Team sieve #36: c162 from 4788:i5141

jrk · 2013-01-21, 01:42

Code:

# aq4788:5141
# sieve with gnfs-lasieve4I14e on -a side from Q=25M to 90M
n: 887205201940910880791354660610517788880512334137737151840425477543967036683766158494641120546745694229868822924877909737118697675474613950106536548327853898707953
# norm 2.053e+22 alpha -7.784 e 1.310e-12 rroots 3
skew: 5051040.10
c5: 362700
c4: 6387646856661
c3: 53624339725773447374
c2: -196675516426754720128099634
c1: -434152361498392616833556546575606
c0: 649980606796839564654107169565430572545
Y1: 21238410121093243597
Y0: -18953682035996144770438879648918
rlim: 50000000
alim: 50000000
lpbr: 30
lpba: 30
mfbr: 60
mfba: 60
rlambda: 2.5
alambda: 2.5

___________________________________________________________________

Dubslow:

I've commandeered this post for organizational purposes.

Code:

Edh                                80-90M ✓
                                   60-64M ✓
                                   58-60M ✓
                                   46-48M ✓
RichD                              76-80M ✓
                                   64-70M ✓
                                   48-50M ✓
Dubslow                            70-76M ✓
                                   30-36M ✓
                                   50-58M ✓
                                   38-44M
                                   44-46M ✓
Andi_HB                            29-30M ✓
jrk                                25-29M ✓
                                   36-38M ✓

free:                              none

EdH · 2013-01-22, 17:02

There doesn't seem to be a great deal of interest in this project at the moment. I don't think I'd be capable of doing post-proc, but I can add some relations. I've started the collection with 88-90M, which will be finished soon, I think.

How much RAM is needed for this c162?

schickel · 2013-01-22, 18:49

Quote:

Originally Posted by EdH

There doesn't seem to be a great deal of interest in this project at the moment. I don't think I'd be capable of doing post-proc, but I can add some relations. I've started the collection with 88-90M, which will be finished soon, I think.

How much RAM is needed for this c162?

I could do the post on this. RAM should be less than 2GB, but I can give you a closer estimate when I start the post on a c166 that I just now finished sieving....

EdH · 2013-01-23, 03:00

Quote:

Originally Posted by schickel

I could do the post on this. RAM should be less than 2GB, but I can give you a closer estimate when I start the post on a c166 that I just now finished sieving....

I might give it a test run if 2GB would be enough.

Something went amiss with one of my programs here and I have a gap in my 88M-90M region that I need to fill in. I should be able to have it (and the offending program) squared away by tomorrow. I'll run 86-88M after that. With a small turnout, I imagine this will be a lengthy factorization. We'll see if my attention span is that long...

schickel · 2013-01-23, 09:15

Quote:

Originally Posted by EdH

I might give it a test run if 2GB would be enough.

Something went amiss with one of my programs here and I have a gap in my 88M-90M region that I need to fill in. I should be able to have it (and the offending program) squared away by tomorrow. I'll run 86-88M after that. With a small turnout, I imagine this will be a lengthy factorization. We'll see if my attention span is that long...

Looks like I'm barely sieved enough. My ETA (c166) was 551 hours with one thread; it's ~450 hours with 2 threads (forgot to start it with 2 threads, but it looks like I might want to move to my larger machine and run with 4 threads anyway....) Task manager shows memory use as 3,996,160K, so closer to 4GB....here's what the matrix looks like:

Code:

commencing linear algebra
matrix starts at (0, 0)
matrix is 8992231 x 8992410 (2753.9 MB) with weight 855926921 (95.18/col)
sparse part has weight 614016717 (68.28/col)
saving the first 48 matrix rows for later
matrix includes 64 packed rows
matrix is 8992183 x 8992410 (2660.2 MB) with weight 689773601 (76.71/col)
sparse part has weight 607419179 (67.55/col)
using block size 21845 for processor cache size 512 kB
commencing Lanczos iteration (2 threads)
memory use: 2213.8 MB

What's really disappointing is that, barring a lucky ECM factor, I have a c168 sitting in my factoring queue.

