maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
282,915 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

3 Commits

Author SHA1 Message Date
Sanjay Patel 7ac2db6a48 [InstCombine] improve folds for icmp gt/lt (shr X, C1), C2 
We can always eliminate the shift in: icmp gt/lt (shr X, C1), C2 --> icmp gt/lt X, C'
This patch was supposed to just be an efficiency improvement because we were doing this 3-step process to fold:

IC: Visiting:   %c = icmp ugt i4 %s, 1
IC: ADD:   %s = lshr i4 %x, 1
IC: ADD:   %1 = udiv i4 %x, 2
IC: Old =   %c = icmp ugt i4 %1, 1
    New =   <badref> = icmp uge i4 %x, 4
IC: ADD:   %c = icmp uge i4 %x, 4
IC: ERASE   %2 = icmp ugt i4 %1, 1
IC: Visiting:   %c = icmp uge i4 %x, 4
IC: Old =   %c = icmp uge i4 %x, 4
    New =   <badref> = icmp ugt i4 %x, 3
IC: ADD:   %c = icmp ugt i4 %x, 3
IC: ERASE   %2 = icmp uge i4 %x, 4
IC: Visiting:   %c = icmp ugt i4 %x, 3
IC: DCE:   %1 = udiv i4 %x, 2
IC: ERASE   %1 = udiv i4 %x, 2
IC: DCE:   %s = lshr i4 %x, 1
IC: ERASE   %s = lshr i4 %x, 1
IC: Visiting:   ret i1 %c

When we could go directly to canonical icmp form:

IC: Visiting:   %c = icmp ugt i4 %s, 1
IC: Old =   %c = icmp ugt i4 %s, 1
    New =   <badref> = icmp ugt i4 %x, 3
IC: ADD:   %c = icmp ugt i4 %x, 3
IC: ERASE   %1 = icmp ugt i4 %s, 1
IC: ADD:   %s = lshr i4 %x, 1
IC: DCE:   %s = lshr i4 %x, 1
IC: ERASE   %s = lshr i4 %x, 1
IC: Visiting:   %c = icmp ugt i4 %x, 3

...but then I noticed that the folds were incomplete too:
https://godbolt.org/g/aB2hLE

Here are attempts to prove the logic with Alive:
https://rise4fun.com/Alive/92o

Name: lshr_ult
Pre: ((C2 << C1) u>> C1) == C2
%sh = lshr i8 %x, C1
%r = icmp ult i8 %sh, C2
  =>
%r = icmp ult i8 %x, (C2 << C1)

Name: ashr_slt
Pre: ((C2 << C1) >> C1) == C2
%sh = ashr i8 %x, C1
%r = icmp slt i8 %sh, C2
  =>
%r = icmp slt i8 %x, (C2 << C1)

Name: lshr_ugt
Pre: (((C2+1) << C1) u>> C1) == (C2+1)
%sh = lshr i8 %x, C1
%r = icmp ugt i8 %sh, C2
  =>
%r = icmp ugt i8 %x, ((C2+1) << C1) - 1

Name: ashr_sgt
Pre: (C2 != 127) && ((C2+1) << C1 != -128) && (((C2+1) << C1) >> C1) == (C2+1)
%sh = ashr i8 %x, C1
%r = icmp sgt i8 %sh, C2
  =>
%r = icmp sgt i8 %x, ((C2+1) << C1) - 1

Name: ashr_exact_sgt
Pre: ((C2 << C1) >> C1) == C2
%sh = ashr exact i8 %x, C1
%r = icmp sgt i8 %sh, C2
  =>
%r = icmp sgt i8 %x, (C2 << C1)

Name: ashr_exact_slt
Pre: ((C2 << C1) >> C1) == C2
%sh = ashr exact i8 %x, C1
%r = icmp slt i8 %sh, C2
  =>
%r = icmp slt i8 %x, (C2 << C1)

Name: lshr_exact_ugt
Pre: ((C2 << C1) u>> C1) == C2
%sh = lshr exact i8 %x, C1
%r = icmp ugt i8 %sh, C2
  =>
%r = icmp ugt i8 %x, (C2 << C1)

Name: lshr_exact_ult
Pre: ((C2 << C1) u>> C1) == C2
%sh = lshr exact i8 %x, C1
%r = icmp ult i8 %sh, C2
  =>
%r = icmp ult i8 %x, (C2 << C1)

We did something similar for 'shl' in D28406.

Differential Revision: https://reviews.llvm.org/D38514

llvm-svn: 315021
 2017-10-05 21:11:49 +00:00
Sanjay Patel 9366b9c53b [InstCombine] add 'exact' variants of all tests; NFC 
We can likely remove most of these as redundant in the near future, 
but I'm trying to make sure I don't introduce any regressions with D38514.

llvm-svn: 314907
 2017-10-04 15:17:25 +00:00
Sanjay Patel 389b7cedc3 [InstCombine] add tests for icmp gt/lt (shr X, C1), C2; NFC 
Surprisingly, we have zero coverage for these patterns.

Many of these are handled in InstSimplify, but it's not obvious
what the rule for folding each case should be, so I've just
stamped out everything.

It should be possible to fold every case, but currently, we
miss these:

int ashr_slt(int x) {
  return (x >> 1) < 1; 
}

int ashr_sgt(int x) {
  return (x >> 1) > 0; 
}

https://godbolt.org/g/aB2hLE

llvm-svn: 314837
 2017-10-03 20:34:20 +00:00