Open kevaundray opened 1 month ago
Testing the difference:
BenchmarkOpenMonomial-10 100 10691245 ns/op 406979 B/op 180 allocs/op
BenchmarkOpenLagrange-10 124 10653541 ns/op 799451 B/op 163 allocs/op
Lagrange is what we currently do and monomial is the change proposed. The difference here is 10.69 vs 10.65 milliseconds
Strangely open monomial uses ~49% less memory
Higher level benchmarks have been collected in #97, #98, #99
Previously I was against changing the 4844 code due to it already been auditted, however the code has changed quite a lot since we added 7594 methods, so it may be feasible to change that code given it will likely need to be auditted again.