[llvm] 12e1936 - [PowerPC] Add XXEVAL TD pattern
Ting Wang via llvm-commits
llvm-commits at lists.llvm.org
Thu Aug 11 22:28:04 PDT 2022
Author: Ting Wang
Date: 2022-08-12T01:27:24-04:00
New Revision: 12e1936f64f8739889dafffd17c826526ce4da5e
URL: https://github.com/llvm/llvm-project/commit/12e1936f64f8739889dafffd17c826526ce4da5e
DIFF: https://github.com/llvm/llvm-project/commit/12e1936f64f8739889dafffd17c826526ce4da5e.diff
LOG: [PowerPC] Add XXEVAL TD pattern
Add xxeval TD pattern for P10 on: eqv, nor, or, xor.
Reviewed By: shchenz
Differential Revision: https://reviews.llvm.org/D131654
Added:
llvm/test/CodeGen/PowerPC/xxeval-eqv-nor-or-xor.ll
Modified:
llvm/lib/Target/PowerPC/PPCInstrP10.td
Removed:
################################################################################
diff --git a/llvm/lib/Target/PowerPC/PPCInstrP10.td b/llvm/lib/Target/PowerPC/PPCInstrP10.td
index 6cf3f1d3341ec..c127f54875527 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrP10.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrP10.td
@@ -2096,6 +2096,52 @@ def : Pat<(f64 nzFPImmAsi64:$A),
def : xxevalPattern<(or (vnot v4i32:$vA), (and v4i32:$vB, v4i32:$vC)),
!sub(255, 14)>;
+ // EQV
+ // (eqv A, B, C)
+ def : xxevalPattern<(or (and v4i32:$vA, (and v4i32:$vB, v4i32:$vC)),
+ (vnot (or v4i32:$vA, (or v4i32:$vB, v4i32:$vC)))),
+ 150>;
+ // (eqv A, (and B, C))
+ def : xxevalPattern<(vnot (xor v4i32:$vA, (and v4i32:$vB, v4i32:$vC))), 225>;
+ // (eqv A, (or B, C))
+ def : xxevalPattern<(vnot (xor v4i32:$vA, (or v4i32:$vB, v4i32:$vC))), 135>;
+
+ // NOR
+ // (nor A, B, C)
+ def : xxevalPattern<(vnot (or v4i32:$vA, (or v4i32:$vB, v4i32:$vC))), 128>;
+ // (nor A, (and B, C))
+ def : xxevalPattern<(vnot (or v4i32:$vA, (and v4i32:$vB, v4i32:$vC))), 224>;
+ // (nor A, (eqv B, C))
+ def : xxevalPattern<(and (vnot v4i32:$vA), (xor v4i32:$vB, v4i32:$vC)), 96>;
+ // (nor A, (nand B, C))
+ def : xxevalPattern<(and (vnot v4i32:$vA), (and v4i32:$vB, v4i32:$vC)), 16>;
+ // (nor A, (nor B, C))
+ def : xxevalPattern<(and (vnot v4i32:$vA), (or v4i32:$vB, v4i32:$vC)), 112>;
+ // (nor A, (xor B, C))
+ def : xxevalPattern<(vnot (or v4i32:$vA, (xor v4i32:$vB, v4i32:$vC))), 144>;
+
+ // OR
+ // (or A, B, C)
+ def : xxevalPattern<(or v4i32:$vA, (or v4i32:$vB, v4i32:$vC)), 127>;
+ // (or A, (and B, C))
+ def : xxevalPattern<(or v4i32:$vA, (and v4i32:$vB, v4i32:$vC)), 31>;
+ // (or A, (eqv B, C))
+ def : xxevalPattern<(or v4i32:$vA, (vnot (xor v4i32:$vB, v4i32:$vC))), 159>;
+ // (or A, (nand B, C))
+ def : xxevalPattern<(or v4i32:$vA, (vnot (and v4i32:$vB, v4i32:$vC))), 239>;
+ // (or A, (nor B, C))
+ def : xxevalPattern<(or v4i32:$vA, (vnot (or v4i32:$vB, v4i32:$vC))), 143>;
+ // (or A, (xor B, C))
+ def : xxevalPattern<(or v4i32:$vA, (xor v4i32:$vB, v4i32:$vC)), 111>;
+
+ // XOR
+ // (xor A, B, C)
+ def : xxevalPattern<(xor v4i32:$vA, (xor v4i32:$vB, v4i32:$vC)), 105>;
+ // (xor A, (and B, C))
+ def : xxevalPattern<(xor v4i32:$vA, (and v4i32:$vB, v4i32:$vC)), 30>;
+ // (xor A, (or B, C))
+ def : xxevalPattern<(xor v4i32:$vA, (or v4i32:$vB, v4i32:$vC)), 120>;
+
// Anonymous patterns to select prefixed VSX loads and stores.
