feat(jolt-core): Adding new multiplication instructions

[mw2000]

Jolt will now support some multiplication instructions - MUL, MULU for the RISC-V M extension.

Addresses parts of #251

[mw2000]

Hey @moodlezoup this PR should be ready for review now. I originally planned to include the MULHU instruction in this as well, but was failing at implementing it (skill issue).

I based the MUL, and MULU instructions closely off what we have for the ADD and SUB instructions, would love to know if this approach makes sense or if I should modify to have a custom sub-table for this.

[mw2000]

Appreciate the direction setting on this in DMs moodlezoup, got the MULHU instruction to work :)

[moodlezoup]

Thanks for working on this! One request for the MULHU instruction

[moodlezoup]

For our current parameters C = 4 and M = 2^16, the TruncateOverflowSubtable here is just all zeros (which is what we want), but I think this would be broken for C and M where C * log2(M) != 2 * WORD_SIZE
For now, let's just add an assert_eq! and remove the TruncateOverflowSubtable

Suggested change

	let msb_chunk_index = C - (WORD_SIZE / log2(M) as usize) - 1;
	// Reversed the order of the subtable indices compared to MUL and MULU
	vec![
	(
	Box::new(TruncateOverflowSubtable::<F, WORD_SIZE>::new()),
	SubtableIndices::from(msb_chunk_index + 1..C),
	),
	(
	Box::new(IdentitySubtable::new()),
	SubtableIndices::from(0..msb_chunk_index + 1),
	),
	]
	assert_eq!(C * log2(M), 2 * WORD_SIZE);
	vec![
	(
	Box::new(IdentitySubtable::new()),
	SubtableIndices::from(0..C / 2),
	),
	]

[moodlezoup]

(there are existing instructions that assume C = 4 and M = 2^16, so I think this is fine for now)

[mw2000]

Sounds good, went with your approach. I also had to change the concatenate looks param to take in vals.len() instead of C for it to not panic.

[moodlezoup]

nice work!