llvm-project.git/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp, branch main Unnamed repository; edit this file 'description' to name the repository. Revert "[AMDGPU] Remove leftover implicit operands from SI_SPILL/SI_RESTORE." (#169068) 2025-11-21T17:52:08+00:00 Nathan Corbyn n_corbyn@apple.com 2025-11-21T17:52:08+00:00 4511c355c35153c6b8f5fd3d0b75f77c126fe8e6 PR causes build failures with expensive checks enabled Reverts llvm/llvm-project#168546 
PR causes build failures with expensive checks enabled

Reverts llvm/llvm-project#168546
 AMDGPU: Don't duplicate implicit operands in 3-address conversion (#168426) 2025-11-21T00:25:47+00:00 Nicolai Hähnle nicolai.haehnle@amd.com 2025-11-21T00:25:47+00:00 ac55d7859fc53d9ca8444a29a93fc45d16ecb26a We previously got a duplicate implicit $exec operand. It didn't really hurt anything (other than being a slight drag on compile-time performance). Still, let's keep things clean. 
We previously got a duplicate implicit $exec operand. It didn't really
hurt anything (other than being a slight drag on compile-time
performance). Still, let's keep things clean.
 [AMDGPU] Remove leftover implicit operands from SI_SPILL/SI_RESTORE. (#168546) 2025-11-19T15:02:03+00:00 LU-JOHN John.Lu@amd.com 2025-11-19T15:02:03+00:00 b79a665f7170fbb631b13175ec747ccfd779bf9e Remove leftover implicit operands from SI_SPILL/SI_RESTORE. --------- Signed-off-by: John Lu <John.Lu@amd.com> 
Remove leftover implicit operands from SI_SPILL/SI_RESTORE.

---------

Signed-off-by: John Lu <John.Lu@amd.com>
 [AMDGPU] When shrinking and/or to bitset*, remove implicit scc def (#168128) 2025-11-15T15:21:43+00:00 LU-JOHN John.Lu@amd.com 2025-11-15T15:21:43+00:00 9fa15ef91601a2b254ee3c52dab03c1c33eb48d5 When shrinking and/or to bitset* remove leftover implicit scc def. bitset* instructions do not set scc. Signed-off-by: John Lu <John.Lu@amd.com> 
When shrinking and/or to bitset* remove leftover implicit scc def.
bitset* instructions do not set scc.

Signed-off-by: John Lu <John.Lu@amd.com>
 AMDGPU: Fix verifier error when waterfall call target is in AV register (#168017) 2025-11-14T17:49:40+00:00 Matt Arsenault Matthew.Arsenault@amd.com 2025-11-14T17:49:40+00:00 b2f12331abeb4a3c42309c5f49273e14c87d31b5 






[AMDGPU] Make use of getFunction and getMF. NFC. (#167872)
2025-11-14T11:00:57+00:00

Jay Foad
jay.foad@amd.com

2025-11-14T11:00:57+00:00

72c69aefbae8bfb087622e642acbd0cba7578747












[AMDGPU] Lower S_ABSDIFF_I32 to VALU instructions (#167691)
2025-11-13T13:35:44+00:00

Mariusz Sikora
mariusz.sikora@amd.com

2025-11-13T13:35:44+00:00

4cd836181f7923e7c51c5ff30eb1b2ca0479b645

Added support for lowering the scalar S_ABSDIFF_I32 instruction to
equivalent VALU operations.




Added support for lowering the scalar S_ABSDIFF_I32 instruction to
equivalent VALU operations.






CodeGen/AMDGPU: Allow 3-address conversion of bundled instructions (#166213)
2025-11-12T22:04:46+00:00

Nicolai Hähnle
nicolai.haehnle@amd.com

2025-11-12T22:04:46+00:00

66366599a9dda37aed9a54c855c5fd7f402ce6eb

This is in preparation for future changes in AMDGPU that will make more
substantial use of bundles pre-RA. For now, simply test this with
degenerate (single-instruction) bundles.




This is in preparation for future changes in AMDGPU that will make more
substantial use of bundles pre-RA. For now, simply test this with
degenerate (single-instruction) bundles.






AMDGPU: Remove override of TargetInstrInfo::getRegClass (#159886)
2025-11-12T00:05:05+00:00

Matt Arsenault
Matthew.Arsenault@amd.com

2025-11-12T00:05:05+00:00

fcba3040107944604904aeb146c26ec0628160f4

This should not be overridable and the special case hacks
have been replaced with RegClassByHwMode




This should not be overridable and the special case hacks
have been replaced with RegClassByHwMode






AMDGPU: Remove wrapper around TRI::getRegClass (#159885)
2025-11-11T23:31:52+00:00

Matt Arsenault
Matthew.Arsenault@amd.com

2025-11-11T23:31:52+00:00

e3a9ac5e24d08cef3160fe3e242a4afe1b6d95a4

This shadows the member in the base class, but differs slightly
in behavior. The base method doesn't check for the invalid case.




This shadows the member in the base class, but differs slightly
in behavior. The base method doesn't check for the invalid case.