// Load / Store f128
def : Pat<(f128 (load PDForm:$src)),
diff --git a/llvm/test/CodeGen/PowerPC/xxeval-eqv-nor-or-xor.ll b/llvm/test/CodeGen/PowerPC/xxeval-eqv-nor-or-xor.ll
new file mode 100644
index 0000000000000..6616a1e6e7e9f
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/xxeval-eqv-nor-or-xor.ll
@@ -0,0 +1,236 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mcpu=pwr10 -mtriple=powerpc64le-unknown-unknown \
+; RUN: -ppc-asm-full-reg-names --ppc-vsr-nums-as-vr < %s | FileCheck %s
+
+; (eqv A, (and B, C)) 225
+define dso_local <2 x i64> @eqvA_andB_C(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: eqvA_andB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 225
+; CHECK-NEXT: blr
+entry:
+ %and = and <2 x i64> %B, %C
+ %xor = xor <2 x i64> %A, %and
+ %neg = xor <2 x i64> %xor, <i64 -1, i64 -1>
+ ret <2 x i64> %neg
+}
+
+; (eqv A, (or B, C)) 135
+define dso_local <4 x i32> @eqvA_orB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: eqvA_orB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 135
+; CHECK-NEXT: blr
+entry:
+ %or = or <4 x i32> %B, %C
+ %xor = xor <4 x i32> %A, %or
+ %neg = xor <4 x i32> %xor, <i32 -1, i32 -1, i32 -1, i32 -1>
+ ret <4 x i32> %neg
+}
+
+; (eqv A, B, C) 150
+define dso_local <8 x i16> @eqvA_B_C(<8 x i16> %A, <8 x i16> %B, <8 x i16> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: eqvA_B_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 150
+; CHECK-NEXT: blr
+entry:
+ %and = and <8 x i16> %B, %C
+ %and1 = and <8 x i16> %A, %and
+ %or = or <8 x i16> %B, %C
+ %or1 = or <8 x i16> %A, %or
+ %neg = xor <8 x i16> %or1, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+ %or2 = or <8 x i16> %and1, %neg
+ ret <8 x i16> %or2
+}
+
+; (nor A, (and B, C)) 224
+define dso_local <16 x i8> @norA_andB_C(<16 x i8> %A, <16 x i8> %B, <16 x i8> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_andB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 224
+; CHECK-NEXT: blr
+entry:
+ %and = and <16 x i8> %B, %C
+ %or = or <16 x i8> %A, %and
+ %neg = xor <16 x i8> %or, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+ ret <16 x i8> %neg
+}
+
+; (nor A, (eqv B, C)) 96
+define dso_local <4 x i32> @norA_eqvB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_eqvB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 96
+; CHECK-NEXT: blr
+entry:
+ %neg = xor <4 x i32> %A, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %xor = xor <4 x i32> %B, %C
+ %and = and <4 x i32> %neg, %xor
+ ret <4 x i32> %and
+}
+
+; (nor A, (nand B, C)) 16
+define dso_local <4 x i32> @norA_nandB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_nandB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 16
+; CHECK-NEXT: blr
+entry:
+ %neg = xor <4 x i32> %A, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %and = and <4 x i32> %B, %C
+ %and1 = and <4 x i32> %neg, %and
+ ret <4 x i32> %and1
+}
+
+; (nor A, (nor B, C)) 112
+define dso_local <4 x i32> @norA_norB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_norB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 112
+; CHECK-NEXT: blr
+entry:
+ %neg = xor <4 x i32> %A, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %or = or <4 x i32> %B, %C
+ %and = and <4 x i32> %neg, %or
+ ret <4 x i32> %and
+}
+
+; (nor A, (xor B, C)) 144
+define dso_local <4 x i32> @norA_xorB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_xorB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 144
+; CHECK-NEXT: blr
+entry:
+ %xor = xor <4 x i32> %B, %C
