[llvm] r330089 - [InstCombine] Enable Add/Sub simplifications with only 'reassoc' FMF
Warren Ristow via llvm-commits
llvm-commits at lists.llvm.org
Sat Apr 14 12:18:28 PDT 2018
Author: wristow
Date: Sat Apr 14 12:18:28 2018
New Revision: 330089
URL: http://llvm.org/viewvc/llvm-project?rev=330089&view=rev
Log:
[InstCombine] Enable Add/Sub simplifications with only 'reassoc' FMF
These simplifications were previously enabled only with isFast(), but that
is more restrictive than required. Since r317488, FMF has 'reassoc' to
control these cases at a finer level.
Modified:
llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp
llvm/trunk/test/Transforms/InstCombine/fast-math.ll
llvm/trunk/test/Transforms/Reassociate/fast-MissedTree.ll
llvm/trunk/test/Transforms/Reassociate/fast-SubReassociate.ll
llvm/trunk/test/Transforms/Reassociate/fast-basictest.ll
Modified: llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp?rev=330089&r1=330088&r2=330089&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp (original)
+++ llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp Sat Apr 14 12:18:28 2018
@@ -511,7 +511,8 @@ Value *FAddCombine::performFactorization
}
Value *FAddCombine::simplify(Instruction *I) {
- assert(I->isFast() && "Expected 'fast' instruction");
+ assert(I->hasAllowReassoc() && I->hasNoSignedZeros() &&
+ "Expected 'reassoc'+'nsz' instruction");
// Currently we are not able to handle vector type.
if (I->getType()->isVectorTy())
@@ -1378,7 +1379,7 @@ Instruction *InstCombiner::visitFAdd(Bin
if (Value *V = SimplifySelectsFeedingBinaryOp(I, LHS, RHS))
return replaceInstUsesWith(I, V);
- if (I.isFast()) {
+ if (I.hasAllowReassoc() && I.hasNoSignedZeros()) {
if (Value *V = FAddCombine(Builder).simplify(&I))
return replaceInstUsesWith(I, V);
}
@@ -1747,7 +1748,7 @@ Instruction *InstCombiner::visitFSub(Bin
if (Value *V = SimplifySelectsFeedingBinaryOp(I, Op0, Op1))
return replaceInstUsesWith(I, V);
- if (I.isFast()) {
+ if (I.hasAllowReassoc() && I.hasNoSignedZeros()) {
if (Value *V = FAddCombine(Builder).simplify(&I))
return replaceInstUsesWith(I, V);
}
Modified: llvm/trunk/test/Transforms/InstCombine/fast-math.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/InstCombine/fast-math.ll?rev=330089&r1=330088&r2=330089&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/InstCombine/fast-math.ll (original)
+++ llvm/trunk/test/Transforms/InstCombine/fast-math.ll Sat Apr 14 12:18:28 2018
@@ -37,16 +37,42 @@ define float @fold2(float %a) {
}
; C * f1 + f1 = (C+1) * f1
+; TODO: The particular case where C is 2 (so the folded result is 3.0*f1) is
+; always safe, and so doesn't need any FMF.
+; That is, (x + x + x) and (3*x) each have only a single rounding.
define double @fold3(double %f1) {
; CHECK-LABEL: @fold3(
-; CHECK-NEXT: [[TMP1:%.*]] = fmul fast double [[F1:%.*]], 3.000000e+00
+; CHECK-NEXT: [[TMP1:%.*]] = fmul fast double [[F1:%.*]], 6.000000e+00
; CHECK-NEXT: ret double [[TMP1]]
;
- %t1 = fmul fast double 2.000000e+00, %f1
+ %t1 = fmul fast double 5.000000e+00, %f1
%t2 = fadd fast double %f1, %t1
ret double %t2
}
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define double @fold3_reassoc_nsz(double %f1) {
+; CHECK-LABEL: @fold3_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc nsz double [[F1:%.*]], 6.000000e+00
+; CHECK-NEXT: ret double [[TMP1]]
+;
+ %t1 = fmul reassoc nsz double 5.000000e+00, %f1
+ %t2 = fadd reassoc nsz double %f1, %t1
+ ret double %t2
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to f1 * 6.0.
