17:11:55 <vthor> https://github.com/monero-project/monero/wiki/URI-Formatting anybody aware that `monero_wallet` is not a valid scheme? 17:13:21 <vthor> https://en.wikipedia.org/wiki/Uniform_Resource_Identifier#Syntax with a hyphen instead of a an underscore it would be valid. Any thoughts of fixing that for the future? 19:31:41 <moneromooo> A URL with keys or seed. What could possibly go wrong... 19:37:14 <sneurlax> https://github.com/l1mey112/randomx.js 19:37:36 <sneurlax> C->JS, nice to see more FFI work 19:37:51 <sneurlax> s/JS/TS 19:48:23 <m-relay> <fede:xmr.mx> this is crazy interesting, trying it right now! Thanks for sharing! 20:06:28 <vthor> moneromooo: :D that was not the question, the URI exists and is be used, and I use it for view wallet QR so you can scan it from the XmrSigner e.g., but I think it should be changed to `monero-wallet` and still supported for a while `monero_wallet`, I solve that with a simple replace at the moment, so QUrl doesn't get derailed... 20:08:18 <m-relay> <jeffro256:monero.social> That RandomX.js implementation is really interesting. The "semifloat" approach seems a little sketchy to someone who hasn't heard of it before but I'm impressed at the reported performance (I haven't tested it myself) 20:09:02 <m-relay> <fede:xmr.mx> i have tested it, getting around 18 H/s on 1 thread 20:10:33 <m-relay> <fede:xmr.mx> on a 2020 intel laptop i7 CPU 20:10:35 <m-relay> <jeffro256:monero.social> For reference, what's the hashrate of the reference impl on your machine? Also ~100 H/s? 20:11:06 <m-relay> <jeffro256:monero.social> (specifically light mode) 20:12:17 <m-relay> <fede:xmr.mx> running it with: ./randomx-benchmark --avx2 --verify --jit --threads 20:12:25 <m-relay> <fede:xmr.mx> Performance: 15.8248 ms per hash 20:13:15 <m-relay> <fede:xmr.mx> ~64 H/s? 20:15:10 <m-relay> <fede:xmr.mx> does make -j4 create an optimized RandomX build? 20:16:58 <m-relay> <jeffro256:monero.social> no it just compiles with 4 threads 20:17:39 <m-relay> <fede:xmr.mx> ok, i need to use cmake -DARCH=native .. 20:18:28 <m-relay> <fede:xmr.mx> performance didn't change much, 14.8659 ms per hash 20:19:19 <m-relay> <fede:xmr.mx> ~67 H/s vs ~18 H/s, the JS version is ~3.7x slower 20:20:31 <m-relay> <jeffro256:monero.social> I refuse to believe it 20:20:54 <m-relay> <jeffro256:monero.social> That's pretty dang decent for a NodeJS program 20:21:10 <m-relay> <jeffro256:monero.social> > Reaching compliance was done by single stepping the virtual machine and diffing the state with the reference implementation (breakpoint function, JIT instrumentation). 20:21:26 <m-relay> <jeffro256:monero.social> I can sense the agony in that sentence 20:28:38 <m-relay> <jeffro256:monero.social> > download a JS impl of RandomX 20:28:39 <m-relay> <jeffro256:monero.social> > look inside 20:28:41 <m-relay> <jeffro256:monero.social> > C 20:28:45 <m-relay> <jeffro256:monero.social> https://matrix.monero.social/_matrix/media/v1/download/monero.social/abKIGzJQDSRReLBwkpsSyNOG 20:32:25 <m-relay> <jeffro256:monero.social> * 21:21:18 <sech1> Yes, because WebAssembly = almost C 23:35:10 <sturdyvacation> re cli wallet (and chainalysis fingerprinting): I'm looking for a way to send my split spends (for example: % transfer <address 1> <amount 1> <address 1> <amount 2> <address 1> <amount 3>) with each spend staggered by a delay of seconds or blocks. Any thoughts? 23:39:07 <vthor> sturdyvacation: I think monero-wallet-rpc with a script language of you choice will get you pretty quick there, but the machine with the spend keys running on it posses a risk (all dependent on volume and time you have in mind)