--- Log opened Wed May 14 00:00:42 2025
02:13 -!- jonatack [~jonatack@user/jonatack] has joined #secp256k1
09:09 -!- pablomartin_ [~pablomart@217.130.254.81] has joined #secp256k1
09:45 -!- pablomartin_ [~pablomart@217.130.254.81] has quit [Ping timeout: 244 seconds]
09:45 -!- pablomartin_ [~pablomart@217.130.254.81] has joined #secp256k1
10:26 -!- pablomartin_ [~pablomart@217.130.254.81] has quit [Ping timeout: 248 seconds]
10:27 -!- pablomartin_ [~pablomart@217.130.254.81] has joined #secp256k1
11:30 -!- pablomartin_ [~pablomart@217.130.254.81] has quit [Ping timeout: 244 seconds]
13:01 < bitcoin-git> [secp256k1] jonasnick pushed 2 commits to master: https://github.com/bitcoin-core/secp256k1/compare/bb597b3d3987...9fab42525676
13:01 < bitcoin-git> secp256k1/master 0544537 Sebastian Falbesoner: bench_ecmult: add benchmark for ecmult_const_xonly
13:01 < bitcoin-git> secp256k1/master 9fab425 Jonas Nick: Merge bitcoin-core/secp256k1#1668: bench_ecmult: add benchmark for ecmult_...
13:01 < bitcoin-git> [secp256k1] jonasnick merged pull request #1668: bench_ecmult: add benchmark for ecmult_const_xonly (master...add-ecmult_const_xonly-benchmark) https://github.com/bitcoin-core/secp256k1/pull/1668
15:47 -!- jonatack [~jonatack@user/jonatack] has quit [Ping timeout: 245 seconds]
15:50 -!- jonatack [~jonatack@user/jonatack] has joined #secp256k1
19:55 < andytoshi> roconnor: do you remember offhand what you did to remove the secp256k1_ge_globalz_set_table_gej function in e5c18892db69b5db44d282225ab4fea788af8035 in 2021?
19:55 < andytoshi> at least, was it some very clever field arithmetic thing, or a simple refactor?
19:56 < roconnor> is that a git commit?
19:56 < andytoshi> yeah
19:56 < andytoshi> https://github.com/bitcoin-core/secp256k1/pull/899 commit 2
19:56 < andytoshi> i am rebasing an old vanitygenner and i think i'm gonna have to rewrite it because of this..
19:57 < roconnor> oh, this isn't simplicity
19:57 < andytoshi> heh, yeah, it's #secp256k1
19:57 < roconnor> I think it was like a complicated refactor
19:58 < andytoshi> ok, cool. i'll study it
19:58 < roconnor> but I don't know why it would affect your vainity generator.
19:58 < roconnor> It was functionally equivalent.
19:58 < andytoshi> ok, awesome, "functionally equivalent" is encouraging
19:59 < roconnor> The process was approximiately, rewrite it in Haskell
19:59 < roconnor> optimize the resulting haskell
19:59 < roconnor> go back and do those optimizations in C now that things are not so opaque.
20:05 < andytoshi> ok, thanks. i think i've gotten it. at least, now all my problems are related to cmake
20:05 < roconnor> even more, I think it was basically the same calculation
20:05 < roconnor> just storing stuff differently
20:05 < andytoshi> yeah, that's what it looks like actually
20:05 < andytoshi> it's just hard, in C, to tell what's going on with in-params and out-params looking basically the same
20:26 < andytoshi> hmm, i think i might have to resurrect the old function ... in the main codebase secp256k1_ecmult_odd_multiples_table and secp256k1_ge_table_set_globalz are always paired, and i think your optimization assumes that
20:26 < andytoshi> but i was using only the latter
20:47 < andytoshi> o nice, i got it!
20:48 < andytoshi> if i am feeling motivated enough to get my old code building on nixos, i'll try to benchmark it againsnt the new stuff
--- Log closed Thu May 15 00:00:43 2025