Need For Speed II - NFS@Home Challenge

[pinhodecarlos]

The (N)eed (F)or (S)peed II Challenge issued by SETI.USA team and will start from 2015-03-15 00:00 UTC to 2015-03-20 00:00 UTC.

Stats can be followed here: http://boincstats.com/en/stats/challenge/team/chat/652

Please note that this is a 5 days challenge with wu deadline of 7 days meaning that people will start bunkering 6-7 days before the start of the challenge. So 8-9 days before we need to feed the applications with work:

14e Lattice Sieve - cache size of at least 70k wu's
15e Lattice Sieve - cache size of at least 70k wu's
16e V5 Lattice Sieve - cache size of at least 400k wu's

Carlos

[pinhodecarlos]

I expect more than 3000 cores to be added before challenge start. SETI.USA has more than 2000 cores by itself. LAF (Biggest French team) will also be on the challenge.

[pinhodecarlos]

Tom, are you reading this? I need guarantee that we will have plenty of 14e and 15e work for the challenge.
I will chase Greg for 16e V5 work.

Thank you,

yoyo and Paul,

What is the ECM status on the GCW candidates?

Carlos

[xilman]

Originally Posted by pinhodecarlos

yoyo and Paul,

What is the ECM status on the GCW candidates?

I'll check it out.

Paul

[Batalov]

Homogeneous Cunningham numbers have >> t50.

One can start from the bottom of this list and go up. The polynomials are easily built with this; you can build a hundred polys in less than a minute and because they are all in the same size class, tag the constant limit part at the bottom:

rlim: 120000000
alim: 120000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 62
rlambda: 2.6
alambda: 2.6

and they are ready to sieve -r from 10M to 100-120-140...M or maybe less. Someone should spend an hour or so trial-sieving them in 3-5-digit buckets (one random poly from any bucket), and use the same parameters on the bucket..

[debrouxl]

While t50 + several hundreds of curves at B1=11e7 yields low enough risk of severe ECM misses (p4x factors), it still leaves a significant risk of p50-p55 factors on SNFS 240+ numbers (which is what I'd rather queue to 14e for the challenge), which would raise some eyebrows in the factoring community

Any chance we could get at least 10K curves at B1=11e7 (15K for a full t55 if you wish) on several Homogeneous Cunningham numbers of SNFS difficulty 242-250 in time for the challenge ?
Say, for instance:
* first batch: 7^293-5^293 , 7^293-6^293 , 7^293-2^293 and 7^293-3^293;
* second batch: 7^293+2^293 , 7^293+3^293 and 7^293+4^293.

[Batalov]

Originally Posted by debrouxl

Any chance we could get at least 10K curves at B1=11e7 (15K for a full t55 if you wish) on several Homogeneous Cunningham numbers of SNFS difficulty 242-250 in time for the challenge ?

This could be a good time to call yoyo for synergetic ECMing.

[R.D. Silverman]

Originally Posted by Batalov

This could be a good time to call yoyo for synergetic ECMing.

I ran 1000 curves with B1 = 500M.

[fivemack]

Carlos: at what point would it make sense to move all the 15e targets from 'queued for sieving' to 'sieving' in preparation for this Need For Speed challenge?

[pinhodecarlos]

Originally Posted by fivemack

Carlos: at what point would it make sense to move all the 15e targets from 'queued for sieving' to 'sieving' in preparation for this Need For Speed challenge?

8 days before the start of the challenge.

[pinhodecarlos]

Tom, please start thinking for the next three days to:

1. move all the 15e targets from 'queued for sieving' to 'sieving' in preparation for the challenge
2. move all the 14e targets from 'queued for sieving' to 'sieving' in preparation for the challenge
3. increase unsent 16e Lattice Sieve V5 wus to 250k.

Also if you get in touch with Greg please tell him to set:

<max_wus_in_progress> 100 </max_wus_in_progress>

Currently is set to 50.

Thank you.

Carlos