Danny Willems

Reimport tests from https://github.com/o1-labs/proof-systems/pull/1371 if you want :)

1

I added in some tests in https://github.com/o1-labs/proof-systems/pull/1371 if you want :) _Originally posted by @dannywillems in https://github.com/o1-labs/proof-systems/pull/1396#discussion_r1416108078_

Unify logic for increasing instruction pointer by 4

1

Given that this operation keeps happening over and over again, perhaps it makes sense to create a wrapper for it. _Originally posted by @querolita in https://github.com/o1-labs/proof-systems/pull/1365#discussion_r1411959202_

Flaky test: tests::lookup::test_runtime_table_only_one_table_with_id_zero_with_non_zero_entries_random_values

2

https://github.com/o1-labs/proof-systems/actions/runs/7047117196/job/19180286206

Follow-up: we could have a routine called `is_positive` which is more specific. _Originally posted by @dannywillems in https://github.com/o1-labs/proof-systems/pull/1434#discussion_r1417171179_

We could improve it later by checking only one bit. We could have `test_is_negative`. _Originally posted by @dannywillems in https://github.com/o1-labs/proof-systems/pull/1421#discussion_r1417149561_

Can we add some ASCII doc like [here](https://github.com/o1-labs/proof-systems/pull/1371/files#diff-07f81252f8ef24c4333665a2666686c4bb46ce667c27b1e3d25e198a93692165R336)? ``` // R-Format: register-to-register // I-Format: register immediate // J-Format: jump format // shamt: shift amount // rs/rt are sources for R-format,...

I believe I would prefer to have separate functions for each type of instruction. Like `decode_instruction` would match 3 branches, each of them with some intervals, and then you have...

Add a test checking the domain size when building a circuit with gates using same table.

1

Approving, but would have been nice to have a test, even it looks fine regarding the code. Opening a follow-up issue for a 10 min task. _Originally posted by @dannywillems...

u32/u64 <--> int32/int64: dangerous to allow?

1

The doc says u32 (resp. u64) is mapped to int32 (resp. u64). Is it not dangerous to allow this regarding overflow? Should it be mapped to a non native Caml...

Call overhead compared to C bindings

8

I have been doing some experimentation to see the actual cost of using ocaml-rs. When writing C bindings, I have been used to have a 4-6ns overhead due to root...