Current Effort

[jasonp]

Originally Posted by bdodson

sieving was done already some
time ago, and the matrix has been grinding away for a year or so?

I think sieving was about half done by last October.

Originally Posted by batalov

As I walk through the valley where I harvest my relns
I take a look at my matrix and realize she's very plain
But that's just perfect for an Amish like me
You know I shun fancy things like B-Wee...

You frighten me sometimes :)

[xilman]

Originally Posted by bdodson

to see where matrix and sieving are ... sieving was done already some
time ago, and the matrix has been grinding away for a year or so? -bd

I sent in the last of my relations in early March. By that time roughly twice as much sieving as the bare minimum had been performed.

AFAIK, the project is still in the filtering/merging phase. At least, I haven't heard of the matrix having started yet.

It would be nice if the factorization was finished by SHARCS 09 in early September. I've no idea whether that is possible,


Paul

[fivemack]

I hadn't realised there was to be a SHARCS 09; thanks! I wonder whether Switzerland or Vienna will prove to be more the nest of spies ...

[bdodson]

Originally Posted by xilman

I sent in the last of my relations in early March. By that time roughly twice as much sieving as the bare minimum had been performed.
...
Paul

So this would be the first number done with Franke, et. al. sieving
that's over-sieved. All of the others (three from Franke's Utrecht
report), plus rsa200, plus snfs1024, were rather under sieved, in favor
of shifting more of the computation over to the matrix. We know
(from Bernstein) that -- in theory -- sieving is the only step that
matters, asymptotically; but this sounds like the first with Franke
sieving that recognizes the facts of our life (pre-asymptotic) that
it's the matrix that currently holds us up.

bdodson (with no disrespect intended to our friend, Thorsten)

[Raman]

1.

Originally Posted by xilman

I sent in the last of my relations in early March. By that time roughly twice as much sieving as the bare minimum had been performed.

AFAIK, the project is still in the filtering/merging phase. At least, I haven't heard of the matrix having started yet.

It would be nice if the factorization was finished by SHARCS 09 in early September. I've no idea whether that is possible,

Paul

Are you all talking about M1061?

2. Prime Net page says that F14 has been tested 9404 curves with B1 = 110 million. So, are we fairly sure that there are no factors for it atleast upto the 50 digit level mark?

3. For all my past SNFS candidates,
I choose a .poly file like this (with GGNFS)
Here is the one for 6,343+

n: 26922772671993809456879907269709743469032275674527231737993761996560079676334743186443456819701189925726786903389849007268675914915973679835240588245224019849525209125823880436459795415367623
m: 134713546244127343440523266742756048896
c6: 1
c5: -1
c4: 1
c3: -1
c2: 1
c1: -1
c0: 1
skew: 1

type: snfs
rlim: 40000000
alim: 40000000
lpbr: 29
lpba: 29
mfbr: 58
mfba: 58
rlambda: 2.6
alambda: 2.6
qintsize: 100000

In the .job file,

rlim: 40000000
alim: 10000000
lpbr: 29
lpba: 29
mfbr: 58
mfba: 58
rlambda: 2.6
alambda: 2.6
q0: 10000001
qintsize: 999999
#q1:11000000

For example, if I am sieving on the algebraic side, within the q0 value = 10M, GGNFS does not allow the value of alim > 10M. Similarly, also for the rational side too.

Otherwise, I will rather get an error like this...
F:\6_343P>"gnfs-lasieve4I14e.exe" -k -o spairs99.out -v -n0 -a 6,343+.job
Special q lower bound 10000001 below FB bound 4e+007

Does the lower value of alim or rlim have any consequence in lowering the rate of relations that are accumulated?

4. Where can I get gnfs-lasieve4I15e.exe from? Does this executive, have any effect at all, if so, for helping up within increasing the rate of yield of relations, that are being gathered so far, thereby increasing up so, the throughput, thus?

[10metreh]

It's not M1061, it's the (harder) RSA768 (I think).

15e? There are copies all over the place. Where did you get 14e from?

[Raman]

Originally Posted by 10metreh

15e? There are copies all over the place. Where did you get 14e from?

14e was available from ggnfs suite at sourceforge.net along within the archive when I downloaded it up at that time. (early November 2006, to be exact)

Seems that I should check again at that place, or atleast search for the postings all over this forum for finding out and then getting 15e, thus.

(Does it improve the rate of yield for getting the relations?)

[10metreh]

Originally Posted by Raman

(Does it improve the rate of yield for getting the relations?)

Depends on the size of the number. It sieves more for each special-q, with the result that there are less duplicates, so even if the speed is slightly lower but there are more relations per q, it may be an idea to use 15e.

[henryzz]

Originally Posted by Raman

14e was available from ggnfs suite at sourceforge.net along within the archive when I downloaded it up at that time. (early November 2006, to be exact)

Seems that I should check again at that place, or atleast search for the postings all over this forum for finding out and then getting 15e, thus.

(Does it improve the rate of yield for getting the relations?)

is suspected as much
there are much newer versions(SVN versions not official releases but almost everyone uses them(your the only one i remember using them recently))
download them from http://gilchrist.ca/jeff/factoring/
you wont have to bother manually lowering the factor base bound with the new binaries as it does it automatically