+define double @fold3_reassoc(double %f1) {
+; CHECK-LABEL: @fold3_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc double [[F1:%.*]], 5.000000e+00
+; CHECK-NEXT: [[TMP2:%.*]] = fadd reassoc double [[TMP1]], [[F1]]
+; CHECK-NEXT: ret double [[TMP2]]
+;
+ %t1 = fmul reassoc double 5.000000e+00, %f1
+ %t2 = fadd reassoc double %f1, %t1
+ ret double %t2
+}
+
; (C1 - X) + (C2 - Y) => (C1+C2) - (X + Y)
define float @fold4(float %f1, float %f2) {
; CHECK-LABEL: @fold4(
@@ -60,8 +86,35 @@ define float @fold4(float %f1, float %f2
ret float %add
}
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold4_reassoc_nsz(float %f1, float %f2) {
+; CHECK-LABEL: @fold4_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc nsz float [[F1:%.*]], [[F2:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fsub reassoc nsz float 9.000000e+00, [[TMP1]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %sub = fsub float 4.000000e+00, %f1
+ %sub1 = fsub float 5.000000e+00, %f2
+ %add = fadd reassoc nsz float %sub, %sub1
+ ret float %add
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to (9.0 - (f1 + f2)).
+define float @fold4_reassoc(float %f1, float %f2) {
+; CHECK-LABEL: @fold4_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub float 4.000000e+00, [[F1:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fsub float 5.000000e+00, [[F2:%.*]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %sub = fsub float 4.000000e+00, %f1
+ %sub1 = fsub float 5.000000e+00, %f2
+ %add = fadd reassoc float %sub, %sub1
+ ret float %add
+}
+
; (X + C1) + C2 => X + (C1 + C2)
-define float @fold5(float %f1, float %f2) {
+define float @fold5(float %f1) {
; CHECK-LABEL: @fold5(
; CHECK-NEXT: [[ADD1:%.*]] = fadd fast float [[F1:%.*]], 9.000000e+00
; CHECK-NEXT: ret float [[ADD1]]
@@ -71,15 +124,65 @@ define float @fold5(float %f1, float %f2
ret float %add1
}
-; (X + X) + X => 3.0 * X
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold5_reassoc_nsz(float %f1) {
+; CHECK-LABEL: @fold5_reassoc_nsz(
+; CHECK-NEXT: [[ADD1:%.*]] = fadd reassoc nsz float [[F1:%.*]], 9.000000e+00
+; CHECK-NEXT: ret float [[ADD1]]
+;
+ %add = fadd float %f1, 4.000000e+00
+ %add1 = fadd reassoc nsz float %add, 5.000000e+00
+ ret float %add1
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to f1 + 9.0
+define float @fold5_reassoc(float %f1) {
+; CHECK-LABEL: @fold5_reassoc(
+; CHECK-NEXT: [[ADD:%.*]] = fadd float [[F1:%.*]], 4.000000e+00
+; CHECK-NEXT: [[ADD1:%.*]] = fadd reassoc float [[ADD]], 5.000000e+00
+; CHECK-NEXT: ret float [[ADD1]]
+;
+ %add = fadd float %f1, 4.000000e+00
+ %add1 = fadd reassoc float %add, 5.000000e+00
+ ret float %add1
+}
+
+; (X + X) + X + X => 4.0 * X
define float @fold6(float %f1) {
; CHECK-LABEL: @fold6(
-; CHECK-NEXT: [[TMP1:%.*]] = fmul fast float [[F1:%.*]], 3.000000e+00
+; CHECK-NEXT: [[TMP1:%.*]] = fmul fast float [[F1:%.*]], 4.000000e+00
; CHECK-NEXT: ret float [[TMP1]]
;
%t1 = fadd fast float %f1, %f1
%t2 = fadd fast float %f1, %t1
- ret float %t2
+ %t3 = fadd fast float %t2, %f1
+ ret float %t3
+}
+
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold6_reassoc_nsz(float %f1) {
+; CHECK-LABEL: @fold6_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc nsz float [[F1:%.*]], 4.000000e+00
+; CHECK-NEXT: ret float [[TMP1]]
+;
+ %t1 = fadd reassoc nsz float %f1, %f1
+ %t2 = fadd reassoc nsz float %f1, %t1
+ %t3 = fadd reassoc nsz float %t2, %f1
+ ret float %t3
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to f1 * 4.0.
