Combined sieving discussion

[japelprime]

What happend when PSP-group members reserve range from SOB and submit through PSP-group. Now I am starting reserving range from Automatic Sieve Range Reservation and I have always submitt sieve ranges for SOB through PSP. Shall I continuing doing so ? I don´t know how often Lars sends the faktors from PSP and I wonder maybe it will take more than this "sieve time limit" for SOB having theyr faktors back. Do I need to tell SOB when I finish my ranges so they don´t send them out again ?
I like to do things easyest way and submitt SOB range through PSP is nice.

[Joe O]

Originally Posted by Joe O

This is just the reservation system for combined ranges above 1200000. How you submit factors remains the same.

Originally Posted by Joe O

How you submit factors remains the same,
only the reservation method changes.

Originally Posted by japelprime

What happend when PSP-group members reserve range from SOB and submit through PSP-group. Now I am starting reserving range from Automatic Sieve Range Reservation and I have always submitt sieve ranges for SOB through PSP. Shall I continuing doing so ? I don´t know how often Lars sends the faktors from PSP and I wonder maybe it will take more than this "sieve time limit" for SOB having theyr faktors back. Do I need to tell SOB when I finish my ranges so they don´t send them out again ?
I like to do things easyest way and submitt SOB range through PSP is nice.

Yes, you do need to tell the Combined Automated Reservation System that your range is finished, just as you needed to post here when your range was finished.
How you submit factors remains the same. If you want to have PSP-group to recieve the SoB credit, then just (continue to) click on

SOB factors reported by PSP

and LTD will send them in at the end of the month to be credited to PSP-group.

[dspdude2000]

I am a relative newcomer, was doing some manual sieving, but am now on BOINC. I am wondering: is there a release schedule for new versions of the combined dat file. As far as I can tell, we have eliminated at least 1% of the current dat file (released in October). Maybe a monthly release of a new dat file would allow us to gain some speed?

[VJS]

dspdude2000, I

I don't think it makes that much of a difference with the new client. I think the next update will be with the next prime.

[hhh]

I can confirm that. Speed increases come only from elimination of "k", not from "n". H.

[benjackson]

To repeat what was said, even the last "k" we eliminated only really netted about a 3% speed gain and that was combined with eliminating a ton of "n"s for the other "k"s too. So a monthly update that eliminated "n"s would really be a tiny fraction of a percent increase in speed and probably not worth deploying to some of the bigger setups.

[dspdude2000]

Thanks for the explanation! Good to know that people are thinking about these things.

[VJS]

dspdude,

I think your comment was probably based upon early early early renditions of the dat. Once upon a time when Joe and I were working at less than 50T we are currently at 1000T+. There were speed improvements from reducing dat size but at that time we would eliminate more factors betweeen 3T and 4T than psp currently eliminates in a month or two.

I wish it were as easy as eliminating a few n to see a speed increase, .

[KriZp]

The only real advantage of getting a new dat right now would be for SOB factoring, less need to check http://www.henleyclan.co.uk/sobsieve/2008/ui/19999.htm and http://www.henleyclan.co.uk/sobsieve/2007/ui/19999.htm for recently submitted factors. I've manually removed 3-4 candidates from the worktodo since I started factoring, and still I managed to test and find a factor for a candidate primegrid had discovered just days before.

[Joe O]

You will notice that the the k n pairs in worktodo format files have been updated today. I try to do this on a monthly basis.

http://www.free-dc.org/forum/showthr...403#post113403

[hhh]

Strange: I just got this input:

User: hhh **SOB Flag: PSP_Only
Input contained 13 total lines.
Input contained 12 valid lines.
Input contained 12 valid lines from PSP k values.
Input contained 0 valid lines from SOB k values.
Input contained 0 invalid lines.
Input contained 0 lines with unknown K.
Input contained 1 lines with already known factor.
Input contained 0 wrong factors.

Who submitted the factor in my range? The factors were:

1492075134762137 | 152267*2^22887483+1
1492088869453037 | 90527*2^49970783+1
1492171548617863 | 237019*2^25008226+1
1492187381850631 | 168451*2^3841236+1
1492187381850631 | 168451*2^3841236+1
1492323441848251 | 168451*2^4577076+1
1492343818482971 | 90527*2^26019959+1
1492358408880583 | 258317*2^3536679+1
1492382190867397 | 79817*2^21157439+1
1492398890888833 | 237019*2^15931690+1
1492432652093759 | 90527*2^1559999+1
1492492933658029 | 79309*2^43978862+1