Skip to content

docs: trim Performance to current implementation vs smalljac (scalar + batched) - #6

Merged
salindne merged 2 commits into
mainfrom
docs-perf-smalljac
Jun 18, 2026
Merged

docs: trim Performance to current implementation vs smalljac (scalar + batched)#6
salindne merged 2 commits into
mainfrom
docs-perf-smalljac

Conversation

@salindne

Copy link
Copy Markdown
Owner

Tightens the README Performance section to describe only how the current implementation compares to smalljac, scalar and batched.

Removed (journey / previous-attempt material):

  • the PrimeField vs MontgomeryField before/after table,
  • the per-field inversion-cost breakdown (529 vs 109 ns, etc.),
  • the "explicit formulas were already sound / the gap was the field layer" narrative,
  • the field-operation-count table (op_counts.rs and its test are unchanged, just not featured in the README).

Kept — one focused table, this crate (MontgomeryField) vs smalljac at a 56-bit prime:

operation this crate (scalar) smalljac (scalar) this crate (batched) smalljac (batched)
add 163 190 40 48
double 168 210 46 54

Second commit replaces all em dashes () in the README with plain punctuation (math minus signs in Vn = −Vpl − h are left as-is).

Docs-only; goes through a PR per the new branch protection on main.

salindne added 2 commits June 18, 2026 18:01
Reduce the Performance section to a single focused comparison — this crate's
g2::ramified::not_char2 (MontgomeryField backend) versus smalljac, scalar and
batched. Drop the journey/explanatory material: the PrimeField vs MontgomeryField
before/after table, the per-field inversion-cost breakdown, the 'formulas were
already sound / the gap was the field layer' narrative, and the field-operation
count table. (op_counts.rs and its test remain; just not featured in the README.)
@salindne
salindne merged commit 87e2ebc into main Jun 18, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant