIX64-NEXT: L..BB0_3: # %for.body
; AIX64-NEXT: # =>This Inner Loop Header: Depth=1
; AIX64-NEXT: lxvwsx 0, 10, 8
; AIX64-NEXT: xxspltib 1, 6
; AIX64-NEXT: lxvwsx 1, 9, 8
; AIX64-NEXT: addi 6, 6, 2
; AIX64-NEXT: xxland 0, 0, 1
; AIX64-NEXT: xscvspdpn 0, 0
; AIX64-NEXT: stfsu 0, 8(7)
; AIX64-NEXT: lxvwsx 0, 10, 9
; AIX64-NEXT: addi 10, 10, 8
; AIX64-NEXT: xxland 0, 0, 1
; AIX64-NEXT: xxsldwi 0, 0, 0, 3
; AIX64-NEXT: stfiwx 0, 7, 11
; AIX64-NEXT: xxland 1, 1, 0
; AIX64-NEXT: xscvspdpn 1, 1
; AIX64-NEXT: stfsu 1, 8(7)
; AIX64-NEXT: lxvwsx 1, 9, 10
; AIX64-NEXT: addi 9, 9, 8
; AIX64-NEXT: xxland 1, 1, 0
; AIX64-NEXT: xxsldwi 1, 1, 1, 3
; AIX64-NEXT: stfiwx 1, 7, 11
; AIX64-NEXT: bdnz L..BB0_3

; LINUX64LE: _Z3fooPfS_Pi: # @_Z3fooPfS_Pi
Expand All @@ -157,6 +157,7 @@ for.body:
; LINUX64LE-NEXT: beq 0, .LBB0_4
; LINUX64LE-NEXT: # %bb.2: # %for.body.preheader.new
; LINUX64LE-NEXT: rlwinm 6, 5, 0, 1, 30
; LINUX64LE-NEXT: xxspltib 0, 6
; LINUX64LE-NEXT: addi 8, 4, -8
; LINUX64LE-NEXT: addi 7, 3, -8
; LINUX64LE-NEXT: li 9, 8
Expand All @@ -170,15 +171,14 @@ for.body:
; LINUX64LE-NEXT: .p2align 4
; LINUX64LE-NEXT: .LBB0_3: # %for.body
; LINUX64LE-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX64LE-NEXT: lxvwsx 0, 8, 9
; LINUX64LE-NEXT: xxspltib 1, 6
; LINUX64LE-NEXT: lxvwsx 1, 8, 9
; LINUX64LE-NEXT: addi 6, 6, 2
; LINUX64LE-NEXT: xxland 0, 0, 1
; LINUX64LE-NEXT: xxsldwi 0, 0, 0, 3
; LINUX64LE-NEXT: xscvspdpn 0, 0
; LINUX64LE-NEXT: stfsu 0, 8(7)
; LINUX64LE-NEXT: lxvwsx 0, 8, 10
; LINUX64LE-NEXT: xxland 1, 1, 0
; LINUX64LE-NEXT: xxsldwi 1, 1, 1, 3
; LINUX64LE-NEXT: xscvspdpn 1, 1
; LINUX64LE-NEXT: stfsu 1, 8(7)
; LINUX64LE-NEXT: lxvwsx 1, 8, 10
; LINUX64LE-NEXT: addi 8, 8, 8
; LINUX64LE-NEXT: xxland 0, 0, 1
; LINUX64LE-NEXT: stxvrwx 0, 7, 11
; LINUX64LE-NEXT: xxland 1, 1, 0
; LINUX64LE-NEXT: stxvrwx 1, 7, 11
; LINUX64LE-NEXT: bdnz .LBB0_3
32 changes: 16 additions & 16 deletions llvm/test/CodeGen/PowerPC/memset-tail.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -380,17 +380,17 @@ define dso_local void @memsetTailV1B2(ptr nocapture noundef writeonly %p) local_
;
; P9-BE-LABEL: memsetTailV1B2:
; P9-BE: # %bb.0: # %entry
