Hi folks ... over the weekend. I am ... be able to help with any ... Hehehehe!!! The 'interface' as it stands at the moment will probably make your stomach...
3
Chris Nash
chris_nash@...
Oct 14, 2000 10:11 pm
Hi there folks, Many thanks to Phil for pointing out his 'direction' for those list members who don't know ... totally removed, ... ideal as it's ... written...
4
Chris Nash
chris_nash@...
Oct 16, 2000 5:27 am
Hi folks, I mentioned earlier that I was looking at constructing an "FFT emulation" using GMP so you can see how the pfgwlib interface works, and have half a...
5
Chris Nash
chris_nash@...
Oct 21, 2000 8:12 pm
Hi folks, Just a few quick remarks about dev-20001012, for those of you who looked at it. 1) Peter Kosinar patched root and pfoo. The root configure script ...
6
Michael Bell
mdb36@...
Oct 21, 2000 11:35 pm
Hi, ... It appears that my build does actually work (very limited test done - I just automatically tried pfgw -b to check the different run lengths and it...
7
Chris Nash
chris_nash@...
Oct 22, 2000 12:43 am
Hi folks ... I just ... caused ... seems ... for ... Here's the diff file produced by Peter's checkin. diff -u -r1.4 -r1.4.4.1 ... +++ f_factor.cpp...
8
Chris Nash
chris_nash@...
Oct 23, 2000 3:37 am
Hi folks, I've just promoted 20001012 to 'stable', and checked in dev-20001022. Things that have changed: - root now includes Chris Caldwell's top 5000 to be...
9
Paul Jobling
Paul.Jobling@...
Oct 23, 2000 8:43 am
... If this is the original version, I managed to build it with no problems. The only bit of oddness that I had to deal with were two .obj files (I forget...
10
Phil Carmody
fatphil@...
Oct 23, 2000 3:39 pm
... All my x10 and x22 primes have been proved on my GMP (karatsuba?)code on my 21164 Alpha - about the same speed as a P166! I'm on holiday at the moment, but...
11
Phil Carmody
fatphil@...
Oct 24, 2000 4:16 pm
OK, I've taken about 1% off the runtime of primegen up to 10^9, but by the time it's up to 3*10^12 it runs in 55% of the time that it used to. Are you...
12
Paul Jobling
Paul.Jobling@...
Oct 24, 2000 4:23 pm
Fortunately PFGW won't usually be trial factoring that high, but any performance improvement is worth while. I wonder if the NewPGen approach could be used -...
13
Chris Nash
chris_nash@...
Oct 24, 2000 5:25 pm
Hi folks ... Well, any performance increase is, after all, an increase. Put in some perspective, the time taken to generate a prime is usually overwhelmed by...
14
Chris Nash
chris_nash@...
Oct 27, 2000 7:43 am
Hi folks, I just downloaded FFTW from http://www.fftw.org/download.html (I love their link 'Get FFTW before the devil does!' - and generally, these MIT bods...
15
Chris Nash
chris_nash@...
Oct 29, 2000 2:25 am
Hi folks, Thanks to David Broadhurst's wonderful maintenance of the 'BigBird' project folder in the primeform files area, I've had some very applicable data to...
16
Chris Nash
chris_nash@...
Oct 30, 2000 3:44 am
Hi folks, A new dev update is coming soon. It adds two major new features (at least, new to pfgw). -t does the N-1 test (33% or more factors of N-1 gives a...
17
Chris Nash
chris_nash@...
Nov 2, 2000 12:21 am
Hi folks Paul Underwood just wrote ... Whoah Paul, that didn't take long! Sounds like we may well have overoptimized.... ... Indeed, I just threw...
18
Chris Nash
chris_nash@...
Nov 2, 2000 6:13 am
Hi folks, Jim Fougeron built a Windows version of pfgw, dev-20001029. Everything seemed to go quite well, and Jim made a few minor source modifications to...
19
Jim Fougeron
jfoug@...
Nov 2, 2000 11:44 pm
Hi all, ... Everything ... modifications ... add ... Chris now has my full Windows development environment. I still had a few things to work out before posting...
20
Phil Carmody
fatphil@...
Nov 3, 2000 12:15 am
... baited breath, and all that. ... AFAIK there is no unnecessary overhead from GMP at all. I use it for all my bignum jobs, and have not noticed anything...
21
Chris Nash
chris_nash@...
Nov 3, 2000 5:28 am
Hi folks ... Beware! Comparing proth and pfgw is very much like comparing apples to oranges - and causes all sorts of strain in the 'diplomatic relations' ...
22
Chris Nash
chris_nash@...
Nov 3, 2000 5:38 am
Hi folks ... holiday. Up until the last couple of days, you didn't miss much. Jim's Windows build seems to have lit a fire under the primeform list, though :) ...
23
Phil Carmody
fatphil@...
Nov 3, 2000 8:36 am
... Primegen works in 4 phases. 1) clear everything. (which is why you can start anywhere trivially, each block of numbers us calculated afresh) 2) toggle...
24
Chris Nash
chris_nash@...
Nov 3, 2000 11:02 am
Thanks for the summary Phil! I looked at it for long enough to realize "well, it generates decent- sized chunks of primes at a time", did a few tests, and it...
25
Phil Carmody
fatphil@...
Nov 3, 2000 11:58 am
... I have no idea how the code and the maths relate to each other. I'm doing black box coding. I'm moving things around, making sure that the results are...
26
Jim Fougeron
jfoug@...
Nov 4, 2000 4:53 am
... I speak with knowledge, having been the ultimate rebellious teenager before (damnit we all have to grow up sometime ;). Sometimes, the "parent" needs to...
27
Chris Nash
chris_nash@...
Nov 4, 2000 7:06 am
Hi folks ... Exactly. Why not indeed. There's no reason for instance why pfgw can't implement the PRP modular reducer. The 'hook' for it is already there - ...
28
Phil Carmody
fatphil@...
Nov 4, 2000 10:57 am
... The perl hacker in me says use "-" as the pseudo-filename for stdin. I can just see it now, one PFGW process hanging off a named pipe, and another process...
29
Phil Carmody
fatphil@...
Nov 4, 2000 3:38 pm
... I'm curious, where does the 64-bit k restriction come from? Surely arbitrary k would be a better goal. For example, I'm currently bogged down in (huge...
30
Chris Nash
chris_nash@...
Nov 4, 2000 9:24 pm
Hi folks ... installed, and that breaks the whole system. ... serious mathmos will already ... *giggles* It's there purely for the darn legalities of the GPL....