+define float @fold6_reassoc(float %f1) {
+; CHECK-LABEL: @fold6_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc float [[F1:%.*]], [[F1]]
+; CHECK-NEXT: [[TMP2:%.*]] = fadd reassoc float [[TMP1]], [[F1]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP2]], [[F1]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fadd reassoc float %f1, %f1
+ %t2 = fadd reassoc float %f1, %t1
+ %t3 = fadd reassoc float %t2, %f1
+ ret float %t3
}
; C1 * X + (X + X) = (C1 + 2) * X
@@ -94,16 +197,72 @@ define float @fold7(float %f1) {
ret float %t3
}
-; (X + X) + (X + X) => 4.0 * X
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold7_reassoc_nsz(float %f1) {
+; CHECK-LABEL: @fold7_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc nsz float [[F1:%.*]], 7.000000e+00
+; CHECK-NEXT: ret float [[TMP1]]
+;
+ %t1 = fmul reassoc nsz float %f1, 5.000000e+00
+ %t2 = fadd reassoc nsz float %f1, %f1
+ %t3 = fadd reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to f1 * 7.0.
+define float @fold7_reassoc(float %f1) {
+; CHECK-LABEL: @fold7_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[F1:%.*]], 5.000000e+00
+; CHECK-NEXT: [[TMP2:%.*]] = fadd reassoc float [[F1]], [[F1]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fmul reassoc float %f1, 5.000000e+00
+ %t2 = fadd reassoc float %f1, %f1
+ %t3 = fadd reassoc float %t1, %t2
+ ret float %t3
+}
+
+; (X + X) + (X + X) + X => 5.0 * X
define float @fold8(float %f1) {
; CHECK-LABEL: @fold8(
-; CHECK-NEXT: [[TMP1:%.*]] = fmul fast float [[F1:%.*]], 4.000000e+00
+; CHECK-NEXT: [[TMP1:%.*]] = fmul fast float [[F1:%.*]], 5.000000e+00
; CHECK-NEXT: ret float [[TMP1]]
;
%t1 = fadd fast float %f1, %f1
%t2 = fadd fast float %f1, %f1
%t3 = fadd fast float %t1, %t2
- ret float %t3
+ %t4 = fadd fast float %t3, %f1
+ ret float %t4
+}
+
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold8_reassoc_nsz(float %f1) {
+; CHECK-LABEL: @fold8_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc nsz float [[F1:%.*]], 5.000000e+00
+; CHECK-NEXT: ret float [[TMP1]]
+;
+ %t1 = fadd reassoc nsz float %f1, %f1
+ %t2 = fadd reassoc nsz float %f1, %f1
+ %t3 = fadd reassoc nsz float %t1, %t2
+ %t4 = fadd reassoc nsz float %t3, %f1
+ ret float %t4
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to f1 * 5.0.
+define float @fold8_reassoc(float %f1) {
+; CHECK-LABEL: @fold8_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc float [[F1:%.*]], [[F1]]
+; CHECK-NEXT: [[TMP2:%.*]] = fadd reassoc float [[F1]], [[F1]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: [[TMP4:%.*]] = fadd reassoc float [[TMP3]], [[F1]]
+; CHECK-NEXT: ret float [[TMP4]]
+;
+ %t1 = fadd reassoc float %f1, %f1
+ %t2 = fadd reassoc float %f1, %f1
+ %t3 = fadd reassoc float %t1, %t2
+ %t4 = fadd reassoc float %t3, %f1
+ ret float %t4
}
; X - (X + Y) => 0 - Y
@@ -117,6 +276,29 @@ define float @fold9(float %f1, float %f2
ret float %t3
}
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @fold9_reassoc_nsz(float %f1, float %f2) {
+; CHECK-LABEL: @fold9_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float -0.000000e+00, [[F2:%.*]]
+; CHECK-NEXT: ret float [[TMP1]]
+;
+ %t1 = fadd float %f1, %f2
+ %t3 = fsub reassoc nsz float %f1, %t1
+ ret float %t3
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to 0 - f2
+define float @fold9_reassoc(float %f1, float %f2) {
+; CHECK-LABEL: @fold9_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd float [[F1:%.*]], [[F2:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fsub reassoc float [[F1]], [[TMP1]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fadd float %f1, %f2
+ %t3 = fsub reassoc float %f1, %t1
+ ret float %t3
+}
+
; Let C3 = C1 + C2. (f1 + C1) + (f2 + C2) => (f1 + f2) + C3 instead of
; "(f1 + C3) + f2" or "(f2 + C3) + f1". Placing constant-addend at the
; top of resulting simplified expression tree may potentially reveal some
@@ -134,6 +316,36 @@ define float @fold10(float %f1, float %f
ret float %t3
}
+; Check again with 'reassoc' and 'nsz'.
