You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With Ed25519 it's possible to do batch verification which optimizes verifying a set of signatures vs a set of public keys all at once. As our committee sizes grow (and we need to start validating SGX commits), it's probably worth thinking about utilizing this technique to reduce the time spent validating commits.
golang.org/x/crypto/ed25519 last I checked is still agl's Go port of djb's ref10, but he left out the batch verification implementation. Porting the rest of it over should be a straight forward drop-in process.
If signature verification/signing ends up being a perf bottleneck, it's possible to implement all of these much faster than, ref10 as well, though doing it without massive pain will require Go 1.12 (for math/bits integer intrinsics).
The text was updated successfully, but these errors were encountered:
With Ed25519 it's possible to do batch verification which optimizes verifying a set of signatures vs a set of public keys all at once. As our committee sizes grow (and we need to start validating SGX commits), it's probably worth thinking about utilizing this technique to reduce the time spent validating commits.
golang.org/x/crypto/ed25519
last I checked is still agl's Go port of djb'sref10
, but he left out the batch verification implementation. Porting the rest of it over should be a straight forward drop-in process.If signature verification/signing ends up being a perf bottleneck, it's possible to implement all of these much faster than,
ref10
as well, though doing it without massive pain will require Go 1.12 (formath/bits
integer intrinsics).The text was updated successfully, but these errors were encountered: