On Jul 2, 4:11 pm, Daryl Hammond <[EMAIL PROTECTED]> wrote: Hi Daryl,
> I recently did a clean install of Fedora 9 (formerly running Fedora 8) > and then > installed sage-3.0.3 from source (formerly running sage-3.0). After > running > "sage -testall" I ran a couple of my own programs. > > I was surprised to see one of the programs (sieve of Erasthenes) run > 70% longer. > I installed sage-3.0 from source and verified that sage-3.0.3 was > indeed taking > 70% longer to run the sieve program than sage-3.0 did. > > To narrow down the problem, I installed sage-3.0.1 and sage-3.0.2, > both from > source. I then determined the performance problem was introduced > going from > sage-3.0.1 to sage-3.0.2. > > The following doc shows my system; output from the sieve progam under > sage-3.0, sage-3.0.1, sage-3.0.2, and sage-3.0.3; and source for the > sieve > program. > > The system has a single user, no load, no swapping, and the > performance monitor > indicates that the CPU (2.8 GHz P4) is pegged during the entire > elapsed time. We have had a speed regression with GMP in Sage 3.0.1->3.0.2, so that might be part of it. We intent to fix that for Sage 3.0.4. Could you measure the individual sections of your code, i.e. counting phase vs. assigning "0" to non-prime cases so that we can get an idea if anything specific slowed down? <SNIP> > -Daryl Hammond Cheers, Michael --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
