[InstCombine] Add transforms for (or/and (icmp eq/ne X,0),(icmp eq/ne X,Pow2OrZero))
#94648
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-llvm-transforms Author: None (goldsteinn) Changes
Full diff: https://github.com/llvm/llvm-project/pull/94648.diff 2 Files Affected:
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
index 8695e9e69df2..89eb583ee427 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
@@ -701,6 +701,38 @@ Value *InstCombinerImpl::simplifyRangeCheck(ICmpInst *Cmp0, ICmpInst *Cmp1,
return Builder.CreateICmp(NewPred, Input, RangeEnd);
}
+// (or (icmp eq X, 0), (icmp eq X, Pow2OrZero))
+// -> (icmp eq (and X, Pow2OrZero), X)
+// (and (icmp ne X, 0), (icmp ne X, Pow2OrZero))
+// -> (icmp ne (and X, Pow2OrZero), X)
+static Value *
+foldAndOrOfICmpsWithPow2AndWithZero(InstCombiner::BuilderTy &Builder,
+ ICmpInst *LHS, ICmpInst *RHS, bool IsAnd,
+ const SimplifyQuery &Q) {
+ CmpInst::Predicate Pred = IsAnd ? CmpInst::ICMP_NE : CmpInst::ICMP_EQ;
+ // Make sure we have right compares for our op.
+ if (LHS->getPredicate() != Pred || RHS->getPredicate() != Pred)
+ return nullptr;
+
+ // Make it so we can match LHS against the (icmp eq/ne X, 0) just for
+ // simplicity.
+ if (match(RHS->getOperand(1), m_Zero()))
+ std::swap(LHS, RHS);
+
+ Value *Pow2, *Op;
+ // Match the desired pattern:
+ // LHS: (icmp eq/ne X, 0)
+ // RHS: (icmp eq/ne X, Pow2OrZero)
+ if (!match(LHS, m_ICmp(Pred, m_Value(Op), m_Zero())) ||
+ !match(RHS, m_c_ICmp(Pred, m_Specific(Op), m_Value(Pow2))) ||
+ !isKnownToBeAPowerOfTwo(Pow2, Q.DL, /*OrZero*/ true, /*Depth*/ 0, Q.AC,
+ Q.CxtI, Q.DT))
+ return nullptr;
+
+ Value *And = Builder.CreateAnd(Op, Pow2);
+ return Builder.CreateICmp(Pred, And, Op);
+}
+
// Fold (iszero(A & K1) | iszero(A & K2)) -> (A & (K1 | K2)) != (K1 | K2)
// Fold (!iszero(A & K1) & !iszero(A & K2)) -> (A & (K1 | K2)) == (K1 | K2)
Value *InstCombinerImpl::foldAndOrOfICmpsOfAndWithPow2(ICmpInst *LHS,
@@ -3240,6 +3272,11 @@ Value *InstCombinerImpl::foldAndOrOfICmps(ICmpInst *LHS, ICmpInst *RHS,
ICmpInst::Predicate PredL = LHS->getPredicate(), PredR = RHS->getPredicate();
Value *LHS0 = LHS->getOperand(0), *RHS0 = RHS->getOperand(0);
Value *LHS1 = LHS->getOperand(1), *RHS1 = RHS->getOperand(1);
+ if (!IsLogical)
+ if (Value *V =
+ foldAndOrOfICmpsWithPow2AndWithZero(Builder, LHS, RHS, IsAnd, Q))
+ return V;
+
const APInt *LHSC = nullptr, *RHSC = nullptr;
match(LHS1, m_APInt(LHSC));
match(RHS1, m_APInt(RHSC));
diff --git a/llvm/test/Transforms/InstCombine/and-or-icmps.ll b/llvm/test/Transforms/InstCombine/and-or-icmps.ll
index c20f48a985b3..14ac3e404774 100644
--- a/llvm/test/Transforms/InstCombine/and-or-icmps.ll