+; TODO: We may be able to remove the 'nsz' requirement.
+define float @fold10_reassoc_nsz(float %f1, float %f2) {
+; CHECK-LABEL: @fold10_reassoc_nsz(
+; CHECK-NEXT: [[T2:%.*]] = fadd reassoc nsz float [[F1:%.*]], [[F2:%.*]]
+; CHECK-NEXT: [[T3:%.*]] = fadd reassoc nsz float [[T2]], -1.000000e+00
+; CHECK-NEXT: ret float [[T3]]
+;
+ %t1 = fadd reassoc nsz float 2.000000e+00, %f1
+ %t2 = fsub reassoc nsz float %f2, 3.000000e+00
+ %t3 = fadd reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Observe that the fold is not done with only reassoc (the instructions are
+; canonicalized, but not folded).
+; TODO: As noted above, 'nsz' may not be required for this to be fully folded.
+define float @fold10_reassoc(float %f1, float %f2) {
+; CHECK-LABEL: @fold10_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc float [[F1:%.*]], 2.000000e+00
+; CHECK-NEXT: [[TMP2:%.*]] = fadd reassoc float [[F2:%.*]], -3.000000e+00
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fadd reassoc float 2.000000e+00, %f1
+ %t2 = fsub reassoc float %f2, 3.000000e+00
+ %t3 = fadd reassoc float %t1, %t2
+ ret float %t3
+}
+
; This used to crash/miscompile.
define float @fail1(float %f1, float %f2) {
@@ -171,40 +383,66 @@ define float @fold13(float %x) {
ret float %sub
}
+; Check again using the minimal subset of FMF.
+define float @fold13_reassoc_nsz(float %x) {
+; CHECK-LABEL: @fold13_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc nsz float [[X:%.*]], 6.000000e+00
+; CHECK-NEXT: ret float [[TMP1]]
+;
+ %mul = fmul reassoc nsz float %x, 7.000000e+00
+ %sub = fsub reassoc nsz float %mul, %x
+ ret float %sub
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fold13_reassoc(float %x) {
+; CHECK-LABEL: @fold13_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[X:%.*]], 7.000000e+00
+; CHECK-NEXT: [[TMP2:%.*]] = fsub reassoc float [[TMP1]], [[X]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %mul = fmul reassoc float %x, 7.000000e+00
+ %sub = fsub reassoc float %mul, %x
+ ret float %sub
+}
+
; -x + y => y - x
+; This is always safe. No FMF required.
define float @fold14(float %x, float %y) {
; CHECK-LABEL: @fold14(
-; CHECK-NEXT: [[ADD:%.*]] = fsub fast float [[Y:%.*]], [[X:%.*]]
+; CHECK-NEXT: [[ADD:%.*]] = fsub float [[Y:%.*]], [[X:%.*]]
; CHECK-NEXT: ret float [[ADD]]
;
- %neg = fsub fast float -0.0, %x
- %add = fadd fast float %neg, %y
+ %neg = fsub float -0.0, %x
+ %add = fadd float %neg, %y
ret float %add
}
; x + -y => x - y
+; This is always safe. No FMF required.
define float @fold15(float %x, float %y) {
; CHECK-LABEL: @fold15(
-; CHECK-NEXT: [[ADD:%.*]] = fsub fast float [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[ADD:%.*]] = fsub float [[X:%.*]], [[Y:%.*]]
; CHECK-NEXT: ret float [[ADD]]
;
- %neg = fsub fast float -0.0, %y
- %add = fadd fast float %x, %neg
+ %neg = fsub float -0.0, %y
+ %add = fadd float %x, %neg
ret float %add
}
; (select X+Y, X-Y) => X + (select Y, -Y)
+; This is always safe. No FMF required.
