[InstCombine] Fold `(icmp eq/ne (add nuw x, y), 0)` -> `(icmp eq/ne (or x, y), 0)` #88088

goldsteinn · 2024-04-09T04:42:41Z

[InstCombine] Add tests for folding (icmp eq/ne (add nuw x, y), 0); NFC
[InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> (icmp eq/ne (or x, y), 0)

(icmp eq/ne (or x, y), 0) is probably easier to analyze than (icmp eq/ne x, -y)

Proof: https://alive2.llvm.org/ce/z/2-VTb6

llvmbot · 2024-04-09T04:43:12Z

@llvm/pr-subscribers-llvm-transforms

Author: None (goldsteinn)

Changes

[InstCombine] Add tests for folding (icmp eq/ne (add nuw x, y), 0); NFC
[InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> (icmp eq/ne (or x, y), 0)

(icmp eq/ne (or x, y), 0) is probably easier to analyze than (icmp eq/ne x, -y)

Proof: https://alive2.llvm.org/ce/z/2-VTb6

Full diff: https://github.com/llvm/llvm-project/pull/88088.diff

2 Files Affected:

(modified) llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp (+5)
(modified) llvm/test/Transforms/InstCombine/icmp-add.ll (+39-24)

diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
index db302d7e526844..53aa84d53f3085 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
@@ -3453,6 +3453,11 @@ Instruction *InstCombinerImpl::foldICmpBinOpEqualityWithConstant(
       if (Value *NegVal = dyn_castNegVal(BOp0))
         return new ICmpInst(Pred, NegVal, BOp1);
       if (BO->hasOneUse()) {
+        // (add nuw A, B) != 0 -> (or A, B) != 0
+        if (match(BO, m_NUWAdd(m_Value(), m_Value()))) {
+          Value *Or = Builder.CreateOr(BOp0, BOp1);
+          return new ICmpInst(Pred, Or, Constant::getNullValue(BO->getType()));
+        }
         Value *Neg = Builder.CreateNeg(BOp1);
         Neg->takeName(BO);
         return new ICmpInst(Pred, BOp0, Neg);
diff --git a/llvm/test/Transforms/InstCombine/icmp-add.ll b/llvm/test/Transforms/InstCombine/icmp-add.ll
index b99ed20d7d431c..e71a3af4b5be4f 100644
--- a/llvm/test/Transforms/InstCombine/icmp-add.ll
+++ b/llvm/test/Transforms/InstCombine/icmp-add.ll
@@ -9,10 +9,8 @@ declare void @use(i32)
 define i1 @cvt_icmp_0_zext_plus_zext_eq_i16(i16 %arg, i16 %arg1) {
 ; CHECK-LABEL: @cvt_icmp_0_zext_plus_zext_eq_i16(
 ; CHECK-NEXT:  bb:
-; CHECK-NEXT:    [[I:%.*]] = zext i16 [[ARG:%.*]] to i32
-; CHECK-NEXT:    [[I2:%.*]] = zext i16 [[ARG1:%.*]] to i32
-; CHECK-NEXT:    [[I3:%.*]] = sub nsw i32 0, [[I]]
-; CHECK-NEXT:    [[I4:%.*]] = icmp eq i32 [[I2]], [[I3]]
+; CHECK-NEXT:    [[TMP0:%.*]] = or i16 [[ARG1:%.*]], [[ARG:%.*]]
+; CHECK-NEXT:    [[I4:%.*]] = icmp eq i16 [[TMP0]], 0
 ; CHECK-NEXT:    ret i1 [[I4]]
 ;
 bb:
@@ -27,10 +25,8 @@ bb:
 define i1 @cvt_icmp_0_zext_plus_zext_eq_i8(i8 %arg, i8 %arg1) {
 ; CHECK-LABEL: @cvt_icmp_0_zext_plus_zext_eq_i8(
 ; CHECK-NEXT:  bb:
-; CHECK-NEXT:    [[I:%.*]] = zext i8 [[ARG:%.*]] to i32
-; CHECK-NEXT:    [[I2:%.*]] = zext i8 [[ARG1:%.*]] to i32
-; CHECK-NEXT:    [[I3:%.*]] = sub nsw i32 0, [[I]]
-; CHECK-NEXT:    [[I4:%.*]] = icmp eq i32 [[I2]], [[I3]]
+; CHECK-NEXT:    [[TMP0:%.*]] = or i8 [[ARG1:%.*]], [[ARG:%.*]]
+; CHECK-NEXT:    [[I4:%.*]] = icmp eq i8 [[TMP0]], 0
 ; CHECK-NEXT:    ret i1 [[I4]]
 ;
 bb:
@@ -1802,22 +1798,17 @@ define i1 @test4(i32 %a) {
   ret i1 %c
 }
 