+++ b/llvm/test/Transforms/InstCombine/and-or-icmps.ll
@@ -3057,9 +3057,9 @@ define i32 @icmp_slt_0_or_icmp_add_1_sge_100_i32_fail(i32 %x) {
define i1 @logical_and_icmps1(i32 %a, i1 %other_cond) {
; CHECK-LABEL: @logical_and_icmps1(
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[CMP3:%.*]] = icmp ult i32 [[A:%.*]], 10086
-; CHECK-NEXT: [[RET2:%.*]] = select i1 [[RET1:%.*]], i1 [[CMP3]], i1 false
-; CHECK-NEXT: ret i1 [[RET2]]
+; CHECK-NEXT: [[TMP0:%.*]] = icmp ult i32 [[A:%.*]], 10086
+; CHECK-NEXT: [[RET:%.*]] = select i1 [[OTHER_COND:%.*]], i1 [[TMP0]], i1 false
+; CHECK-NEXT: ret i1 [[RET]]
;
entry:
%cmp1 = icmp sgt i32 %a, -1
@@ -3085,9 +3085,9 @@ entry:
define <4 x i1> @logical_and_icmps_vec1(<4 x i32> %a, <4 x i1> %other_cond) {
; CHECK-LABEL: @logical_and_icmps_vec1(
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[CMP3:%.*]] = icmp ult <4 x i32> [[A:%.*]], <i32 10086, i32 10086, i32 10086, i32 10086>
-; CHECK-NEXT: [[RET2:%.*]] = select <4 x i1> [[RET1:%.*]], <4 x i1> [[CMP3]], <4 x i1> zeroinitializer
-; CHECK-NEXT: ret <4 x i1> [[RET2]]
+; CHECK-NEXT: [[TMP0:%.*]] = icmp ult <4 x i32> [[A:%.*]], <i32 10086, i32 10086, i32 10086, i32 10086>
+; CHECK-NEXT: [[RET:%.*]] = select <4 x i1> [[OTHER_COND:%.*]], <4 x i1> [[TMP0]], <4 x i1> zeroinitializer
+; CHECK-NEXT: ret <4 x i1> [[RET]]
;
entry:
%cmp1 = icmp sgt <4 x i32> %a, <i32 -1, i32 -1, i32 -1, i32 -1 >
@@ -3113,3 +3113,147 @@ entry:
%ret = select i1 %logical_and, i1 %cmp2, i1 false
ret i1 %ret
}
+
+
+define i1 @icmp_eq_or_z_or_pow2orz(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_eq_or_z_or_pow2orz(
+; CHECK-NEXT: [[NY:%.*]] = sub i8 0, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and i8 [[NY]], [[Y]]
+; CHECK-NEXT: [[TMP1:%.*]] = and i8 [[POW2ORZ]], [[X:%.*]]
+; CHECK-NEXT: [[R:%.*]] = icmp eq i8 [[TMP1]], [[X]]
+; CHECK-NEXT: ret i1 [[R]]
+;
+ %ny = sub i8 0, %y
+ %pow2orz = and i8 %ny, %y
+
+ %c0 = icmp eq i8 %x, 0
+ %cp2 = icmp eq i8 %x, %pow2orz
+ %r = or i1 %c0, %cp2
+ ret i1 %r
+}
+
+
+define i1 @icmp_eq_or_z_or_pow2orz_fail_logic_or(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_eq_or_z_or_pow2orz_fail_logic_or(
+; CHECK-NEXT: [[NY:%.*]] = sub i8 0, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and i8 [[NY]], [[Y]]
+; CHECK-NEXT: [[C0:%.*]] = icmp eq i8 [[X:%.*]], 0
+; CHECK-NEXT: [[CP2:%.*]] = icmp eq i8 [[POW2ORZ]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = select i1 [[C0]], i1 true, i1 [[CP2]]
+; CHECK-NEXT: ret i1 [[R]]
+;
+ %ny = sub i8 0, %y
+ %pow2orz = and i8 %ny, %y
+
+ %c0 = icmp eq i8 %x, 0
+ %cp2 = icmp eq i8 %x, %pow2orz
