♟ qiucf

I'm identical to @ecnelises .

[SelectionDAG] Add guard to automatically insert flags

In D88063#2287835, @spatel wrote:

What, if any, interaction is there between this and D87361?
Can we solve this with the more general flag propagation scheme proposed there?

[PowerPC] Pass nofpexcept flag to custom lowered constrained ops

The differential revision should be https://reviews.llvm.org/D87698

[SelectionDAG] Check any use of negation result before removal

[PowerPC] Fix store-fptoi combine of f128 on Power8

Remove flag passing in methods visiting floating-point nodes.

Revert "[SelectionDAG] Remove unused FP constant in getNegatedExpression"

After running sanitizer, I think the bug is not totally covered by this patch.

[SelectionDAG] Remove unused FP constant in getNegatedExpression

Update test using script

Keep flags in current methods and introduce new methods to auto-insert flags. (maybe can use variadic template methods?)

[DAGCombiner] Propagate FMF flags in FMA folding

Reland "[PowerPC] Implement instruction clustering for stores"

[FPEnv] [Clang] Enable constrained FP support for PowerPC

In D87223#2261086, @kpn wrote:

Over in D80952 there was a comment from @steven.zhang: "Agree that we can turn on it later when it is done." But your comment here is that it is "almost done".

Is it fair to say that the PowerPC work is at _parity_ with X86 and SystemZ? I think that's a better cutover point than "done".

[PowerPC] [FPEnv] Disable strict FP mutation by default

[PowerPC] Fix STRICT_FRINT/STRICT_FNEARBYINT lowering

Revert "[PowerPC] Implement instruction clustering for stores"

[PowerPC] Fix getMemOperandWithOffsetWidth

[PowerPC] Implement instruction clustering for stores

Fix clang to align the change

Update tests
Keep redundant code (irrelevant to this patch)

[PowerPC] Expand constrained ppc_fp128 to i32 conversion

Only pass nofpexcept flag.

Move out constants
Transfer nofpexcept flag. If this is a problem, other lowering methods should also be fixed
Transfer nofpexcept to FADD if FADDRtz is nofpexcept

Introduce strict version of FADDRTZ
Use algorithm similar to expandFP_TO_UINT

[NFC] [PowerPC] Add FMA flag propagation test

define i32 @test(double %a, double %b) {
entry:
  %r = fcmp ogt double %a, %b
  %g = zext i1 %r to i32
  ret i32 %g
}

[PowerPC] Handle STRICT_FSETCC(S) in more cases

Fix chain of fcmp instruction

[NFC] [DAGCombiner] Refactor bitcast folding within fabs/fneg

[NFC] [DAGCombiner] Remove unnecessary negation in visitFNEG

Address style comments.

Use FusionFeatures.

Is this still relevant?

[PowerPC] Support lowering int-to-fp on ppc_fp128

[PowerPC] Allow constrained FP intrinsics in mightUseCTR

[PowerPC] Support constrained vector fp/int conversion

[PowerPC] Support constrained scalar sitofp/uitofp

use script to refresh test

[PowerPC] Add readflm/setflm intrinsics to Clang

Use PPCany_* to restructure tblgen file
Add mayRaiseFPException flag to fcfid*
Add default argument to convertIntToFp

The original test case is too big to get something useful. I happened to grab a basic block from it. It's not small enough to be a test, but it would be much easier to analyze: https://reviews.llvm.org/P8230

qiucf (Qiu Chaofan)
User

Projects

User Details

Recent Activity
View All

Fri, Sep 25

Tue, Sep 22

Sun, Sep 20

Thu, Sep 17

Wed, Sep 16

Tue, Sep 15

Mon, Sep 14

Sun, Sep 13

Fri, Sep 11

Thu, Sep 10

Wed, Sep 9

Tue, Sep 8

Mon, Sep 7

Sun, Sep 6

Fri, Sep 4

Wed, Sep 2

Tue, Sep 1

Mon, Aug 31

Sat, Aug 29

Aug 27 2020

Aug 26 2020

Aug 23 2020

Aug 21 2020

Aug 20 2020

qiucf (Qiu Chaofan)User