by Guest on 2026/03/26 02:06:09 PM
I’d like to request that the developer consider switching Tixati back to the SAM protocol instead of I2CP.
I do a significant amount of uploading over I2P. Over time, I’ve moved from qBittorrent 4.6 to BiglyBT, and eventually to Tixati when it introduced I2P support.
About an hour ago, I decided to try FrostWire after seeing that its latest update included I2P support. To my surprise, I’ve been consistently reaching speeds of around 600 KB/s on several of the same torrents I was seeding in Tixati, where speeds rarely exceeded 250–380 KB/s.
Could you please look into this?
by Guest on 2026/03/27 10:04:39 AM
So why exactly do you think that the command and data exchange protocol between the program and the I2P node (on the same system, mind it) is the slowest link? Do you see any hints that it might be so, like 100% single core usage or process being in the I/O wait state too much?
Let's start with making sure that tunnel lengths and counts are the same in both tests. Are they?
As far as I can see, FrostWire uses libtorrent trough a Java wrapper. Therefore, it should be no different to any other client using a matching version of libtorrent.
Also, when I see that emoji vomit in documentation, I turn 720° and walk away.