-define { i32, i1 } @test4multiuse(i32 %a) {
-; CHECK-LABEL: @test4multiuse(
-; CHECK-NEXT:    [[B:%.*]] = add nsw i32 [[A:%.*]], -2147483644
-; CHECK-NEXT:    [[C:%.*]] = icmp slt i32 [[A]], 2147483640
-; CHECK-NEXT:    [[TMP:%.*]] = insertvalue { i32, i1 } undef, i32 [[B]], 0
-; CHECK-NEXT:    [[RES:%.*]] = insertvalue { i32, i1 } [[TMP]], i1 [[C]], 1
-; CHECK-NEXT:    ret { i32, i1 } [[RES]]
+define {
+i32, i1 } @test4multiuse(i32 %a) {
 ;
 
-  %b = add nsw i32 %a, -2147483644
-  %c = icmp slt i32 %b, -4
+%b = add nsw i32 %a, -2147483644
+%c = icmp slt i32 %b, -4
 
-  %tmp = insertvalue { i32, i1 } undef, i32 %b, 0
-  %res = insertvalue { i32, i1 } %tmp, i1 %c, 1
+%tmp = insertvalue { i32, i1 } undef, i32 %b, 0
+%res = insertvalue { i32, i1 } %tmp, i1 %c, 1
 
-  ret { i32, i1 } %res
+ret { i32, i1 } %res
 }
 
 define <2 x i1> @test4vec(<2 x i32> %a) {
@@ -2857,7 +2848,7 @@ define i1 @icmp_add_add_C_comm2(i32 %X, i32 %b) {
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp ugt i32 [[A]], [[TMP1]]
 ; CHECK-NEXT:    ret i1 [[CMP]]
 ;
-  %a = udiv i32 42, %X ; thwart complexity-based canonicalization
+  %a = udiv i32 42, %X  ; thwart complexity-based canonicalization
   %add1 = add i32 %a, %b
   %add2 = add i32 %add1, -1
   %cmp = icmp ugt i32 %a, %add2
@@ -2871,7 +2862,7 @@ define i1 @icmp_add_add_C_comm2_pred(i32 %X, i32 %b) {
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp ule i32 [[A]], [[TMP1]]
 ; CHECK-NEXT:    ret i1 [[CMP]]
 ;
-  %a = udiv i32 42, %X ; thwart complexity-based canonicalization
+  %a = udiv i32 42, %X  ; thwart complexity-based canonicalization
   %add1 = add i32 %a, %b
   %add2 = add i32 %add1, -1
   %cmp = icmp ule i32 %a, %add2
@@ -2886,7 +2877,7 @@ define i1 @icmp_add_add_C_comm2_wrong_pred(i32 %X, i32 %b) {
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[A]], [[ADD2]]
 ; CHECK-NEXT:    ret i1 [[CMP]]
 ;
-  %a = udiv i32 42, %X ; thwart complexity-based canonicalization
+  %a = udiv i32 42, %X  ; thwart complexity-based canonicalization
   %add1 = add i32 %a, %b
   %add2 = add i32 %add1, -1
   %cmp = icmp ult i32 %a, %add2
@@ -2900,7 +2891,7 @@ define i1 @icmp_add_add_C_comm3(i32 %X, i32 %b) {
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp ugt i32 [[A]], [[TMP1]]
 ; CHECK-NEXT:    ret i1 [[CMP]]
 ;
-  %a = udiv i32 42, %X ; thwart complexity-based canonicalization
+  %a = udiv i32 42, %X  ; thwart complexity-based canonicalization
   %add1 = add i32 %b, %a
   %add2 = add i32 %add1, -1
   %cmp = icmp ugt i32 %a, %add2
@@ -3003,4 +2994,28 @@ define i1 @icmp_dec_notnonzero(i8 %x) {
   ret i1 %c
 }
 
