RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-19 02:03:11 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

[SCEV] Make howFarToZero use a simpler formula for max backedge-taken count.

Browse Source 
This is both easier to understand, and produces a tighter bound in certain
cases.

Differential Revision: https://reviews.llvm.org/D28393

llvm-svn: 291701
This commit is contained in:
Eli Friedman 2017-01-11 20:55:48 +00:00
parent 8069f39125
commit 9136368f80
2 changed files with 85 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
lib/Analysis/ScalarEvolution.cpp
View File

@ -7206,17 +7206,8 @@ ScalarEvolution::howFarToZero(const SCEV *V, const Loop *L, bool ControlsExit,
// 1*N = -Start; -1*N = Start (mod 2^BW), so:
// N = Distance (as unsigned)
if (StepC->getValue()->equalsInt(1) || StepC->getValue()->isAllOnesValue()) {
ConstantRange CR = getUnsignedRange(Start);
const SCEV *MaxBECount;
if (!CountDown && CR.getUnsignedMin().isMinValue())
// When counting up, the worst starting value is 1, not 0.
MaxBECount = CR.getUnsignedMax().isMinValue()
? getConstant(APInt::getMinValue(CR.getBitWidth()))
: getConstant(APInt::getMaxValue(CR.getBitWidth()));
else
MaxBECount = getConstant(CountDown ? CR.getUnsignedMax()
: -CR.getUnsignedMin());
return ExitLimit(Distance, MaxBECount, false, Predicates);
APInt MaxBECount = getUnsignedRange(Distance).getUnsignedMax();
return ExitLimit(Distance, getConstant(MaxBECount), false, Predicates);
}
// As a special case, handle the instance where Step is a positive power of

83
test/Analysis/ScalarEvolution/max-trip-count.ll
View File

@ -207,3 +207,86 @@ for.cond.i: ; preds = %for.body.i
bar.exit: ; preds = %for.cond.i, %for.body.i
ret i32 0
}
; CHECK-LABEL: @ne_max_trip_count_1
; CHECK: Loop %for.body: max backedge-taken count is 7
define i32 @ne_max_trip_count_1(i32 %n) {
entry:
%masked = and i32 %n, 7
br label %for.body
for.body:
%i = phi i32 [ 0, %entry ], [ %add, %for.body ]
%add = add nsw i32 %i, 1
%cmp = icmp ne i32 %i, %masked
br i1 %cmp, label %for.body, label %bar.exit
bar.exit:
ret i32 0
}
; CHECK-LABEL: @ne_max_trip_count_2
; CHECK: Loop %for.body: max backedge-taken count is -1
define i32 @ne_max_trip_count_2(i32 %n) {
entry:
%masked = and i32 %n, 7
br label %for.body
for.body:
%i = phi i32 [ 0, %entry ], [ %add, %for.body ]
%add = add nsw i32 %i, 1
%cmp = icmp ne i32 %add, %masked
br i1 %cmp, label %for.body, label %bar.exit
bar.exit:
ret i32 0
}
; TODO: Improve count based on guard.
; CHECK-LABEL: @ne_max_trip_count_3
; CHECK: Loop %for.body: max backedge-taken count is -1
define i32 @ne_max_trip_count_3(i32 %n) {
entry:
%masked = and i32 %n, 7
%guard = icmp eq i32 %masked, 0
br i1 %guard, label %exit, label %for.preheader
for.preheader:
br label %for.body
for.body:
%i = phi i32 [ 0, %for.preheader ], [ %add, %for.body ]
%add = add nsw i32 %i, 1
%cmp = icmp ne i32 %add, %masked
br i1 %cmp, label %for.body, label %loop.exit
loop.exit:
br label %exit
exit:
ret i32 0
}
; TODO: Improve count based on guard.
; CHECK-LABEL: @ne_max_trip_count_4
; CHECK: Loop %for.body: max backedge-taken count is -1
define i32 @ne_max_trip_count_4(i32 %n) {
entry:
%guard = icmp eq i32 %n, 0
br i1 %guard, label %exit, label %for.preheader
for.preheader:
br label %for.body
for.body:
%i = phi i32 [ 0, %for.preheader ], [ %add, %for.body ]
%add = add nsw i32 %i, 1
%cmp = icmp ne i32 %add, %n
br i1 %cmp, label %for.body, label %loop.exit
loop.exit:
br label %exit
exit:
ret i32 0
}