; P9-BE-NEXT: xxspltib 0, 15
; P9-BE-NEXT: li 4, 3855
; P9-BE-NEXT: sth 4, 16(3)
; P9-BE-NEXT: xxspltib 0, 15
; P9-BE-NEXT: stxv 0, 0(3)
; P9-BE-NEXT: blr
;
; P10-BE-LABEL: memsetTailV1B2:
; P10-BE: # %bb.0: # %entry
; P10-BE-NEXT: xxspltib 0, 15
; P10-BE-NEXT: li 4, 3855
; P10-BE-NEXT: sth 4, 16(3)
; P10-BE-NEXT: xxspltib 0, 15
; P10-BE-NEXT: stxv 0, 0(3)
; P10-BE-NEXT: blr
;
Expand All @@ -404,17 +404,17 @@ define dso_local void @memsetTailV1B2(ptr nocapture noundef writeonly %p) local_
;
; P9-LE-LABEL: memsetTailV1B2:
; P9-LE: # %bb.0: # %entry
; P9-LE-NEXT: xxspltib 0, 15
; P9-LE-NEXT: li 4, 3855
; P9-LE-NEXT: sth 4, 16(3)
; P9-LE-NEXT: xxspltib 0, 15
; P9-LE-NEXT: stxv 0, 0(3)
; P9-LE-NEXT: blr
;
; P10-LE-LABEL: memsetTailV1B2:
; P10-LE: # %bb.0: # %entry
; P10-LE-NEXT: xxspltib 0, 15
; P10-LE-NEXT: li 4, 3855
; P10-LE-NEXT: sth 4, 16(3)
; P10-LE-NEXT: xxspltib 0, 15
; P10-LE-NEXT: stxv 0, 0(3)
; P10-LE-NEXT: blr
entry:
Expand All @@ -433,17 +433,17 @@ define dso_local void @memsetTailV1B1(ptr nocapture noundef writeonly %p) local_
;
; P9-BE-LABEL: memsetTailV1B1:
; P9-BE: # %bb.0: # %entry
; P9-BE-NEXT: xxspltib 0, 15
; P9-BE-NEXT: li 4, 15
; P9-BE-NEXT: stb 4, 16(3)
; P9-BE-NEXT: xxspltib 0, 15
; P9-BE-NEXT: stxv 0, 0(3)
; P9-BE-NEXT: blr
;
; P10-BE-LABEL: memsetTailV1B1:
; P10-BE: # %bb.0: # %entry
; P10-BE-NEXT: xxspltib 0, 15
; P10-BE-NEXT: li 4, 15
; P10-BE-NEXT: stb 4, 16(3)
; P10-BE-NEXT: xxspltib 0, 15
; P10-BE-NEXT: stxv 0, 0(3)
; P10-BE-NEXT: blr
;
Expand All @@ -457,17 +457,17 @@ define dso_local void @memsetTailV1B1(ptr nocapture noundef writeonly %p) local_
;
; P9-LE-LABEL: memsetTailV1B1:
; P9-LE: # %bb.0: # %entry
; P9-LE-NEXT: xxspltib 0, 15
; P9-LE-NEXT: li 4, 15
; P9-LE-NEXT: stb 4, 16(3)
; P9-LE-NEXT: xxspltib 0, 15
; P9-LE-NEXT: stxv 0, 0(3)
; P9-LE-NEXT: blr
;
; P10-LE-LABEL: memsetTailV1B1:
; P10-LE: # %bb.0: # %entry
; P10-LE-NEXT: xxspltib 0, 15
; P10-LE-NEXT: li 4, 15
; P10-LE-NEXT: stb 4, 16(3)
; P10-LE-NEXT: xxspltib 0, 15
; P10-LE-NEXT: stxv 0, 0(3)
; P10-LE-NEXT: blr
entry:
Expand Down Expand Up @@ -861,17 +861,17 @@ define dso_local void @memset2TailV1B2(ptr nocapture noundef writeonly %p) local