define float @fold16(float %x, float %y) {
; CHECK-LABEL: @fold16(
; CHECK-NEXT: [[CMP:%.*]] = fcmp ogt float [[X:%.*]], [[Y:%.*]]
-; CHECK-NEXT: [[TMP1:%.*]] = fsub fast float -0.000000e+00, [[Y]]
+; CHECK-NEXT: [[TMP1:%.*]] = fsub float -0.000000e+00, [[Y]]
; CHECK-NEXT: [[R_P:%.*]] = select i1 [[CMP]], float [[Y]], float [[TMP1]]
-; CHECK-NEXT: [[R:%.*]] = fadd fast float [[R_P]], [[X]]
+; CHECK-NEXT: [[R:%.*]] = fadd float [[R_P]], [[X]]
; CHECK-NEXT: ret float [[R]]
;
%cmp = fcmp ogt float %x, %y
- %plus = fadd fast float %x, %y
- %minus = fsub fast float %x, %y
+ %plus = fadd float %x, %y
+ %minus = fsub float %x, %y
%r = select i1 %cmp, float %plus, float %minus
ret float %r
}
@@ -331,6 +569,33 @@ define float @fact_mul1(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_mul1_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul1_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc nsz float [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc nsz float [[TMP1]], [[Z:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fmul reassoc nsz float %x, %z
+ %t2 = fmul reassoc nsz float %y, %z
+ %t3 = fadd reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_mul1_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul1_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[X:%.*]], [[Z:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc float [[Y:%.*]] [[Z]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fmul reassoc float %x, %z
+ %t2 = fmul reassoc float %y, %z
+ %t3 = fadd reassoc float %t1, %t2
+ ret float %t3
+}
+
; z*x + y*z => (x+y) * z
define float @fact_mul2(float %x, float %y, float %z) {
; CHECK-LABEL: @fact_mul2(
@@ -344,6 +609,33 @@ define float @fact_mul2(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_mul2_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul2_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc nsz float [[TMP1]], [[Z:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fmul reassoc nsz float %z, %x
+ %t2 = fmul reassoc nsz float %y, %z
+ %t3 = fsub reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_mul2_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul2_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[Z:%.*]], [[X:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc float [[Y:%.*]], [[Z]]
+; CHECK-NEXT: [[TMP3:%.*]] = fsub reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fmul reassoc float %z, %x
+ %t2 = fmul reassoc float %y, %z
+ %t3 = fsub reassoc float %t1, %t2
+ ret float %t3
+}
+
; z*x - z*y => (x-y) * z
define float @fact_mul3(float %x, float %y, float %z) {
; CHECK-LABEL: @fact_mul3(
@@ -357,6 +649,33 @@ define float @fact_mul3(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_mul3_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul3_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc nsz float [[TMP1]], [[Z:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t2 = fmul reassoc nsz float %z, %y
+ %t1 = fmul reassoc nsz float %z, %x
+ %t3 = fsub reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_mul3_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul3_reassoc(
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc float [[Z:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[Z]], [[X:%.*]]
+; CHECK-NEXT: [[TMP3:%.*]] = fsub reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t2 = fmul reassoc float %z, %y
+ %t1 = fmul reassoc float %z, %x
+ %t3 = fsub reassoc float %t1, %t2
+ ret float %t3
+}
+
; x*z - z*y => (x-y) * z
define float @fact_mul4(float %x, float %y, float %z) {
; CHECK-LABEL: @fact_mul4(
@@ -370,6 +689,33 @@ define float @fact_mul4(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_mul4_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul4_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float [[X:%.*]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc nsz float [[TMP1]], [[Z:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fmul reassoc nsz float %x, %z
+ %t2 = fmul reassoc nsz float %z, %y
+ %t3 = fsub reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_mul4_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_mul4_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fmul reassoc float [[X:%.*]], [[Z:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc float [[Z]], [[Y:%.*]]
+; CHECK-NEXT: [[TMP3:%.*]] = fsub reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fmul reassoc float %x, %z
+ %t2 = fmul reassoc float %z, %y
+ %t3 = fsub reassoc float %t1, %t2
+ ret float %t3
+}
+
; x/y + x/z, no xform
define float @fact_div1(float %x, float %y, float %z) {
; CHECK-LABEL: @fact_div1(
@@ -411,6 +757,33 @@ define float @fact_div3(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_div3_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_div3_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fadd reassoc nsz float [[Y:%.*]], [[Z:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fdiv reassoc nsz float [[TMP1]], [[X:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fdiv reassoc nsz float %y, %x
+ %t2 = fdiv reassoc nsz float %z, %x
+ %t3 = fadd reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_div3_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_div3_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fdiv reassoc float [[Y:%.*]], [[X:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fdiv reassoc float [[Z:%.*]], [[X]]
+; CHECK-NEXT: [[TMP3:%.*]] = fadd reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fdiv reassoc float %y, %x
+ %t2 = fdiv reassoc float %z, %x
+ %t3 = fadd reassoc float %t1, %t2
+ ret float %t3
+}
+
; y/x - z/x => (y-z)/x
define float @fact_div4(float %x, float %y, float %z) {
; CHECK-LABEL: @fact_div4(
@@ -424,6 +797,33 @@ define float @fact_div4(float %x, float
ret float %t3
}
+; Check again using the minimal subset of FMF.
