10 Commits

Author SHA1 Message Date
Matt Arsenault
3f62718c4a
AMDGPU: Migrate some tests away from undef (#131277) 2025-03-15 00:29:10 +07:00
Matt Arsenault
ad993687b1
AMDGPU: Replace ptr addrspace(4) undef uses with poison in tests (#131095) 2025-03-14 09:47:54 +07:00
Matt Arsenault
6705d812b8
AMDGPU: Replace ptr addrspace(1) undefs with poison (#130900)
Many tests use store to undef as a placeholder use, so just replace
all of these with poison.
2025-03-13 08:25:02 +07:00
Matt Arsenault
da42b2f67d
AMDGPU: Replace insertelement poison with insertelement undef (#130896)
This is the bulk update with perl, with cases which require additional
update left for later.
2025-03-12 20:33:33 +07:00
Matt Arsenault
36a0f20ac3
AMDGPU/NewPM: Fill out addPreISelPasses (#102814)
This specific callback should now be at parity with the old
pass manager version. There are still some missing IR passes
before this point.

Also I don't understand the need for the RequiresAnalysisPass at the
end. SelectionDAG should just be using the uncached getResult?
2024-08-14 20:57:00 +04:00
paperchalice
7652a59407
Reland "[NewPM][CodeGen] Port selection dag isel to new pass manager" (#94149)
- Fix build with `EXPENSIVE_CHECKS`
- Remove unused `PassName::ID` to resolve warning
- Mark `~SelectionDAGISel` virtual so AArch64 backend can work properly
2024-06-04 08:10:58 +08:00
paperchalice
8917afaf0e
Revert "[NewPM][CodeGen] Port selection dag isel to new pass manager" (#94146)
This reverts commit de37c06f01772e02465ccc9f538894c76d89a7a1 to
de37c06f01772e02465ccc9f538894c76d89a7a1

It still breaks EXPENSIVE_CHECKS build. Sorry.
2024-06-02 14:31:52 +08:00
paperchalice
d2cdc8ab45
[NewPM][CodeGen] Port selection dag isel to new pass manager (#83567)
Port selection dag isel to new pass manager.
Only `AMDGPU` and `X86` support new pass version. `-verify-machineinstrs` in new pass manager belongs to verify instrumentation, it is enabled by default.
2024-06-02 09:12:33 +08:00
Nikita Popov
bdf2fbba9c [AMDGPU] Convert some tests to opaque pointers (NFC) 2022-12-19 12:41:13 +01:00
Carl Ritson
fbaa35e169 [AMDGPU] Add SelectionDAG support for insert_subvector on v4f64
Enable custom insert_subvector for larger vector types.
This is necessary now that SelectionDAG can attempt v3f64 insert
to v4f64, etc.

Reviewed By: foad

Differential Revision: https://reviews.llvm.org/D105385
2021-07-27 10:11:34 +09:00