Help needed for doublecheck sieving!

[mdettweiler]

Hi all,

A while back, Gary asked me to coordinate some doublechecking for the range 100K<n<260K for the NPLB range of k's (300<k<1001). I've started the sieving and taken it to p=10G, but since Gary estimates optimal depth at p=550G, it is way more than I can handle, with only one machine that isn't on 24/7. So, I've decided to open it up to the members of the NPLB project. This will serve as a reservation/status thread for the effort; to claim a range, simply post a reply and specify what range you're taking.

Reservations of all sizes are welcome--feel free to take whatever you feel you can complete in a reasonable amount of time. (Please give status reports at least every two weeks, preferably more frequently).

Update: We have now expanded this sieving effort to include k<300 (for the same n-range). Chris (Flatlander) and I (Anonymous) started the sieve for k<300, and got it caught up to this one (at p=130G). I've now combined the two sieve files; thus, any reservations past 130G will be for the combined sieve file. (I've updated the link below to point to the new combined sieve file.)

The sieve file can be downloaded here. It contains one file, "sr2data.txt", which is our sieve file. Extract it to the folder containing your copy of sr2sieve (if you don't have the latest version already, get it here). Then, make a new text file called "sr2work.txt". Open it in a text editor, and enter your range on the first line, like in this example (for 10G-20G):

10-20

You can queue up additional ranges by adding more lines, each containing another range in that same format.

When your range is done, email the factors.txt file to me at bugmesticky@googlemail.com. Then, report the range as complete here in this thread, and, if you wish, reserve another range and start the process over again!

Status of reserved ranges:

Range      Reserved by     Status
----------------------------------
550G-600G  gd_barnes       complete
310G-550G  (various)*      complete
300G-310G  gd_barnes       complete
270G-300G  Kman1293        complete
130G-270G  Flatlander      complete
----sieve file combined w/k<300 here----
120G-130G  Anonymous       complete
80G-120G   Flatlander      complete
70G-80G    Anonymous       complete
50G-70G    Flatlander      complete
30G-50G    AES             complete
20G-30G    Flatlander      complete
10G-20G    Anonymous       complete
0-10G      Anonymous       complete

*This range was partially done in a hodgepodge manner by jasong. Gaps were filled by me (Anonymous), Mini-Geek, and IronBits; the complete list of gaps can be found here.

Sieving is complete! Thanks to all who helped out! If you'd like to help out with the LLRing please see the Doublecheck Drive #1 thread.

Let's get rolling!

[mdettweiler]

I'll take 10G-20G to get us started.

[Flatlander]

How long do you estimate a 10G range would take on a single 3GHz C2D core?

If I reboot the machine will it find its place and continue?

[mdettweiler]

Originally Posted by Flatlander

How long do you estimate a 10G range would take on a single 3GHz C2D core?

If I reboot the machine will it find its place and continue?

I would guess that it would take around 2-3 days, running 24/7.

Yes, it will remember its place. Simply press Ctrl-C to interrupt srsieve before you shut your computer down, and then when you start back up, just run the batch file/shell script again. srsieve will read from the checkpoint.txt file that it's saved (it saves it every five minutes, so even if your computer is shut down abruptly, it will still not lose much work), and pick up from there.

[Flatlander]

I'll take 20G-30G.

Do I need the "...i686" srsieve? (XP 32 bit, C2D)

[mdettweiler]

Originally Posted by Flatlander

I'll take 20G-30G.

Do I need the "...i686" srsieve? (XP 32 bit, C2D)

Yes, you'll want to get i686. I think the "i686" means that it has a few more optimizations included than the "i586" version does, but most (if not all) modern processors (i.e. P4 and later) support the i686 optimizations, so yes, you'll want the i686 version.

Thanks for helping out!

[Flatlander]

Originally Posted by Flatlander

How long do you estimate a 10G range would take on a single 3GHz C2D core?

I will take about 24hrs if I don't run anything else.

[mdettweiler]

Originally Posted by Flatlander

I will take about 24hrs if I don't run anything else.

Okay, great! That's quicker than I expected! Are you running it on both cores, or just one?

[Flatlander]

Just on one core at 3013MHz.

[mdettweiler]

Originally Posted by Flatlander

Just on one core at 3013MHz.

Okay. I'm using a 2.2Ghz Core 2 Duo, so it shouldn't take too much longer. Looks like my estimate of 2-3 days of CPU time was a little off--but I'm not complaining, since it should actually take less time than that.

[AES]

Is there a reason you're not using sr2sieve?

I'll take 30e9 to 50e9