More evidence that Freenet can be reasonably fast when it feels like it... IMHO if we solve the data retention problems and then introduce bloom filter sharing, we could be looking at a considerable performance gain...
[21:03:31] <evanbd> saces: Your test files are all downloading successfully here. Three have finished, three still going. [21:04:50] <evanbd> Looks like overall rate is something like 80 KiB/s. [22:17:58] <toad_> <evanbd> Looks like overall rate is something like 80 KiB/s. [22:18:00] <toad_> evanbd: woah [22:18:33] <toad_> evanbd: high transfer rate only for recently inserted data??? [22:19:23] <evanbd> Yeah. Something like that. [22:19:47] <toad_> evanbd: so sorting out data persistence should also improve performance considerably [22:19:53] <evanbd> Also, it looked like it might have been cycling as it ran over the input bw limit (100 KiB/s) and back down [22:19:59] <evanbd> Probably. [22:19:59] <toad_> 80KB/sec on how big? [22:20:17] <toad_> you said that's an overall rate computed from the time taken to fetch the data? [22:20:57] <evanbd> 6 files, totalling 327MiB [22:21:17] <toad_> nice, and that's overall average from time taken or instantanous transfer rate? [22:21:27] <evanbd> Not precisely computed. That was based on Freenet's total input bw being a bit higher than that. [22:21:39] <toad_> ah well that's not necessarily the same thing :) [22:21:53] <toad_> can you get a timing for the next one? or ... hmmm, i think we report it in FCP? [22:22:06] <evanbd> It's not inconsistent with the total time to fetch, though. [22:22:08] <toad_> i think we report time started and finished in FCP - connect and do a ListPersistentRequests [22:22:28] <toad_> hmmm, you are fairly sure it's in that range?? [22:24:41] <evanbd> Looks like LPR only works on this client's requests [22:24:52] <toad_> huh? [22:25:05] <toad_> WatchGlobal [22:25:06] <toad_> Enabled=true [22:25:07] <toad_> End [22:25:10] <toad_> do that first [22:28:16] <toad_> evanbd: got some numbers? [22:29:16] <evanbd> I don't see times in here... [22:29:29] <toad_> evanbd: they would be in millis [22:37:24] <evanbd> Regardless, I can tell from irc timestamps that it was not less than 47 KiB/s. But the end time on that is fairly conservative. Probably more like 60-70 average. [22:39:00] <toad_> evanbd: i thought we had timestamps ... odd -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 835 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20091010/6c639df9/attachment.pgp>
