New thresholds in table
nisse at lysator.liu.se
Thu Nov 17 15:27:27 CET 2011
nisse at lysator.liu.se (Niels Möller) writes:
> Hmm. I think the threshold should be in the same ballpark as
> HGCD_THRESHOLD. Which makes small values for all x86_64, 64-bit ABI
> builds suspicious. The smallest value for builds in that class is 78,
> for tom.gmplib.org
> One would need to look at sime time/size graphs for those machines to
> figure out what's right.
I've checked a few machines, and there the very small thresholds seem
bogus. Some day ago I changed min_size up again, from 10 to 50. But I
still got some suspiciously small thresholds on several machines,
including on shell. Now I've checked in another tweak, to increasing
stop_since_change from the default 80 to 150. Seems to solve the problem
on shell, we'll see what happens on the other machines.
However, this made no difference to the HGCD_REDUCE threshold. So maybe
it doesn't matter very much if the HGCD_APPR_THRESHOLD is set to 10 or
Niels Möller. PGP-encrypted email is preferred. Keyid C0B98E26.
Internet email is subject to wholesale government surveillance.
More information about the gmp-devel