docs: localize shared-core deep path (GftSadd=54, GftSmul=44) — pipeline the RNE cascade (Refs #1764) - #1898
Merged
Conversation
#1764) Measured the critical logic depth so the eventual pipeline cut lands in the right place. GftSmul is purely combinational (no read-before-ready bug). Yosys ltp: GftSadd=54, GftSmul=44. The depth is NOT the multiplier width — magmul's 32-iteration shift-add has 10-bit operands (512+mant), so yosys prunes the dead upper iterations; hand-narrowing to 10 leaves depth unchanged (44->45), a dead end (tested before touching the verified spec). The real depth is the dependent normalize/round cascade (magsub priority-shift + <<14 + RNE; magmul post-product RNE). Concrete pipeline cut: split each cascade into two registered stages (~27/~22), which halves the path AND resynchronises it (kills the placement hazard). Reinforces that the ~47ns nominal path is ~100x under the settle window, so the glitch is a hazard static timing can't see. Refs #1764 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributor
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Measures where the shared-core critical combinational depth actually is, so the eventual pipeline cut lands in the right place instead of guessing.
GftSmulis purely combinational (assign result = smul(a,b); theclk/en/readyports are unused) — so there is no read-before-ready handshake bug.ltp(longest topological path):GftSadd= 54,GftSmul= 44.magmul's*lowers to a 32-iteration shift-add, but its operands are512+mant ∈ [512,1023](10-bit), so yosys already prunes the dead upper iterations. Hand-narrowing the loop to 10 leaves the depth unchanged (44 → 45) and area flat — a dead end, tested before touching the verified spec (observe-before-mutate).magsub's 4-stage priority-shift +<<14fixed-point + round-to-nearest-even inGftSadd, andmagmul's post-product RNE carry inGftSmul.Concrete pipeline cut (for the Variant-1 work): split each cascade into two registered stages (~27 / ~22 deep), which both halves the path and resynchronises it (killing the placement hazard); the microsequencer then waits a fixed 2-cycle latency instead of a
settlecounter. The ~47 ns nominal path is already ~100× under the µs settle window, which is why the glitch is a placement hazard a static-timing fix cannot see.Docs-only; no code or gate changes. Refs #1764