Closed Sladuca closed 2 years ago
Use the blst's implementation of Pippenger to perform MSMs, which is ~4-5X faster for 1000 elements on my 2019 MB Pro 16. Due to the unsafe interfaces I'm cloning the vectors, though I'm open to suggestions for ways to not do this.
blst
See #28 for more information.
Use the
blst
's implementation of Pippenger to perform MSMs, which is ~4-5X faster for 1000 elements on my 2019 MB Pro 16. Due to the unsafe interfaces I'm cloning the vectors, though I'm open to suggestions for ways to not do this.See #28 for more information.