Sieving almost done...

[frmky]

Looking at the stats, we have about 8 million relations to go to complete the required 82 million. Since the beginning of February, NFSNet's been averaging over 1 million relations per day...even if we ignore the three recent near 2 million days. So it looks like sieving will be done in about a week or so! Sieving for a record factorization in 4.5 months! Nice job!

Greg

[Wacky]

Originally Posted by frmky

NFSNet's been averaging over 1 million relations per day...even if we ignore the three recent near 2 million days.

Don't ignore them. We expect them to continue for the next week.

Originally Posted by frmky

So it looks like sieving will be done in about a week or so!

Actually, we are in the process of doing an analysis of the data in hand to revise the estimate of 82 million relations. My preliminary thought is that we may need to add a few more. I'm currently guessing 84 million, but will have a better idea after tonight's filter run finishes. Unless my estimates are off, we might start switching projects as early as Sunday.

I encourage everyone who is running on Microsoft Windows systems to upgrade to the version 1.0 release if they have not done so. (If you are running in the "Cantab" pool, you have not done the upgrade.)

[xilman]

Originally Posted by Wacky

Unless my estimates are off, we might start switching projects as early as Sunday.

The next project will be to factor the Cunningham Most Wanted Number 10,223+.c201. This is the 201-digit cofactor of 10^223+1. A few small factors are already known.

One of my machines began sieving a few thousand lines yesterday so that we can estimate how much effort will be needed to complete the factorization. First results suggests that about 25 million lines will be enough and that this will require about 9000 WU of effort. More precise figures will be available when we've analyzed the results of the survey sieving.

Paul

[junky]

Originally Posted by xilman

One of my machines began sieving a few thousand lines yesterday so that we can estimate how much effort will be needed to complete the factorization. First results suggests that about 25 million lines will be enough and that this will require about 9000 WU of effort. More precise figures will be available when we've analyzed the results of the survey sieving.
Paul

only 25 M lines and 9000 WUs ? is a line takes the same time to complete ? if yes, that shouldn't take so long to complete

that's great that's ya already started work on the next project, in this case, we won't lose CPU cycles :)

so give us some details when ya'll finish analyzing the results of the survey sieving.

later

[xilman]

Originally Posted by junky

only 25 M lines and 9000 WUs ? is a line takes the same time to complete ? if yes, that shouldn't take so long to complete

that's great that's ya already started work on the next project, in this case, we won't lose CPU cycles :)

so give us some details when ya'll finish analyzing the results of the survey sieving.

later

Still surveying.

Yes, the next project should rather less effort than the current one.

Paul

[frmky]

Originally Posted by Wacky

I'm currently guessing 84 million, but will have a better idea after tonight's filter run finishes.

The stats say we're currently at 86.5 million. Apparently the filter run deemed a few more necessary. Looking forward to the switch soon though since the next target is a nice entry for the database of factors of base-10 repunits that I maintain (and need to update...but anyway).

Greg

[junky]

i saw the whole project is currently at 95.76% completed.
so its good.

keep us in touch about the surveying.

[xilman]

Originally Posted by frmky

The stats say we're currently at 86.5 million. Apparently the filter run deemed a few more necessary. Looking forward to the switch soon though since the next target is a nice entry for the database of factors of base-10 repunits that I maintain (and need to update...but anyway).

Greg

We've switched!

We discovered overnight that we have enough relations to proceed with the next phases of the computation and so switched to the next project, 10_223P, at around noon UT.

I'm snowed under at the moment, so more later...


Paul