refactor: sdk improvements #580
Conversation
…hris/sdk-improvements
ctian1
force-pushed
the
chris/sdk-improvements
branch
from
e53aede to
61b3ea6
Compare
Co-authored-by: Uma Roy <[email protected]>
prover/src/lib.rs
Outdated
| let config = CoreSC::default(); | ||
| let machine = RiscvAir::machine(config); | ||
| let (pk, vk) = machine.setup(&program); | ||
| let (pk, vk) = self.core_machine.setup(&program); |
There was a problem hiding this comment.
Are we including the ELF in the pk now? I feel like it should go in there
There was a problem hiding this comment.
yea program is in SP1ProvingKey already actually
prover/src/lib.rs
Outdated
| reduced_proof: SP1ReduceProof<InnerSC>, | ||
| ) -> ShardProof<OuterSC> { | ||
| // Get verify_start_challenger from the reduce proof's public values. | ||
| let pv = RecursionPublicValues::from_vec(reduced_proof.proof.public_values.clone()); |
There was a problem hiding this comment.
So do we have to do this in verification as well? (Like similar initialization of the challenger)
There was a problem hiding this comment.
no, it's checked in the program when is_complete=1 is committed, so we can just check that
prover/src/lib.rs
Outdated
|
|
||
| // Convert pv.vkey_digest to a bn254 field element | ||
| let mut vkey_hash = Bn254Fr::zero(); | ||
| for (i, word) in pv.sp1_vk_digest.iter().enumerate() { |
There was a problem hiding this comment.
lol can we have more comments as to what is going on here and why (also maybe a link to the corresponding parts in the groth16 circuit code)
There was a problem hiding this comment.
this is not stable yet, it's not constrained in circuit
sdk/src/types.rs
Outdated
| } | ||
|
|
||
| impl Prover for LocalProver { | ||
| fn prove(&self, elf: &[u8], stdin: SP1Stdin) -> Result<SP1DefaultProof> { |
There was a problem hiding this comment.
In the future, it might be nice to have a way to setup a prover with just a pk instead of the ELF, so we don't have to regenerate all of the tables/commitments.
|
This should probably have been squash-merged, or its history been cleaned up before a merge. |
No description provided.