Shuffling for 32 bit platforms

[ec2]

Issue Addressed

Proposed Changes

• In shuffling, a the raw_pivot (u64) is cast to a usize which will break on 32 bit systems. Now it is modulo'ed with the list_size first then cast to a usize.
• ruint doesn't implement shifting with u64's on 32-bit arch. Since prefix_bits is u8 and NODE_ID_BITS = 256, we use them as u32's instead.

See: https://docs.rs/ruint/latest/src/ruint/bits.rs.html#711

Additional Info

Please provide any additional information. For example, future considerations
or information useful for reviewers.

[cla-assistant]

All committers have signed the CLA.

[cla-assistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[michaelsproul]

We currently do not support 32-bit platforms, and have a compile-time check that prevents compilation on 32-bit arch: https://github.com/sigp/lighthouse/blob/stable/common/target_check/src/lib.rs

Is this required for RISC0? I thought RISCV would be primarily 64-bit?

It's possible we could roll out 32-bit support one crate at a time, but I would like to do this in a way that is testable, rather than just updating the code which is prone to regressions.

[ec2]

We only need 32 bit support for the some of the types in the types crate. We technically don't need it for the shuffling stuff, but then we would just be copy pasta-ing that whole file which seems a little silly.

EDIT: There's something in milhouse that breaks on 32 bit platforms as well. I will be opening a PR there soon.

[michaelsproul]

Ok, I think we should add a CI job that runs the tests for the relevant crates on a 32-bit machine. I'm worried things will just regress otherwise.

Let me discuss with the rest of the team

[ec2]

Thanks boss 🫡 I always appreciate you and the team <3

[ec2]

Related-ish PR: sigp/milhouse#72

[ec2]

@michaelsproul can i get a CI run pl0x <3

[michaelsproul]

I gotchu

[mergify]

Some required checks have failed. Could you please take a look @ec2? 🙏

[michaelsproul]

Beta compiler fail looks unrelated. We'll fix that in another PR

[michaelsproul]

Just needs a back-merge of unstable for CI to pass.

[ec2]

@michaelsproul done! thanks 🙏🏼