04:04:19 <ridauhebauye[m]> I've been reading a fair bit of monerod code trying to figure some things out, it's starting to get very time consuming so hopefully I can get a few answers here instead. My end goal is to have the safest possible Tor-only configuration for monerod. Initially I was hoping that I could set --proxy and monerod... (full message at <https://libera.ems.host/_matrix/media/r0/download/libera.chat/8df2c6fe6c0eb9a83d6a6f453494c04d05049c0d>)
06:25:10 <jadeydi[m]> hi, I want to invoke functions from monero, list sign transaction, address format etc, using other languages, for go here. But there's not sdk right now, so I want to use CGO (it not support c++), does there c library for monero?
06:42:48 <stretch1> > <@ridauhebauye:matrix.org> I've been reading a fair bit of monerod code trying to figure some things out, it's starting to get very time consuming so hopefully I can get a few answers here instead. My end goal is to have the safest possible Tor-only configuration for monerod. Initially I was hoping that I could... (full message at <https://libera.ems.host/_matrix/media/r0/download/libera.chat/bcde5e5c4c55790ef75b42e9dab2e1ce13742fd0>)
06:43:05 <stretch1>  * read & contribute https://github.com/monero-project/monero/pull/8423
09:47:43 <NorrinRadd> ridauhebauye[m] you will not have incoming connections but once you make an outgoing connection, transactions should be bidirectional. 
10:38:46 <ridauhebauye[m]> <NorrinRadd> "ridauhebauye you will not have..." <- But Dandelion++ compliant nodes only send out to outgoing connections, so while they can be bidirectional they won't be in practice. Or did I get that wrong?
10:45:45 <ridauhebauye[m]> <stretch1> "> <@ridauhebauye:matrix.org> I'..." <- Thanks that's what I  was looking for.
14:27:32 <NorrinRadd> ridauhebauye[m] sorry i tagged the wrong person
15:43:23 <r4v3r23[m]> while testing proxy settings in an android wallet, it looks like monero doesnt respect the proxy and falls back to clearnet when proxy connection cant be made (like shutting tor down after initializing wallet)
15:43:38 <r4v3r23[m]> is this expected behavior?
16:12:46 <ofrnxmr[m]> I dont see why this would have anything to do with monero - the proxy is the issueb
16:16:01 <NorrinRadd> ofrnxmr[m] it shouldn't resort back to clearnet.  that's a privacy leak.  if you set it to a proxy it should only use that proxy or just fail 100%
16:20:58 <moneromooo> I'm not familiar with the monero proxy code, but if it's a real proxy (and not --tx-proxy), then it should definitely not connect direct.
16:21:32 <moneromooo> I assume this is a monerod option, not a separate proxy like torsocks, right ?
16:21:57 <moneromooo> ie, monero doesn't check the HTTP_PROXY env var.
16:25:14 <NorrinRadd> r4v3r23[m] ^
16:33:09 <ofrnxmr[m]> Im under the impression the proxy is not a monerod optiob
16:33:09 <ofrnxmr[m]> --proxy definitely doesnt fall back to clear 
16:51:36 <duggavo[m]> RandomX uses 2MiB scratchpad to be GPU-resistant.
16:51:36 <duggavo[m]> However, the RTX 4090 has 24 GBs of memory, which is about 11500 times 2 MiB.
16:51:36 <duggavo[m]> The RTX 4090 has 16384 CUDA cores.
16:51:36 <duggavo[m]> Does this mean that RandomX is not really GPU-resistant anymore, or GPUs are still noticeably slower because GPU memory has more latency than CPU core cache?
16:56:33 <NorrinRadd> duggavo[m] it has l3 cache? 
16:58:03 <r4v3r23[m]> <NorrinRadd> "ofrnxmr it shouldn't resort back..." <- exactly
16:59:18 <r4v3r23[m]> <moneromooo> "I'm not familiar with the monero..." <- im using `--proxy 127.0.0.1:9050`
16:59:33 <r4v3r23[m]> <moneromooo> "I assume this is a monerod..." <- yes
17:03:45 <NorrinRadd> duggavo[m]  https://github.com/tevador/RandomX/blob/master/README.md?plain=1#L107
17:32:06 <sech1> duggavo[m] you can always try and test https://github.com/SChernykh/RandomX_CUDA
17:32:56 <sech1> memory is only one of many obstacles for an efficient RandomX GPU implementation
18:14:22 <hyc> latency is a big one for GPUs, since their memory is optimized for sequential access not random access
20:15:38 <selsta> r4v3r23[m]: just tested and  --proxy does not fall back to clearnet, you have to be more specific
20:22:53 <selsta> tested both daemon and wallet with --proxy
20:51:00 <pepin[m]> How does p2p discovery work? I am trying to get my node to connect to other nodes via IPv6, however monerod uses ipv4 only even when it's not bound to an address. 
20:52:30 <pepin[m]> Perhaps I can somehow discover IPv6 enabled nodes manually? (Maybe monerod just prefers IPv4 when dual-stack is available)
23:44:30 <r4v3r23[m]> <selsta> "r4v3r23: just tested and  --..." <- ill look into it deeper. just wanted to be sure it wasnt meant to fall back to clearnet
23:45:11 <selsta> if you can reproduce it then i need detailed steps, just running with --proxy and closing tor doesn't make it fallback to clearnet for me