arch-amdgpu: support ACC-selected DS data operands - #7
Open
Basemism wants to merge 1 commit into
Open
Conversation
Decode the CDNA DS ACC bit and map DS data operands, but not address operands, through the AGPR window of the unified vector register file. Apply the same accumulator offset during operand mapping and instruction execution for 32-, 64-, 96-, and 128-bit DS reads and writes. Also accept all DS primary-encoding variants used by the ACC form and check that mapped AGPR operands remain inside the wavefront's reserved vector register allocation.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add support for the CDNA DS ACC encoding, which selects accumulator registers for DS instruction data operands.
Before, DS instructions emitted with ACC=1 are either rejected by the decoder or incorrectly read and write ordinary VGPRs instead of the wavefront’s AGPR allocation.
The DS encoding now decodes the former reserved bit as ACC.
When ACC=1:
Applying the offset in both paths ensures that dependency tracking and the instruction data path refer to the same physical registers.
The change covers 32-, 64-, 96-, and 128-bit DS reads and writes, including multi-data-register instructions.