Hi! I ran the new 0.60.9 version with a small number and the perl script. During the relation phase it once did a error -193 error, but recovered. Earlier it...
... Hi, Murphy's thesis is available online at Richard Brent's website: http://web.comlab.ox.ac.uk/oucl/work/richard.brent/ftp/Murphy- thesis.ps.gz Chris...
hi ! I have posted (and Sander too) about the incapacity of v 0.54.5b to find the square root of a gnfs composite. Chris Monico answers me it was a "matrix...
Oh darn - I wish I would have seen this before. But I suppose I can just go into the header where I've defined u32,s32,... and define or typedef these to...
Hi Tim, Sorry it took me a while to get back to you - I haven't checked this list in a couple of weeks (my wife and I are having a baby very soon, and so...
Hello, Yep - this is certainly possible. The 'q' value is just a particular large prime (norm) being used to lattice sieve with. Each 'q' value could (and...
Hi, This problem has been fixed, as of 0.61.1. I did something very stupid in the relation filtering code. The net effect was that procrels failed to filter...
Hello, I am running this one right now with 0.61.2 (insignificant changes over 0.61.1). If it does crash, I will find and fix the problem. I am hopeful that...
Hello, Okay - I am more hopeful that the bug which caused this problem is gone. The factorization did complete for me: $ more summary.txt Number: test ...
Hello, I tried again with the new 0.61.2 verion under CYGWIN on my W2000 box (P-III). The CYGWIN seems to compile as Athlon. Each time all OLD code is removed...
Hi ! I still have problems with the sqrt module under cygwin. Since I understand that reproducing problem I encountered with a c104 is a big task, I have made...
Hi All, Tetsuya Kobayashi reported that a 165 digit near-repdigit had been factored by GGNFS-0.54.4. Four sets of PCs were used to do it. 10^165-9 = 99...991 =...
We are proud to announce we factored an 397 bit composite number with ggnfs 0.60.6. We sieved around 150 million relations which gave us 26 millon after...
Hi! I tried to make the new version (0.61.3) as usual under cygwin. It bombed out in building Lanzcos64 code. gcc -I../include -O3 -Wall -march=athlon -o...
Hi, This is because Cygwin does not support posix_memalign, but it does support memalign. I made these changes and compiled a working version of 0.61.3, and...
Hi again, The script files which had been used in the factoring of 10^165-9 is available at: http://homepage2.nifty.com/m_kamada/math/nu200411.htm#041110_01 A...
... version ... It appears in Msys / MinGW it also doesn't recognize posix_memalign but it also doesn't recognize memalign. Anyone with some more expertise can...
Hi! There seems to be something fishy. I tried my old C81 again after fixing the changes by Greg. Everything worked nice until the matrix step. End of screen: ...
That's odd. Everything looks good below until the crash. Here's the output on my Cygwin run. All the numbers seem to match yours, only mine didn't crash. I...
Hmmm...it appears that the authors of FFTW (www.fftw.org) have been down this aligned-malloc road and figured it out on most platforms. Their code is GPLed, so...
Hi! Thanks for the tips, the make file had athlon instead of pentium3 now I run a new test with hopefully better result. Thanks for checking. Torbjörn Alm ...
I've successfully finished the sieving on a general c105, and then it tells me that depinf has been written, but it goes on to do another sieving cycle! I...
Hi Bob, ... The latests factlat script has a new setting: # The more extra FF's there are, the smaller the resulting pruned # matrix will be (upto a certain...
Hello Sander, ... then it ... another ... matrix ... depinf and ... Thanks for the advice, so I set it back to zero, and tried again. This time it got into the...
... Hi ! Sieving 2 or 3 times the same number with different polynomials, dumping and processing relations again and again with different ggnfs versions, I...
This'll be fixed in the next version. In the meantime, you can enable the #define NEED_UNDERSCORES in src/lasieve4/piii/asm_mul64.S and recompile. That...
