asraa

> Now that we have SecretnessAnalysis, we could update the --heir-simd-vectorizer to only apply its transformations to tensors that are indeed secret. Wouldn't it also be beneficial to vectorize plaintext...

I'd like to talk at the HEIR WG meeting this week on the LWE type - we know we need to revamp the encoding attribute to accommodate prime modulus anyway,...

Hey! @ahmedshakill what's the state of this? Do you need me to take a review of it?

Once https://gist.github.com/asraa/6105c144aa10295aa4f021438d5b046d is inscribed into LWE types, I assume we'll have a place for that, and then the Action Item for this issue would be to update the conversion pass.

https://gist.github.com/asraa/6105c144aa10295aa4f021438d5b046d I started a small scratch doc that I'll convert into a draft PR for commenting purposes.. BTW: > Maybe not necesary, unless there are situations where dimension doesn't uniquely...

Right! My feeling was putting a manual description in a README.md as a section.

The internal code that runs the stableHLO to standard MLIR lowering does: 1. stablehlo-legalize-to-hlo 2. Create XLA HLO module 3. Partition computation using XLA utils 4. Convert subgraphs to MLIR...

Thank you! Re-visiting this now after the break. There's two parts to this really (1) Lowering from JAX/PyTorch/TF to StableHLO (watching the video right now!) (2) Lowering StableHLO to standard...

> I would guess this should be the very first pass, but I'm not sure. Same here - I would expect that there would be smarter ways of ordering the...

Just realized the tensor encoding attribute can have a map :)