+ %r = select i1 %c0, i1 true, i1 %cp2
+ ret i1 %r
+}
+
+
+define <2 x i1> @icmp_ne_and_z_and_pow2orz(<2 x i8> %x, <2 x i8> %y) {
+; CHECK-LABEL: @icmp_ne_and_z_and_pow2orz(
+; CHECK-NEXT: [[NY:%.*]] = sub <2 x i8> zeroinitializer, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and <2 x i8> [[NY]], [[Y]]
+; CHECK-NEXT: [[TMP1:%.*]] = and <2 x i8> [[POW2ORZ]], [[X:%.*]]
+; CHECK-NEXT: [[R:%.*]] = icmp ne <2 x i8> [[TMP1]], [[X]]
+; CHECK-NEXT: ret <2 x i1> [[R]]
+;
+ %ny = sub <2 x i8> zeroinitializer, %y
+ %pow2orz = and <2 x i8> %ny, %y
+
+ %c0 = icmp ne <2 x i8> %x, zeroinitializer
+ %cp2 = icmp ne <2 x i8> %x, %pow2orz
+ %r = and <2 x i1> %c0, %cp2
+ ret <2 x i1> %r
+}
+
+define <2 x i1> @icmp_ne_and_z_and_pow2orz_fail_logic_and(<2 x i8> %x, <2 x i8> %y) {
+; CHECK-LABEL: @icmp_ne_and_z_and_pow2orz_fail_logic_and(
+; CHECK-NEXT: [[NY:%.*]] = sub <2 x i8> zeroinitializer, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and <2 x i8> [[NY]], [[Y]]
+; CHECK-NEXT: [[C0:%.*]] = icmp ne <2 x i8> [[X:%.*]], zeroinitializer
+; CHECK-NEXT: [[CP2:%.*]] = icmp ne <2 x i8> [[POW2ORZ]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = select <2 x i1> [[C0]], <2 x i1> [[CP2]], <2 x i1> zeroinitializer
+; CHECK-NEXT: ret <2 x i1> [[R]]
+;
+ %ny = sub <2 x i8> zeroinitializer, %y
+ %pow2orz = and <2 x i8> %ny, %y
+
+ %c0 = icmp ne <2 x i8> %x, zeroinitializer
+ %cp2 = icmp ne <2 x i8> %x, %pow2orz
+ %r = select <2 x i1> %c0, <2 x i1> %cp2, <2 x i1> zeroinitializer
+ ret <2 x i1> %r
+}
+
+define i1 @icmp_eq_or_z_or_pow2orz_fail_not_pow2(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_eq_or_z_or_pow2orz_fail_not_pow2(
+; CHECK-NEXT: [[NY:%.*]] = sub i8 1, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and i8 [[NY]], [[Y]]
+; CHECK-NEXT: [[C0:%.*]] = icmp eq i8 [[X:%.*]], 0
+; CHECK-NEXT: [[CP2:%.*]] = icmp eq i8 [[POW2ORZ]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = or i1 [[C0]], [[CP2]]
+; CHECK-NEXT: ret i1 [[R]]
+;
+ %ny = sub i8 1, %y
+ %pow2orz = and i8 %ny, %y
+
+ %c0 = icmp eq i8 %x, 0
+ %cp2 = icmp eq i8 %x, %pow2orz
+ %r = or i1 %c0, %cp2
+ ret i1 %r
+}
+
+define i1 @icmp_eq_or_z_or_pow2orz_fail_nonzero_const(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_eq_or_z_or_pow2orz_fail_nonzero_const(
+; CHECK-NEXT: [[NY:%.*]] = sub i8 0, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and i8 [[NY]], [[Y]]
+; CHECK-NEXT: [[C0:%.*]] = icmp eq i8 [[X:%.*]], 1
+; CHECK-NEXT: [[CP2:%.*]] = icmp eq i8 [[POW2ORZ]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = or i1 [[C0]], [[CP2]]
+; CHECK-NEXT: ret i1 [[R]]
+;
+ %ny = sub i8 0, %y
+ %pow2orz = and i8 %ny, %y
+
