asraa

Hey, I'm currently exploring this. For extract_slice / insert_slice that are simply rank-reducing or expanding, I'm adding patterns that would canonicalize them to expand/collapse shape ops that would just be...

The remaining part of this (that https://github.com/google/heir/pull/2036 pushes) is that RNS types are not supported in polynomial-to-mod-arith now. Since poly coeff types can be RNS, poly operations are still operating...

Oooh I like this! I'm just curious - does hardware natively support a vector of different modular operations? Or do they break them down into individual single mod Z_p operations?

Ah right, yes. mod_arith ops take RNS types too. I think then maybe an RNS-unroll transform pass at the mod_arith level is what this should be

Thank you for starting this! Cloned it and I''ll push some changes to make everything work with bazel. > cli argument string, which people interact with regularly, Agree, although I'll...

> Thanks! Btw, I still see two files in tests/old- did you push all of your most recent changes? Oops! Rebased and pushed now. > @ polynomial runner: I think...

> @asraa Can you take a look at why the verilog tests fail? I don't see any obvious changes from how things were set up before. Done! The lit runner...

> Right! There is also some in tools/*.bzl now too, but I don't really see a difference in marking these macros as test vs non-test. I think we probably want...

> Did you already start the internal process or should I update with mod_arith tests and rebase? I'll continue to rebase this, the internal transforms had to change to import...

> Btw, where do we currently use secret.conceal for trivial encryptions? This will only happen if there's some reason to reconcile types - rarely happens, but see https://github.com/google/heir/blob/da734afdeaf5e4a1e4f4fb594d0e036bec7ae260/lib/Dialect/Secret/Transforms/DistributeGeneric.cpp#L262 for the...