fix(vpto): lower MTE store atomics through CTRL - #569
Open
peanutchan wants to merge 2 commits into
Open
Conversation
Add the minimal PTODSL and VPTO lowering needed for TileLang UB-to-GM atomic add, max, and min copies. Co-authored-by: Cursor <cursoragent@cursor.com>
Expose CCE-style atomic configuration operations and preserve ordered CTRL reads so dtype and operation state reaches asynchronous MTE stores. Co-authored-by: Cursor <cursoragent@cursor.com>
peanutchan
changed the title
5 tasks
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
pto.set_atomic_{dtype,op,none}ops in the explicit PTODSL surfaceget_ctrl/set_ctrlupdates that match the existing L0C→GM FIXP atomic encoding (CTRL[8:6]dtype,CTRL[10:9]op)get_ctrlnon-pure so CSE cannot reuse a stale CTRL across interveningset_ctrlpto.set_store_atomic_cfgfor the scalarST_ATOMIC_CFGpath; MTE UB→GM atomics do not use itWhy
TileLang PR #75 now emits
pto.set_atomic_*around UB→GM copies. Directllvm.hivm.SET.ATOMIC.*is not selectable by bisheng; CCEset_atomic_*lowers to CTRL bit updates, which is also how L0C→GM FIXP atomics already work in PTOAS.Verified on camodel opsim for int32/float32
add/max/min/none.Test plan
python ptodsl/tests/test_jit_compile.pyptoas --pto-arch=a5 --pto-backend=vpto --emit-vpto --mlir-print-ir-after=vpto-expand-wrapper-ops test/lit/vpto/store_atomic_ctrl_expand.ptoptoas_runtime_stagingand Python bindings