+ %c0 = icmp eq i8 %x, 1
+ %cp2 = icmp eq i8 %x, %pow2orz
+ %r = or i1 %c0, %cp2
+ ret i1 %r
+}
+
+define <2 x i1> @icmp_ne_and_z_and_pow2orz_fail_bad_pred(<2 x i8> %x, <2 x i8> %y) {
+; CHECK-LABEL: @icmp_ne_and_z_and_pow2orz_fail_bad_pred(
+; CHECK-NEXT: [[NY:%.*]] = sub <2 x i8> zeroinitializer, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and <2 x i8> [[NY]], [[Y]]
+; CHECK-NEXT: [[TMP1:%.*]] = or <2 x i8> [[POW2ORZ]], [[X:%.*]]
+; CHECK-NEXT: [[R:%.*]] = icmp eq <2 x i8> [[TMP1]], zeroinitializer
+; CHECK-NEXT: ret <2 x i1> [[R]]
+;
+ %ny = sub <2 x i8> zeroinitializer, %y
+ %pow2orz = and <2 x i8> %ny, %y
+
+ %c0 = icmp eq <2 x i8> %x, zeroinitializer
+ %cp2 = icmp eq <2 x i8> %x, %pow2orz
+ %r = and <2 x i1> %c0, %cp2
+ ret <2 x i1> %r
+}
+
+define i1 @icmp_eq_or_z_or_pow2orz_fail_bad_pred2(i8 %x, i8 %y) {
+; CHECK-LABEL: @icmp_eq_or_z_or_pow2orz_fail_bad_pred2(
+; CHECK-NEXT: [[NY:%.*]] = sub i8 0, [[Y:%.*]]
+; CHECK-NEXT: [[POW2ORZ:%.*]] = and i8 [[NY]], [[Y]]
+; CHECK-NEXT: [[C0:%.*]] = icmp slt i8 [[X:%.*]], 1
+; CHECK-NEXT: [[CP2:%.*]] = icmp sge i8 [[POW2ORZ]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = or i1 [[C0]], [[CP2]]
+; CHECK-NEXT: ret i1 [[R]]
+;
+ %ny = sub i8 0, %y
+ %pow2orz = and i8 %ny, %y
+
+ %c0 = icmp sle i8 %x, 0
+ %cp2 = icmp sle i8 %x, %pow2orz
+ %r = or i1 %c0, %cp2
+ ret i1 %r
+}
|
goldsteinn
changed the title
(or/and (icmp eq/ne X,0),(icmp eq/ne X,Pow2OrZero))
|
NB: this was approved by nikic a while back: https://reviews.llvm.org/D157312 Hopefully its time to get this in :) |
|
Implementation still looks good to me. I think @dtcxzyw's tests should give us a much better answer about whether this canonicalization is problematic or not now :) |
dtcxzyw
added a commit
to dtcxzyw/llvm-opt-benchmark
that referenced
this pull request
Jun 7, 2024
dtcxzyw
reviewed
Jun 7, 2024
We need a one-use check to avoid creating more instructions than before. |
goldsteinn
force-pushed
the
goldsteinn/and-or-p2-w-zero
branch
from
bac02a8 to
52823f5
Compare
|
@dtcxzyw Could you please rerun tests with the new one-use checks? |
dtcxzyw
added a commit
to dtcxzyw/llvm-opt-benchmark
that referenced
this pull request
Jun 8, 2024
dtcxzyw
reviewed
Jun 8, 2024
…cmp eq/ne X,Pow2OrZero))`; NFC
goldsteinn
force-pushed
the
goldsteinn/and-or-p2-w-zero
branch
from
52823f5 to
71aa5ab
Compare
dtcxzyw
added a commit
to dtcxzyw/llvm-opt-benchmark
that referenced
this pull request
Jun 8, 2024
nikic
reviewed
Jun 8, 2024
| if (!IsLogical) | ||
| if (Value *V = | ||
| foldAndOrOfICmpsWithPow2AndWithZero(Builder, LHS, RHS, IsAnd, Q)) | ||
| return V; |
There was a problem hiding this comment.