EdH · 2013-01-23, 16:10

Quote:

Originally Posted by schickel

Looks like I'm barely sieved enough. My ETA (c166) was 551 hours with one thread; it's ~450 hours with 2 threads (forgot to start it with 2 threads, but it looks like I might want to move to my larger machine and run with 4 threads anyway....) Task manager shows memory use as 3,996,160K, so closer to 4GB....here's what the matrix looks like:

Code:

...

What's really disappointing is that, barring a lucky ECM factor, I have a c168 sitting in my factoring queue.

Bummer! I guess I can't do it here, anyway. Unless, I can figure out openmpi... msieve has mpi code in it for the LA. I just don't know anything about implementing it. Right now I'm running other things like ECM and sieving across several machines via binaries and scripts that I wrote, that are very specific to my machines.

Good Luck on your c168...

Dubslow · 2013-01-23, 16:36

I can do the post proc, presuming the sieving takes longer than the SNFS 221 I'm currently doing.

EdH · 2013-01-23, 22:18

Quote:

Originally Posted by Dubslow

I can do the post proc, presuming the sieving takes longer than the SNFS 221 I'm currently doing.

Thanks for the offer. At the rate rels are being compiled, it will probably be a while... Actually, if my attention doesn't wane and I can settle out some variations between machines here so they're a little more efficient, I might be able to do the sieve range in 2-3 weeks. I started high so others could grab the low end, but this sequence isn't all that interesting right now.

schickel · 2013-01-23, 23:16

Quote:

Originally Posted by EdH

Thanks for the offer. At the rate rels are being compiled, it will probably be a while... Actually, if my attention doesn't wane and I can settle out some variations between machines here so they're a little more efficient, I might be able to do the sieve range in 2-3 weeks. I started high so others could grab the low end, but this sequence isn't all that interesting right now.

If you do sputter out, post here and I'm sure one of us will pick it up; though 2-3 weeks is better than my time: 11/21-1/22 for sieving...

schickel · 2013-01-23, 23:24

Quote:

Originally Posted by schickel

My ETA (c166) was 551 hours with one thread; it's ~450 hours with 2 threads (forgot to start it with 2 threads, but it looks like I might want to move to my larger machine and run with 4 threads anyway....)

Wow, in this case AMD sure beats out Intel!

Quote:

Originally Posted by msieve

linear algebra completed 605163 of 8992410 dimensions (6.7%, ETA 116h22m)

RichD · 2013-01-24, 04:20

It looks like yoyo@home will run out of curves in a day or two. I'll pick up a few M Qs, say 76-80M unless someone thinks I should do a different range. I'll check back in less than 24 hours.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Team Sieve #28: c162 from 4788:2714	jrk	Aliquot Sequences	27	2011-10-25 08:51
Team sieve #23: c172 from 4788:i2617	schickel	Aliquot Sequences	64	2011-02-19 02:28
Team sieve #21: c162 from 4788:2602	jrk	Aliquot Sequences	31	2010-12-30 21:33
Team sieve #20: c170 from 4788:2549	schickel	Aliquot Sequences	153	2010-11-09 07:39
Team sieve #5: c140 from 4788:2407	10metreh	Aliquot Sequences	77	2009-05-27 20:39

2013-01-22, 17:02	#2
EdH "Ed Hall" Dec 2009 Adirondack Mtns 11×347 Posts	There doesn't seem to be a great deal of interest in this project at the moment. I don't think I'd be capable of doing post-proc, but I can add some relations. I've started the collection with 88-90M, which will be finished soon, I think. How much RAM is needed for this c162?

2013-01-23, 16:36	#7
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 16065₈ Posts	I can do the post proc, presuming the sieving takes longer than the SNFS 221 I'm currently doing.

2013-01-24, 04:20	#11
RichD Sep 2008 Kansas 2⁴·211 Posts	It looks like yoyo@home will run out of curves in a day or two. I'll pick up a few M Qs, say 76-80M unless someone thinks I should do a different range. I'll check back in less than 24 hours.