Hi, On 11/18/2010 08:58 AM, Zooko O'Whielacronx wrote:
> Could you please do the same experiment with Tahoe-LAFS v1.7.1? If it > was just as ill-behaved back then, I guess this isn't a recent > regression and we don't have to hold up 1.8.1 to investigate it. Here are the kind of performance I see with v1.7.1. 1-of-3 51 2-of-3 54 3-of-3 35 10-of-30 105 20-of-60 224 I consistently see almost half download times for a 100 MB file with K=20 with v1.7.1 than with v1.8.1. This looks like a regression of the new immutable file downloader for large values of K. > http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1200# package up Brian's > New Visualization of immutable download > > François: if you could apply Brian's New Visualizer and post > screenshots of the visualization of the performance when you have > 20-of-60 encoding, that would be cool! I'll do that. François _______________________________________________ tahoe-dev mailing list [email protected] http://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
