Fix: drop the stale N_UNROLL value from the unroll orchestration headers - #1572
Conversation
Four copies of the unroll paged-attention orchestration open with
"Paged Attention Orchestration Function V2 - N_UNROLL=8, 4 Tasks Per Group"
while `#define N_UNROLL 64` sits twenty lines below. The comment has been
wrong since the constant was raised, and it is the first line a reader sees.
The value is dropped rather than corrected to 64. The line below already says
"Batches up to N_UNROLL blocks per group" without a number, and the `#define`
is the only place the value belongs — restating it in prose is what let the
two drift apart.
All four copies land together: `examples/{a2a3,a5}/.../
paged_attention_unroll_manual_scope/` and `tests/st/{a2a3,a5}/.../
paged_attention_unroll/`. `grep -rn "N_UNROLL=8"` now returns nothing.
The `paged_attention_unroll_manual_scope` README noted the discrepancy; that
sentence goes with it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (5)
📝 WalkthroughWalkthroughThe changes align paged attention orchestration comments and README text with the current unroll configuration. Executable code, interfaces, control flow, and runtime behavior remain unchanged. ChangesPaged attention documentation alignment
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Summary
Four copies of the unroll paged-attention orchestration open with
while
#define N_UNROLL 64sits twenty lines below. The comment has been wrong since the constant was raised, and it is the first line a reader sees.The value is dropped rather than corrected to 64. The next line already reads "Batches up to N_UNROLL blocks per group" with no number, and the
#defineis the only place the value belongs — restating it in prose is exactly what let the two drift apart. Per.claude/rules/comments.md, a comment should state a fact that survives on its own.All four copies land in one commit:
examples/a2a3/tensormap_and_ringbuffer/paged_attention_unroll_manual_scope/examples/a5/tensormap_and_ringbuffer/paged_attention_unroll_manual_scope/tests/st/a2a3/tensormap_and_ringbuffer/paged_attention_unroll/tests/st/a5/tensormap_and_ringbuffer/paged_attention_unroll/The
paged_attention_unroll_manual_scopeREADME (added in #1571) called out the discrepancy; that sentence goes with it, so the docs do not describe a state that no longer exists.Found by scanning
examples/,tests/, andsrc/for#define NAME <int>whose value disagrees with aNAME = <int>mention elsewhere in the same file. These four were the only true positives.Testing
grep -rn "N_UNROLL=8"returns nothing repo-wide#define N_UNROLL 64unchanged in all four filesclang-format --output-replacements-xmlclean on all four.cppNote for rebasing: #1475 touches two of these files and #1521 touches five paths under
paged_attention_unroll*. The overlap here is a single comment line in each.