On Mon, Aug 13, 2012 at 01:32:31PM -0400, Kevin A. McGrail wrote: > Anyone have some feedback on the memory footprint for their masscheck?
For a sample of 1400 messages, mass-check just took a maximum of about 50 MB (resident set size; 56 MB virtual) on my machine. Thinking that maybe messages are kept in memory, I tried removing all e-mails above 300 KB from the corpus. This did not reduce the memory used at all, however; in fact, this time mass-check used 8 KB more. I then compiled a sample of about 17000 (unsorted) messages, totalling 221 MB in file size -- so the average message size is much smaller for me than for Michael. During mass-check for these collected folders, the maximum observed memory use (checkpoints at 60 second intervals) was less than 62 MB virtual. Run time was nearly 90 minutes. This is an ageing Linux i386 system; mass-check was run with -j 1. In short, I cannot reproduce the problem, even though I tried. Sorry. ;) Regards, Marc