+define float @fact_div4_reassoc_nsz(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_div4_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float [[Y:%.*]], [[Z:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fdiv reassoc nsz float [[TMP1]], [[X:%.*]]
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %t1 = fdiv reassoc nsz float %y, %x
+ %t2 = fdiv reassoc nsz float %z, %x
+ %t3 = fsub reassoc nsz float %t1, %t2
+ ret float %t3
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
+define float @fact_div4_reassoc(float %x, float %y, float %z) {
+; CHECK-LABEL: @fact_div4_reassoc(
+; CHECK-NEXT: [[TMP1:%.*]] = fdiv reassoc float [[Y:%.*]], [[X:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fdiv reassoc float [[Z:%.*]], [[X]]
+; CHECK-NEXT: [[TMP3:%.*]] = fsub reassoc float [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret float [[TMP3]]
+;
+ %t1 = fdiv reassoc float %y, %x
+ %t2 = fdiv reassoc float %z, %x
+ %t3 = fsub reassoc float %t1, %t2
+ ret float %t3
+}
+
; y/x - z/x => (y-z)/x is disabled if y-z is denormal.
define float @fact_div5(float %x) {
; CHECK-LABEL: @fact_div5(
@@ -735,12 +1135,14 @@ define fp128 @min4(fp128 %a, fp128 %b) {
ret fp128 %c
}
+; ((which ? 2.0 : a) + 1.0) => (which ? 3.0 : (a + 1.0))
+; This is always safe. No FMF required.
define float @test55(i1 %which, float %a) {
; CHECK-LABEL: @test55(
; CHECK-NEXT: entry:
; CHECK-NEXT: br i1 [[WHICH:%.*]], label [[FINAL:%.*]], label [[DELAY:%.*]]
; CHECK: delay:
-; CHECK-NEXT: [[PHITMP:%.*]] = fadd fast float [[A:%.*]], 1.000000e+00
+; CHECK-NEXT: [[PHITMP:%.*]] = fadd float [[A:%.*]], 1.000000e+00
; CHECK-NEXT: br label [[FINAL]]
; CHECK: final:
; CHECK-NEXT: [[A:%.*]] = phi float [ 3.000000e+00, [[ENTRY:%.*]] ], [ [[PHITMP]], [[DELAY]] ]
@@ -754,7 +1156,6 @@ delay:
final:
%A = phi float [ 2.0, %entry ], [ %a, %delay ]
- %value = fadd fast float %A, 1.0
+ %value = fadd float %A, 1.0
ret float %value
}
-
Modified: llvm/trunk/test/Transforms/Reassociate/fast-MissedTree.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Reassociate/fast-MissedTree.ll?rev=330089&r1=330088&r2=330089&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/Reassociate/fast-MissedTree.ll (original)
+++ llvm/trunk/test/Transforms/Reassociate/fast-MissedTree.ll Sat Apr 14 12:18:28 2018
@@ -12,7 +12,19 @@ define float @test1(float %A, float %B)
}
; Check again using minimal subset of FMF.
+; Both 'reassoc' and 'nsz' are required.
+define float @test1_reassoc_nsz(float %A, float %B) {
+; CHECK-LABEL: @test1_reassoc_nsz(
+; CHECK-NEXT: [[Z:%.*]] = fadd reassoc nsz float %A, %B
+; CHECK-NEXT: ret float [[Z]]
+;
+ %W = fadd reassoc nsz float %B, -5.0
+ %Y = fadd reassoc nsz float %A, 5.0
+ %Z = fadd reassoc nsz float %W, %Y
+ ret float %Z
+}
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
define float @test1_reassoc(float %A, float %B) {
; CHECK-LABEL: @test1_reassoc(
; CHECK-NEXT: [[W:%.*]] = fadd reassoc float %B, -5.000000e+00
@@ -25,4 +37,3 @@ define float @test1_reassoc(float %A, fl
%Z = fadd reassoc float %W, %Y
ret float %Z
}
-
Modified: llvm/trunk/test/Transforms/Reassociate/fast-SubReassociate.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Reassociate/fast-SubReassociate.ll?rev=330089&r1=330088&r2=330089&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/Reassociate/fast-SubReassociate.ll (original)
+++ llvm/trunk/test/Transforms/Reassociate/fast-SubReassociate.ll Sat Apr 14 12:18:28 2018
@@ -29,7 +29,20 @@ define float @test2(float %A, float %B)
}
; Check again using minimal subset of FMF.
