Skip to content

cggi: remove memrefs from cggi pipeline #1910

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.

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

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

copybara-service[bot]
Copy link
Contributor

@copybara-service copybara-service bot commented Jun 13, 2025

cggi: remove memrefs from cggi pipeline

This PR makes the following changes:

  • Uses linalg-to-affine-loops-tensor to lower linalg operations with tensor semantics
  • Convert yosys RTLIL to MLIR with tensor semantics
  • A few fixes to the secret dialect / patterns to handle secret vs plaintext tensors and insertion operations
  • Remove the option for submodule optimization in Yosys - it never worked quite well, and was suited mostly for memref semantics. I think if we want to invest in Yosys compilation more, we should separate that out into a separate pass that prepares the secret.generic functions so that yosys-optimizer isn't doing all these rewrites within its body

@copybara-service copybara-service bot force-pushed the test_757889057 branch 2 times, most recently from c90d985 to cc6d654 Compare June 13, 2025 20:04
@copybara-service copybara-service bot force-pushed the test_757889057 branch 2 times, most recently from 2301b0d to 69793a3 Compare June 30, 2025 20:34
@copybara-service copybara-service bot changed the title wip: refactor to remove memrefs in boolean pipeline cggi: remove memrefs from cggi pipeline Jun 30, 2025
@copybara-service copybara-service bot force-pushed the test_757889057 branch 6 times, most recently from 177e3ba to 4eb6961 Compare July 7, 2025 19:37
This PR makes the following changes:
* Uses linalg-to-affine-loops-tensor to lower linalg operations with tensor semantics
* Convert yosys RTLIL to MLIR with tensor semantics
* A few fixes to the secret dialect / patterns to handle secret vs plaintext tensors and insertion operations
* Remove the option for submodule optimization in Yosys - it never worked quite well, and was suited mostly for memref semantics. I think if we want to invest in Yosys compilation more, we should separate that out into a separate pass that prepares the secret.generic functions so that yosys-optimizer isn't doing all these rewrites within its body

PiperOrigin-RevId: 757889057
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