C208_124_107 (15e) factored
 

124 hours to solve a 12.5M matrix using -t 4, w/ TD=136.
[CODE]p72 factor: 267334869844255969519671159489493536778014788561751429286114214298677333
p137 factor: 16217889812789483905608630327898136888641850755006409303847873215768945002467084682896646682958962902963089931703732754668219847298474393[/CODE]
[url]https://pastebin.com/3AuvHezX[/url]

130_121 is done.

L1541 complete
 

[code]
Mon Sep 18 08:28:54 2017 p59 factor: 24572721564342894443628333904991213494122501676911797198531
Mon Sep 18 08:28:54 2017 p119 factor: 52087648482887152059927693557818086077317342653765194430601196177094018313728235984233018849831070102864030668470782279
[/code]

229.2 hours on 7 cores E5-2650v2 for 20.85M density-130 (132 didn't work) matrix. Log attached and at [url]https://pastebin.com/kZmGTbHg[/url]

Taking L3795a, ETA 26 September

C168_11040_10047 completed - 37 hours on 6.5M matrix (TD=140).

[CODE]prp60 factor: 491603674812510082226272789729122345012757309242463575729529
prp108 factor: 335100558920055696578425554604534422765141390009127644899659908857715878388956356577572131424574905751694487[/CODE][URL]https://pastebin.com/EkJUqW12[/URL]

C164_142_95 completed - 28.5 hours on 5.7M matrix (TD=140).

[CODE]prp62 factor: 20760201695936308939892001143689750992739762375605352068278831
prp103 factor: 1584326367334797141238866046488685093621775889979686456722229705173469907516345331732370833183022919813[/CODE][URL]https://pastebin.com/XvrxAK1u[/URL]

frmky is willing to run 135_128, which solves a bit of our resourcing problem at 15e.

148_83 Progress
 

Currently crunching in LA, 15946799 of 49068517 dimensions (32.5%, ETA 2809h12m). So it should finish ~15th of January.

I am more than willing to take another 15e/33 job once 148_83 is factored.

Thank you!

The problem with the enormous 15e jobs isn't so much finding willing crunchers, as that they tie up the largest fastest computers of the most dedicated crunchers for months at a time, and so reduce disproportionately the effort available for the routine 14e jobs - any of the 15e/33 jobs represents about the same amount of compute as [i]all[/i] the outstanding 14e jobs.

And we pretty much have to keep adding 14e jobs at the rate we have been, because there's a lot of BOINC effort that wants to crunch on them. I think, after my current personal SNFS270, that I'll need to take four 14e jobs at a time and run them on my cluster.

No worries. But you raise a good point - my biggest/best machine will only accomplish 2 post processing jobs in a year. Which means it's not doing 10-20 smaller jobs in the same year.

As such, I've tried personally to avoid suggesting any more 15e/33 jobs. Maybe next summer after all the current big jobs are completed. And then only if I'm willing to run the postprocessing as well. My mess, my cleanup...

Reserving 113990869481^19-1 cofactor.

C190_3372531985651_17 is delayed since 4GB of memory on my i3 was not enough for the matrix, and my 8GB machine is finishing a C153 today.

[QUOTE=fivemack;468171]And we pretty much have to keep adding 14e jobs at the rate we have been, because there's a lot of BOINC effort that wants to crunch on them. I think, after my current personal SNFS270, that I'll need to take four 14e jobs at a time and run them on my cluster.[/QUOTE]

I'll be making a dent on the 14e queue, as I just completed a GNFS-172 factorization on my home machine that began 9 July.

I accidentally set that as 15e/33 job rather than the intended 15e/32; it took 529M raw relations to build a 9.0M matrix at density 104. The matrix built on the first try (I use the factmsieve.py script). alim was 90M, sieved Q something like 22 to 82M. I was out of disk space for any 14e jobs, so I deleted the relations before considering that I could have tried some other filtering settings (such as fewer relations, higher TD) to get some more 33LP data. Next time...

Another data point to suggest LP bounds don't have much effect on matrix size.