+; Both 'reassoc' and 'nsz' are required.
+define float @test2_minimal(float %A, float %B) {
+; CHECK-LABEL: @test2_minimal(
+; CHECK-NEXT: [[Z:%.*]] = fsub reassoc nsz float %A, %B
+; CHECK-NEXT: ret float [[Z]]
+;
+ %W = fadd reassoc nsz float %B, 5.000000e+00
+ %X = fadd reassoc nsz float %A, -7.000000e+00
+ %Y = fsub reassoc nsz float %X, %W
+ %Z = fadd reassoc nsz float %Y, 1.200000e+01
+ ret float %Z
+}
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
define float @test2_reassoc(float %A, float %B) {
; CHECK-LABEL: @test2_reassoc(
; CHECK-NEXT: [[W:%.*]] = fadd reassoc float %B, 5.000000e+00
Modified: llvm/trunk/test/Transforms/Reassociate/fast-basictest.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Reassociate/fast-basictest.ll?rev=330089&r1=330088&r2=330089&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/Reassociate/fast-basictest.ll (original)
+++ llvm/trunk/test/Transforms/Reassociate/fast-basictest.ll Sat Apr 14 12:18:28 2018
@@ -12,6 +12,19 @@ define float @test1(float %arg) {
ret float %t2
}
+; Check again using the minimal subset of FMF.
+; Both 'reassoc' and 'nsz' are required.
+define float @test1_minimal(float %arg) {
+; CHECK-LABEL: @test1_minimal(
+; CHECK-NEXT: [[ARG_NEG:%.*]] = fsub reassoc nsz float -0.000000e+00, [[ARG:%.*]]
+; CHECK-NEXT: ret float [[ARG_NEG]]
+;
+ %t1 = fsub reassoc nsz float -1.200000e+01, %arg
+ %t2 = fadd reassoc nsz float %t1, 1.200000e+01
+ ret float %t2
+}
+
+; Verify the fold is not done with only 'reassoc' ('nsz' is required).
define float @test1_reassoc(float %arg) {
; CHECK-LABEL: @test1_reassoc(
; CHECK-NEXT: [[T1:%.*]] = fsub reassoc float -1.200000e+01, [[ARG:%.*]]
@@ -210,6 +223,18 @@ define float @test9(float %X) {
ret float %Z
}
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @test9_reassoc_nsz(float %X) {
+; CHECK-LABEL: @test9_reassoc_nsz(
+; CHECK-NEXT: [[FACTOR:%.*]] = fmul reassoc nsz float [[X:%.*]], 9.400000e+01
+; CHECK-NEXT: ret float [[FACTOR]]
+;
+ %Y = fmul reassoc nsz float %X, 4.700000e+01
+ %Z = fadd reassoc nsz float %Y, %Y
+ ret float %Z
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to X * 94.0
define float @test9_reassoc(float %X) {
; CHECK-LABEL: @test9_reassoc(
; CHECK-NEXT: [[Y:%.*]] = fmul reassoc float [[X:%.*]], 4.700000e+01
@@ -221,25 +246,44 @@ define float @test9_reassoc(float %X) {
ret float %Z
}
+; Side note: (x + x + x) and (3*x) each have only a single rounding. So
+; transforming x+x+x to 3*x is always safe, even without any FMF.
+; To avoid that special-case, we have the addition of 'x' four times, here.
