llvm-project.git/llvm/lib/Target/AMDGPU/SILowerControlFlow.cpp, branch users/nico/python-2

[AMDGPU][NewPM] Port SILowerControlFlow pass into NPM. (#123045)

2025-01-16T05:36:38+00:00

[AMDGPU] Use LV wrapperPass in getAnalysisUsage. (#123044)

2025-01-16T05:33:14+00:00

[AMDGPU][NFC] Correct description (#112847)

2024-10-18T08:41:16+00:00

[AMDGPU][NFC] Update comment referring to SIRemoveShortExecBranches pass (#109756)

2024-09-24T07:46:39+00:00

That pass no longer exists, since
5df2af8b0ef33f48b1ee72bcd27bc609b898da52 has merged it into
SIPreEmitPeephole.

[CodeGen][NewPM] Port `LiveIntervals` to new pass manager (#98118)

2024-07-10T11:34:48+00:00

- Add `LiveIntervalsAnalysis`.
- Add `LiveIntervalsPrinterPass`.
- Use `LiveIntervalsWrapperPass` in legacy pass manager.
- Use `std::unique_ptr` instead of raw pointer for `LICalc`, so
destructor and default move constructor can handle it correctly.

This would be the last analysis required by `PHIElimination`.

[CodeGen][NewPM] Port `SlotIndexes` to new pass manager (#97941)

2024-07-09T04:09:11+00:00

- Add `SlotIndexesAnalysis`.
- Add `SlotIndexesPrinterPass`.
- Use `SlotIndexesWrapperPass` in legacy pass.

[CodeGen][NewPM] Port `LiveVariables` to new pass manager (#97880)

2024-07-09T02:50:43+00:00

- Port `LiveVariables` to new pass manager.
- Convert to `LiveVariablesWrapperPass` in legacy pass manager.

[CodeGen][NewPM] Split `MachineDominatorTree` into a concrete analysis result (#94571)

2024-06-11T13:27:14+00:00

Prepare for new pass manager version of `MachineDominatorTreeAnalysis`.
We may need a machine dominator tree version of `DomTreeUpdater` to
handle `SplitCriticalEdge` in some CodeGen passes.

[AMDGPU] Move INIT_EXEC lowering from SILowerControlFlow to SIWholeQuadMode (#94452)

2024-06-06T09:29:55+00:00

NFCI; this just preserves SI_INIT_EXEC and SI_INIT_EXEC_FROM_INPUT
instructions a little longer so that we can reliably identify them in
SIWholeQuadMode.

[NFC][CodeGen] Change CodeGenOpt::Level/CodeGenFileType into enum classes (#66295)

2023-09-14T21:10:14+00:00

This will make it easy for callers to see issues with and fix up calls
to createTargetMachine after a future change to the params of
TargetMachine.

This matches other nearby enums.

For downstream users, this should be a fairly straightforward
replacement,
e.g. s/CodeGenOpt::Aggressive/CodeGenOptLevel::Aggressive
or s/CGFT_/CodeGenFileType::