Cached modular inverses

[michiexile]

For relatively small primes (say < 65k or so), precalculating all inverses and storing them in a lookup table is often faster (amortized over a long calculation in the same finite field) than recalculating each inverse whenever it is needed.

[PoslavskySV]

Thanks for the suggestion! I'm not sure that currently this is really needed, but at some point I'll certainly return to this.

[michiexile]

I'm looking to adopt the library for a large persistent homology project I'm planning - and for that, there will be sufficiently many computations within a single finite field that using cached inverses is critical to get useful performance.

…

On Mon, Aug 26, 2019, 16:39 Stanislav Poslavsky ***@***.***> wrote: Thanks for the suggestion! I'm not sure that currently this is really needed, but at some point I'll certainly return to this. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#57?email_source=notifications&email_token=AAAH6LUTUQUCHEVLY5E3SSTQGQ5RRA5CNFSM4IMKWGJKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5FTO3Y#issuecomment-525023087>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAAH6LRMHFRLHN2FXTPPEYTQGQ5RRANCNFSM4IMKWGJA> .

[PoslavskySV]

I see. I've created a small PR #59 which addresses this issue. Is that what you need?