+define i1 @icmp_addnuw_nonzero(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_addnuw_nonzero(
+; CHECK-NEXT:    [[TMP1:%.*]] = or i8 [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT:    [[C:%.*]] = icmp eq i8 [[TMP1]], 0
+; CHECK-NEXT:    ret i1 [[C]]
+;
+  %i = add nuw i8 %x, %y
+  %c = icmp eq i8 %i, 0
+  ret i1 %c
+}
+
+define i1 @icmp_addnuw_nonzero_fail_multiuse(i32 %x, i32 %y) {
+; CHECK-LABEL: @icmp_addnuw_nonzero_fail_multiuse(
+; CHECK-NEXT:    [[I:%.*]] = add nuw i32 [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT:    [[C:%.*]] = icmp eq i32 [[I]], 0
+; CHECK-NEXT:    call void @use(i32 [[I]])
+; CHECK-NEXT:    ret i1 [[C]]
+;
+  %i = add nuw i32 %x, %y
+  %c = icmp eq i32 %i, 0
+  call void @use(i32 %i)
+  ret i1 %c
+}
+
 declare void @llvm.assume(i1)

nikic · 2024-04-09T05:39:27Z

llvm/test/Transforms/InstCombine/icmp-add.ll

-; CHECK-NEXT:    [[RES:%.*]] = insertvalue { i32, i1 } [[TMP]], i1 [[C]], 1
-; CHECK-NEXT:    ret { i32, i1 } [[RES]]
+define {
+i32, i1 } @test4multiuse(i32 %a) {


Something went wrong here -- did UTC break this function?

no, I have a crappy regex to try to format my IR that failed miserably here. Ill fix

nikic

This LGTM with the test fixed. I agree that converting this to or is better.

PR Link: llvm/llvm-project#88088

… NFC

…or x, y), 0)` `(icmp eq/ne (or x, y), 0)` is probably easier to analyze than `(icmp eq/ne x, -y)` Proof: https://alive2.llvm.org/ce/z/2-VTb6

dtcxzyw · 2024-04-09T06:54:11Z

Missing fold: https://alive2.llvm.org/ce/z/DShc49

define i1 @src(i1 %x, i32 %z) {
  %22 = select i1 %x, i32 0, i32 255
  %23 = or i32 %z, %22
  %eq = icmp eq i32 %23, 0
  ret i1 %eq
}

define i1 @tgt(i1 %x, i32 %z) {
  %eq = icmp eq i32 %z, 0
  %and = and i1 %x, %eq
  ret i1 %and
}

See also https://github.com/dtcxzyw/llvm-opt-benchmark/pull/485/files#r1557018422

dtcxzyw

LGTM.

Four cases: `(icmp eq (or (select cond, 0, NonZero), Other))` -> `(and cond, (icmp eq Other, 0))` `(icmp ne (or (select cond, NonZero, 0), Other))` -> `(or cond, (icmp ne Other, 0))` `(icmp ne (or (select cond, 0, NonZero), Other))` -> `(or (not cond), (icmp ne Other, 0))` `(icmp eq (or (select cond, NonZero, 0), Other))` -> `(and (not cond), (icmp eq Other, 0))` These cases came up in tests on: llvm#88088 Proofs: https://alive2.llvm.org/ce/z/ojGo_J

goldsteinn · 2024-04-09T19:58:40Z

Missing fold: https://alive2.llvm.org/ce/z/DShc49
define i1 @src(i1 %x, i32 %z) {
  %22 = select i1 %x, i32 0, i32 255
  %23 = or i32 %z, %22
  %eq = icmp eq i32 %23, 0
  ret i1 %eq
}

define i1 @tgt(i1 %x, i32 %z) {
  %eq = icmp eq i32 %z, 0
  %and = and i1 %x, %eq
  ret i1 %and
}
See also https://github.com/dtcxzyw/llvm-opt-benchmark/pull/485/files#r1557018422