Why did you add this fold in this position? New exotic folds usually get added to the end, not the start of functions. Especially weird to see this interleaved with a bunch of variable declarations that you aren't using.
There was a problem hiding this comment.
Moving to the end.
|
Did this fold produce any positive diffs? It seems like it's always either neutral (inverting a predicate) or a regression, but maybe I didn't see some. |
dtcxzyw
reviewed
Jun 8, 2024
…e X,Pow2OrZero))`
`(or (icmp eq X, 0), (icmp eq X, Pow2OrZero))`
--> `(icmp eq (and X, Pow2OrZero), X)`
`(and (icmp ne X, 0), (icmp ne X, Pow2OrZero))`
--> `(icmp ne (and X, Pow2OrZero), X)`
Proofs: https://alive2.llvm.org/ce/z/nPo2BN
goldsteinn
force-pushed
the
goldsteinn/and-or-p2-w-zero
branch
from
71aa5ab to
d82214b
Compare
goldsteinn
added a commit
to goldsteinn/llvm-project
that referenced
this pull request
Jun 8, 2024
`(icmp eq/ne (and X, P2), (and X, -P2))`
-> `(icmp ult/uge X, P2 * 2)`
If `P2` is constant, we can perform this fold profitably even if the
`and` ops are multi-use.
If `P2` is not constant, then we only perform the fold if the `and`
ops are multi-use. This, however, saves an instruction compared to the
`xor` + `and` variant.
NB: This came up in some of the diffs resulting from llvm#94648
Proofs: https://alive2.llvm.org/ce/z/mfd3G9
goldsteinn
added a commit
to goldsteinn/llvm-project
that referenced
this pull request
Jun 8, 2024
`(icmp eq/ne (and X, P2), (and X, -P2))`
-> `(icmp ult/uge X, P2 * 2)`
If `P2` is constant, we can perform this fold profitably even if the
`and` ops are multi-use.
NB: This came up in some of the diffs resulting from llvm#94648
Proofs: https://alive2.llvm.org/ce/z/mfd3G9
See bench/php/optimized/php_date.ll |
|
ping |
dtcxzyw
added a commit
to dtcxzyw/llvm-opt-benchmark
that referenced
this pull request
Jul 9, 2024
nikic
approved these changes
Jul 9, 2024
| // be one-use. We don't create additional instructions if only one | ||
| // of them is one-use. So cases where one is one-use and the other | ||
| // is two-use might be profitable. | ||
| if (!match(LHS, m_OneUse(m_ICmp(Pred, m_Value(Op), m_Zero()))) || |
There was a problem hiding this comment.
As a side note, we should probably add m_ICmp variants without predicate, as we have a bunch of places where the fetched predicate is just discarded, but kind of looks like it's part of the match. (Or maybe alternatively add something like m_SpecificICmp that matches the passed predicate instead of it being an out parameter.)
aaryanshukla
pushed a commit
to aaryanshukla/llvm-project
that referenced
this pull request
Jul 14, 2024
…e X,Pow2OrZero))`
`(or (icmp eq X, 0), (icmp eq X, Pow2OrZero))`
--> `(icmp eq (and X, Pow2OrZero), X)`
`(and (icmp ne X, 0), (icmp ne X, Pow2OrZero))`
--> `(icmp ne (and X, Pow2OrZero), X)`
Proofs: https://alive2.llvm.org/ce/z/nPo2BN
Closes llvm#94648
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
(or/and (icmp eq/ne X,0),(icmp eq/ne X,Pow2OrZero)); NFC(or/and (icmp eq/ne X,0),(icmp eq/ne X,Pow2OrZero))(or (icmp eq X, 0), (icmp eq X, Pow2OrZero))-->
(icmp eq (and X, Pow2OrZero), X)(and (icmp ne X, 0), (icmp ne X, Pow2OrZero))-->
(icmp ne (and X, Pow2OrZero), X)Proofs: https://alive2.llvm.org/ce/z/nPo2BN