On Thu, Jun 12, 2008 at 06:10:10PM -0700, Tim Bray wrote: > In /opt/coolstack/bin
> No idea how much (if any) it'll help, but I'd be eager to hear. -Tim
Interesting. It may be making a /slight/ improvement over my home-built 2.5 32 bit on O.10m, but probably within the margin of error. Strangely, I'm seeing faster results on my 2.5 than previously, so I'm going to have to do some more digging.
James
-- /-------------------------------------------------------------------------- \ James Aylett xapian.org ja...@tartarus.org uncertaintydivision.org
On Fri, Jun 13, 2008 at 11:33:55AM +0100, James Aylett wrote: > > In /opt/coolstack/bin > > No idea how much (if any) it'll help, but I'd be eager to hear. -Tim
> Interesting. It may be making a /slight/ improvement over my > home-built 2.5 32 bit on O.10m, but probably within the margin of > error. Strangely, I'm seeing faster results on my 2.5 than previously, > so I'm going to have to do some more digging.
There isn't much between the Coolstack version of 32 bit 2.5 and mine, but it seems slightly in its favour. Over the times we're talking about for the top of the WF-2 table, the difference will be in the seconds (around 1-2%); Coolstack versus the built-in 32 bit 2.4.4, I'm seeing a 10% improvement, so that's a big win.
James
-- /-------------------------------------------------------------------------- \ James Aylett xapian.org ja...@tartarus.org uncertaintydivision.org
> On Fri, Jun 13, 2008 at 11:33:55AM +0100, James Aylett wrote:
>>> In /opt/coolstack/bin >>> No idea how much (if any) it'll help, but I'd be eager to hear. - >>> Tim
>> Interesting. It may be making a /slight/ improvement over my >> home-built 2.5 32 bit on O.10m, but probably within the margin of >> error. Strangely, I'm seeing faster results on my 2.5 than >> previously, >> so I'm going to have to do some more digging.
> There isn't much between the Coolstack version of 32 bit 2.5 and mine, > but it seems slightly in its favour. Over the times we're talking > about for the top of the WF-2 table, the difference will be in the > seconds (around 1-2%); Coolstack versus the built-in 32 bit 2.4.4, I'm > seeing a 10% improvement, so that's a big win.
So that's probably caused by a faster interpreter in Python 2.5?