define float @test10(float %X) {
; CHECK-LABEL: @test10(
-; CHECK-NEXT: [[FACTOR:%.*]] = fmul fast float [[X:%.*]], 3.000000e+00
+; CHECK-NEXT: [[FACTOR:%.*]] = fmul fast float [[X:%.*]], 4.000000e+00
; CHECK-NEXT: ret float [[FACTOR]]
;
%Y = fadd fast float %X ,%X
%Z = fadd fast float %Y, %X
- ret float %Z
+ %W = fadd fast float %Z, %X
+ ret float %W
+}
+
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @test10_reassoc_nsz(float %X) {
+; CHECK-LABEL: @test10_reassoc_nsz(
+; CHECK-NEXT: [[FACTOR:%.*]] = fmul reassoc nsz float [[X:%.*]], 4.000000e+00
+; CHECK-NEXT: ret float [[FACTOR]]
+;
+ %Y = fadd reassoc nsz float %X ,%X
+ %Z = fadd reassoc nsz float %Y, %X
+ %W = fadd reassoc nsz float %Z, %X
+ ret float %W
}
+; TODO: This doesn't require 'nsz'. It should fold to 4 * x
define float @test10_reassoc(float %X) {
; CHECK-LABEL: @test10_reassoc(
; CHECK-NEXT: [[Y:%.*]] = fadd reassoc float [[X:%.*]], [[X]]
; CHECK-NEXT: [[Z:%.*]] = fadd reassoc float [[Y]], [[X]]
-; CHECK-NEXT: ret float [[Z]]
+; CHECK-NEXT: [[W:%.*]] = fadd reassoc float [[Z]], [[X]]
+; CHECK-NEXT: ret float [[W]]
;
%Y = fadd reassoc float %X ,%X
%Z = fadd reassoc float %Y, %X
- ret float %Z
+ %W = fadd reassoc float %Z, %X
+ ret float %W
}
define float @test11(float %W) {
@@ -253,6 +297,20 @@ define float @test11(float %W) {
ret float %Z
}
+; Check again using the minimal subset of FMF.
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @test11_reassoc_nsz(float %W) {
+; CHECK-LABEL: @test11_reassoc_nsz(
+; CHECK-NEXT: [[FACTOR:%.*]] = fmul reassoc nsz float [[W:%.*]], 3.810000e+02
+; CHECK-NEXT: ret float [[FACTOR]]
+;
+ %X = fmul reassoc nsz float %W, 127.0
+ %Y = fadd reassoc nsz float %X ,%X
+ %Z = fadd reassoc nsz float %Y, %X
+ ret float %Z
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to W*381.0.
define float @test11_reassoc(float %W) {
; CHECK-LABEL: @test11_reassoc(
; CHECK-NEXT: [[X:%.*]] = fmul reassoc float [[W:%.*]], 1.270000e+02
@@ -280,6 +338,22 @@ define float @test12(float %X) {
ret float %Z
}
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @test12_reassoc_nsz(float %X) {
+; CHECK-LABEL: @test12_reassoc_nsz(
+; CHECK-NEXT: [[FACTOR:%.*]] = fmul reassoc nsz float [[X:%.*]], 3.000000e+00
+; CHECK-NEXT: [[Z:%.*]] = fsub reassoc nsz float 6.000000e+00, [[FACTOR]]
+; CHECK-NEXT: ret float [[Z]]
+;
+ %A = fsub reassoc nsz float 1.000000e+00, %X
+ %B = fsub reassoc nsz float 2.000000e+00, %X
+ %C = fsub reassoc nsz float 3.000000e+00, %X
+ %Y = fadd reassoc nsz float %A ,%B
+ %Z = fadd reassoc nsz float %Y, %C
+ ret float %Z
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to (6.0 - 3.0*x)
define float @test12_reassoc(float %X) {
; CHECK-LABEL: @test12_reassoc(
; CHECK-NEXT: [[A:%.*]] = fsub reassoc float 1.000000e+00, [[X:%.*]]
@@ -337,6 +411,21 @@ define float @test14(float %X1, float %X
ret float %D
}
+; (x1 * 47) + (x2 * -47) => (x1 - x2) * 47
+; Check again with 'reassoc' and 'nsz' ('nsz' not technically required).
+define float @test14_reassoc_nsz(float %X1, float %X2) {
+; CHECK-LABEL: @test14_reassoc_nsz(
+; CHECK-NEXT: [[TMP1:%.*]] = fsub reassoc nsz float [[X1:%.*]], [[X2:%.*]]
+; CHECK-NEXT: [[TMP2:%.*]] = fmul reassoc nsz float [[TMP1]], 4.700000e+01
+; CHECK-NEXT: ret float [[TMP2]]
+;
+ %B = fmul reassoc nsz float %X1, 47. ; X1*47
+ %C = fmul reassoc nsz float %X2, -47. ; X2*-47
+ %D = fadd reassoc nsz float %B, %C ; X1*47 + X2*-47 -> 47*(X1-X2)
+ ret float %D
+}
+
+; TODO: This doesn't require 'nsz'. It should fold to ((x1 - x2) * 47.0)
define float @test14_reassoc(float %X1, float %X2) {
; CHECK-LABEL: @test14_reassoc(
; CHECK-NEXT: [[B:%.*]] = fmul reassoc float [[X1:%.*]], 4.700000e+01
More information about the llvm-commits
mailing list