See: #88183

Four cases: `(icmp eq (or (select cond, 0, NonZero), Other))` -> `(and cond, (icmp eq Other, 0))` `(icmp ne (or (select cond, NonZero, 0), Other))` -> `(or cond, (icmp ne Other, 0))` `(icmp ne (or (select cond, 0, NonZero), Other))` -> `(or (not cond), (icmp ne Other, 0))` `(icmp eq (or (select cond, NonZero, 0), Other))` -> `(and (not cond), (icmp eq Other, 0))` These cases came up in tests on: llvm#88088 Proofs: https://alive2.llvm.org/ce/z/ojGo_J

goldsteinn requested a review from nikic as a code owner April 9, 2024 04:42

llvmbot added the llvm:transforms label Apr 9, 2024

goldsteinn changed the title ~~goldsteinn/add nuw to or~~ [InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> (icmp eq/ne (or x, y), 0) Apr 9, 2024

goldsteinn requested a review from dtcxzyw April 9, 2024 04:43

nikic reviewed Apr 9, 2024

View reviewed changes

dtcxzyw added a commit to dtcxzyw/llvm-opt-benchmark that referenced this pull request Apr 9, 2024

pre-commit: test PR88088

3959efc

PR Link: llvm/llvm-project#88088

dtcxzyw mentioned this pull request Apr 9, 2024

pre-commit: test PR88088 dtcxzyw/llvm-opt-benchmark#485

Open

goldsteinn added 2 commits April 9, 2024 01:06

[InstCombine] Add tests for folding (icmp eq/ne (add nuw x, y), 0);…

98bfa63

… NFC

[InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> `(icmp eq/ne (…

08ea7f3

…or x, y), 0)` `(icmp eq/ne (or x, y), 0)` is probably easier to analyze than `(icmp eq/ne x, -y)` Proof: https://alive2.llvm.org/ce/z/2-VTb6

goldsteinn force-pushed the goldsteinn/add-nuw-to-or branch from 4f0aed1 to 08ea7f3 Compare April 9, 2024 06:06

dtcxzyw approved these changes Apr 9, 2024

View reviewed changes

nikic approved these changes Apr 9, 2024

View reviewed changes

goldsteinn closed this in 7599d47 Apr 9, 2024

goldsteinn mentioned this pull request Apr 9, 2024

[InstCombine] Fold (icmp eq/ne (or (select cond, 0/NZ, 0/NZ), X), 0) #88183

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[InstCombine] Fold `(icmp eq/ne (add nuw x, y), 0)` -> `(icmp eq/ne (or x, y), 0)` #88088

[InstCombine] Fold `(icmp eq/ne (add nuw x, y), 0)` -> `(icmp eq/ne (or x, y), 0)` #88088

goldsteinn commented Apr 9, 2024 •

edited

llvmbot commented Apr 9, 2024

nikic Apr 9, 2024

goldsteinn Apr 9, 2024 •

edited

goldsteinn Apr 9, 2024

nikic left a comment

dtcxzyw commented Apr 9, 2024

dtcxzyw left a comment

goldsteinn commented Apr 9, 2024

[InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> (icmp eq/ne (or x, y), 0) #88088

[InstCombine] Fold (icmp eq/ne (add nuw x, y), 0) -> (icmp eq/ne (or x, y), 0) #88088

Conversation

goldsteinn commented Apr 9, 2024 • edited

llvmbot commented Apr 9, 2024

nikic Apr 9, 2024

Choose a reason for hiding this comment

goldsteinn Apr 9, 2024 • edited

Choose a reason for hiding this comment

goldsteinn Apr 9, 2024

Choose a reason for hiding this comment

nikic left a comment

Choose a reason for hiding this comment

dtcxzyw commented Apr 9, 2024

dtcxzyw left a comment

Choose a reason for hiding this comment

goldsteinn commented Apr 9, 2024

[InstCombine] Fold `(icmp eq/ne (add nuw x, y), 0)` -> `(icmp eq/ne (or x, y), 0)` #88088

[InstCombine] Fold `(icmp eq/ne (add nuw x, y), 0)` -> `(icmp eq/ne (or x, y), 0)` #88088

goldsteinn commented Apr 9, 2024 •

edited

goldsteinn Apr 9, 2024 •

edited