;
; P9-BE-LABEL: memset2TailV1B2:
; P9-BE: # %bb.0: # %entry
; P9-BE-NEXT: xxspltib 0, 165
; P9-BE-NEXT: li 4, -23131
; P9-BE-NEXT: sth 4, 16(3)
; P9-BE-NEXT: xxspltib 0, 165
; P9-BE-NEXT: stxv 0, 0(3)
; P9-BE-NEXT: blr
;
; P10-BE-LABEL: memset2TailV1B2:
; P10-BE: # %bb.0: # %entry
; P10-BE-NEXT: xxspltib 0, 165
; P10-BE-NEXT: li 4, -23131
; P10-BE-NEXT: sth 4, 16(3)
; P10-BE-NEXT: xxspltib 0, 165
; P10-BE-NEXT: stxv 0, 0(3)
; P10-BE-NEXT: blr
;
Expand All @@ -887,17 +887,17 @@ define dso_local void @memset2TailV1B2(ptr nocapture noundef writeonly %p) local
;
; P9-LE-LABEL: memset2TailV1B2:
; P9-LE: # %bb.0: # %entry
; P9-LE-NEXT: xxspltib 0, 165
; P9-LE-NEXT: li 4, -23131
; P9-LE-NEXT: sth 4, 16(3)
; P9-LE-NEXT: xxspltib 0, 165
; P9-LE-NEXT: stxv 0, 0(3)
; P9-LE-NEXT: blr
;
; P10-LE-LABEL: memset2TailV1B2:
; P10-LE: # %bb.0: # %entry
; P10-LE-NEXT: xxspltib 0, 165
; P10-LE-NEXT: li 4, -23131
; P10-LE-NEXT: sth 4, 16(3)
; P10-LE-NEXT: xxspltib 0, 165
; P10-LE-NEXT: stxv 0, 0(3)
; P10-LE-NEXT: blr
entry:
Expand All @@ -917,17 +917,17 @@ define dso_local void @memset2TailV1B1(ptr nocapture noundef writeonly %p) local
;
; P9-BE-LABEL: memset2TailV1B1:
; P9-BE: # %bb.0: # %entry
; P9-BE-NEXT: xxspltib 0, 165
; P9-BE-NEXT: li 4, -91
; P9-BE-NEXT: stb 4, 16(3)
; P9-BE-NEXT: xxspltib 0, 165
; P9-BE-NEXT: stxv 0, 0(3)
; P9-BE-NEXT: blr
;
; P10-BE-LABEL: memset2TailV1B1:
; P10-BE: # %bb.0: # %entry
; P10-BE-NEXT: xxspltib 0, 165
; P10-BE-NEXT: li 4, -91
; P10-BE-NEXT: stb 4, 16(3)
; P10-BE-NEXT: xxspltib 0, 165
; P10-BE-NEXT: stxv 0, 0(3)
; P10-BE-NEXT: blr
;
Expand All @@ -943,17 +943,17 @@ define dso_local void @memset2TailV1B1(ptr nocapture noundef writeonly %p) local
;
; P9-LE-LABEL: memset2TailV1B1:
; P9-LE: # %bb.0: # %entry
; P9-LE-NEXT: xxspltib 0, 165
; P9-LE-NEXT: li 4, -91
; P9-LE-NEXT: stb 4, 16(3)
; P9-LE-NEXT: xxspltib 0, 165
; P9-LE-NEXT: stxv 0, 0(3)
; P9-LE-NEXT: blr
;
; P10-LE-LABEL: memset2TailV1B1:
; P10-LE: # %bb.0: # %entry
; P10-LE-NEXT: xxspltib 0, 165
; P10-LE-NEXT: li 4, -91
; P10-LE-NEXT: stb 4, 16(3)
; P10-LE-NEXT: xxspltib 0, 165
; P10-LE-NEXT: stxv 0, 0(3)
; P10-LE-NEXT: blr
entry:
Expand Down