+ %or = or <4 x i32> %A, %xor
+ %neg = xor <4 x i32> %or, <i32 -1, i32 -1, i32 -1, i32 -1>
+ ret <4 x i32> %neg
+}
+
+; (nor A, B, C) 128
+define dso_local <4 x i32> @norA_B_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: norA_B_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 128
+; CHECK-NEXT: blr
+entry:
+ %or = or <4 x i32> %B, %C
+ %or1 = or <4 x i32> %A, %or
+ %neg = xor <4 x i32> %or1, <i32 -1, i32 -1, i32 -1, i32 -1>
+ ret <4 x i32> %neg
+}
+
+; (or A, (and B, C)) 31
+define dso_local <4 x i32> @orA_andB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_andB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 31
+; CHECK-NEXT: blr
+entry:
+ %and = and <4 x i32> %B, %C
+ %or = or <4 x i32> %A, %and
+ ret <4 x i32> %or
+}
+
+; (or A, (eqv B, C)) 159
+define dso_local <4 x i32> @orA_eqvB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_eqvB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 159
+; CHECK-NEXT: blr
+entry:
+ %xor = xor <4 x i32> %B, %C
+ %neg = xor <4 x i32> %xor, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %or = or <4 x i32> %A, %neg
+ ret <4 x i32> %or
+}
+
+; (or A, (nand B, C)) 239
+define dso_local <4 x i32> @orA_nandB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_nandB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 239
+; CHECK-NEXT: blr
+entry:
+ %and = and <4 x i32> %B, %C
+ %neg = xor <4 x i32> %and, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %or = or <4 x i32> %A, %neg
+ ret <4 x i32> %or
+}
+
+; (or A, (nor B, C)) 143
+define dso_local <4 x i32> @orA_norB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_norB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 143
+; CHECK-NEXT: blr
+entry:
+ %or = or <4 x i32> %B, %C
+ %neg = xor <4 x i32> %or, <i32 -1, i32 -1, i32 -1, i32 -1>
+ %or1 = or <4 x i32> %A, %neg
+ ret <4 x i32> %or1
+}
+
+; (or A, (xor B, C)) 111
+define dso_local <4 x i32> @orA_xorB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_xorB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 111
+; CHECK-NEXT: blr
+entry:
+ %xor = xor <4 x i32> %B, %C
+ %or = or <4 x i32> %A, %xor
+ ret <4 x i32> %or
+}
+
+; (or A, B, C) 127
+define dso_local <4 x i32> @orA_B_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: orA_B_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 127
+; CHECK-NEXT: blr
+entry:
+ %or = or <4 x i32> %B, %C
+ %or1 = or <4 x i32> %A, %or
+ ret <4 x i32> %or1
+}
+
+; (xor A, (and B, C)) 30
+define dso_local <4 x i32> @xorA_andB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: xorA_andB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 30
+; CHECK-NEXT: blr
+entry:
+ %and = and <4 x i32> %B, %C
+ %xor = xor <4 x i32> %A, %and
+ ret <4 x i32> %xor
+}
+
+; (xor A, (or B, C)) 120
+define dso_local <4 x i32> @xorA_orB_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: xorA_orB_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 120
+; CHECK-NEXT: blr
+entry:
+ %or = or <4 x i32> %B, %C
+ %xor = xor <4 x i32> %A, %or
+ ret <4 x i32> %xor
+}
+
+; (xor A, B, C) 105
+define dso_local <4 x i32> @xorA_B_C(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C) local_unnamed_addr #0 {
+; CHECK-LABEL: xorA_B_C:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xxeval v2, v2, v3, v4, 105
+; CHECK-NEXT: blr
+entry:
+ %xor = xor <4 x i32> %B, %C
+ %xor1 = xor <4 x i32> %A, %xor
+ ret <4 x i32> %xor1
+}
+
+attributes #0 = { nounwind }
More information about the llvm-commits
mailing list