Skip to content

feat(gpu_prover): [WIP] Stage 3 support for unrolled circuits #156

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 26 commits into
base: dev
Choose a base branch
from
Open

feat(gpu_prover): [WIP] Stage 3 support for unrolled circuits #156

wants to merge 26 commits into from

Conversation

@mcarilli
Copy link

@mcarilli mcarilli commented Oct 27, 2025

What ❔

  • Modularize constraint evaluation kernels to reuse shared pieces across unrolled and non-unrolled paths
  • Add unrolled-specific kernels

Why ❔

Is this a breaking change?

  • Yes
  • No

Checklist

  • PR title corresponds to the body of PR (we generate changelog entries from PRs).
  • Tests for the changes have been added / updated.
  • Documentation comments have been added / updated.
  • Code has been formatted.

@mcarilli mcarilli changed the base branch from main to dev October 27, 2025 16:50
mcarilli added 25 commits October 27, 2025 16:25
split range check and timestamp expressions
5c111f3
stage for incremental debugging
f449989
unify GpuComparisonArgs
2bf5924
nit
c20bfb5
Merge remote-tracking branch 'origin/dev' into mc-stage-3-unrolled
037f1f6
use gpu transpose for tests, much faster debugging
426c4f2
baseline works
5a7bb00
FINALLY found stupid decoder lookup bug
addd6c4
fix all multiplicities args and some non-unrolled assumptions
4ec3174
support new and old delegation conventions
ecf6d66
debugging shuffle ram constraints
858386c
test non-unrolled circuits too
1c71b78
found stupid bug in delegation requests
9408ce8
fix another dumb bug
7b6c6c0
fix another dumb bug in ram accesses
c65c65c
machine state permutation constraint
7796839
masking argument works
0e235cb
lazy init and teardown grand product contributions work
24b831f
re-enable register and indirect accesses, use gpu transpose for speed…
5d3bb2b 
… in stage 2 test
re-enable full grand product
4d30688
constraints at every row except last two work
a4c802f
only decoder multiplicities sums left
e766b07
it's over. it's done
28be26e
undo some debugging diffs
39edbb4
undo more debugging diffs
90795cd
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@mcarilli