Author lemburg
Recipients Arfrever, Giovanni.Bajo, PaulMcMillan, Vlado.Boza, alex, arigo, benjamin.peterson, camara, christian.heimes, dmalcolm, koniiiik, lemburg, serhiy.storchaka, vstinner
Date 2012-11-07.11:19:16
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
In-reply-to <>
On 07.11.2012 12:06, Armin Rigo wrote:
> Armin Rigo added the comment:
> Marc-André: estimating the risks of giving up on a valid query for a truly random hash, at an overestimated one billion queries per second, in a 2/3 full dictionary:
> * for 1000: 4E159 years between mistakes
> * for 100: 12.9 years between mistakes
> * for 150: 8E9 years between mistakes
> * for 200: 5E18 years between mistakes
> So while it seems that 100 might be a bit too small, using 150 to 200 is perfectly safe (and that's "perfect" in the sense that a computer will encounter random hardware errors at a higher rate than that).

I used the 1000 limit only as example. In tests Victor and I ran (see the
original ticket from a few months ago), 200 turned out to be a reasonable
number for the default maximum hash collision value.

I'm not sure about the slot collision limit. We'd have to run more tests
on those.
Date User Action Args
2012-11-07 11:19:16lemburgsetrecipients: + lemburg, arigo, vstinner, christian.heimes, benjamin.peterson, Arfrever, alex, dmalcolm, Giovanni.Bajo, PaulMcMillan, serhiy.storchaka, Vlado.Boza, koniiiik, camara
2012-11-07 11:19:16lemburglinkissue14621 messages
2012-11-07 11:19:16lemburgcreate