Current Effort

[Batalov]

I have now built a nice 8Gb Phenom940 machine at home, and will be willing to help anyone with LA of any size.

--Serge
__________

AMD Phenom II X4 940 3.0GHz
GIGABYTE GA-MA790GP-UD4H
Patriot Viper 8GB DDR2/1066
openSUSE 11.1

[fivemack]

weather.com tells me it's 20 degrees centigrade cooler in Cambridge than in Chennai at the moment.

The animation suggests you ought to have been doing the linear algebra 120 million years ago, when the place Chennai is now was in reasonably chilly waters somewhere between current Capetown and current Antarctica ... though the paleoclimatologists reckon that everything was then three degrees hotter than it is now.

[jasonp]

Starting the LA 120 million years ago means you would be using a 80386, or something...

[Andi47]

Originally Posted by jasonp

Starting the LA 120 million years ago means you would be using a 80386, or something...

...or SNFSing a really BIG number - and having started sieving in the early Silurian...

[Raman]

I have just sent the DVD by post. Saturday and Sunday were holidays (weekends) and I have just sent it today (Monday). I think that it will take a week to reach you. Inform me as soon as you receive it, Safely.

Snail mail (post) was the cheapest (only Rs 51) and others were extremely costly: Speed post: Rs 478, Courier: > Rs 2500. A courier from Trichy to Chennai is hardly Rs 20. The above are the international rates.

I wish that I were born away from the equator, where it is much cooler. Not in extreme conditions like the Sahara desert, or in the ice caps of polar regions. Somewhere in the temperate latitudes, where it is not vulnerable to earthquakes and tornadoes.

Venus is constant in temperature throughout the equator to the polar regions, eventhough the polar regions receive very little sunlight. Why in Earth are the polar regions cooler? Venus has constant temperature of 475°C, generated by the strong greenhouse effect of thick atmosphere of CO₂ under 93 atmospheres pressure. Mercury and Mars have very thin atmosphere, owing to their low gravity. In Mercury, the day and night temperatures are at extreme. 350°C during the day, to -170°C during the night. Mercury is tidally locked to the Sun in a 3:2 ratio, as Moon is to Earth, Pluto to Charon, to each other.

Mr. Tom, what work are you doing? Are you a university professor? How much money do you get as income? In my opinion, the courier and speed post rates to foreign are very expensive.

Did you have enough money to buy your resources (computers) for this contribution, before you started earning? I know that you would have only bought your resources after acquiring enough money, but I can't wait to do so.

[Raman]

Originally Posted by fivemack

Tue Mar 17 01:33:47 2009
Tue Mar 17 01:33:47 2009
Tue Mar 17 01:33:47 2009 Msieve v. 1.39
Tue Mar 17 01:33:47 2009 random seeds: 539cc0f0 794d2fe0
Tue Mar 17 01:33:47 2009 factoring 43530842794654712369936752314716583228552850852880185876967\
53593002074624697573400982036418055827897146317140063377813885557624999850231373576930085296121\
779853543537188567707144180933943323891018955967519960133495049773807349 (226 digits)
Tue Mar 17 01:33:51 2009 no P-1/P+1/ECM available, skipping
Tue Mar 17 01:33:51 2009 commencing number field sieve (226-digit input)
Tue Mar 17 01:33:51 2009 R0: 134713546244127343440523266742756048896
Tue Mar 17 01:33:51 2009 R1: -1
Tue Mar 17 01:33:51 2009 A0: 1
Tue Mar 17 01:33:51 2009 A1: 1
Tue Mar 17 01:33:51 2009 A2: 1
Tue Mar 17 01:33:51 2009 A3: 1
Tue Mar 17 01:33:51 2009 A4: 1
Tue Mar 17 01:33:51 2009 A5: 1
Tue Mar 17 01:33:51 2009 A6: 1
Tue Mar 17 01:33:51 2009 skew 0.00, size 5.844430e-11, alpha 2.427706, combined = 2.920803e-11
Tue Mar 17 01:33:52 2009
Tue Mar 17 01:33:52 2009 commencing relation filtering
Tue Mar 17 01:33:52 2009 commencing duplicate removal, pass 1
...
Tue Mar 17 01:56:16 2009 found 24832201 hash collisions in 78058702 relations
Tue Mar 17 01:56:40 2009 added 2438526 free relations
Tue Mar 17 01:56:43 2009 commencing duplicate removal, pass 2
Tue Mar 17 02:08:22 2009 found 32176118 duplicates and 48321110 unique relations
Tue Mar 17 02:08:22 2009 memory use: 504.8 MB
Tue Mar 17 02:08:22 2009 reading rational ideals above 48300032
Tue Mar 17 02:08:22 2009 reading algebraic ideals above 48300032
Tue Mar 17 02:08:22 2009 commencing singleton removal, pass 1
Tue Mar 17 02:25:05 2009 relations with 0 large ideals: 1765513
Tue Mar 17 02:25:05 2009 relations with 1 large ideals: 6731846
Tue Mar 17 02:25:05 2009 relations with 2 large ideals: 14527106
Tue Mar 17 02:25:05 2009 relations with 3 large ideals: 15203739
Tue Mar 17 02:25:05 2009 relations with 4 large ideals: 7077019
Tue Mar 17 02:25:05 2009 relations with 5 large ideals: 993805
Tue Mar 17 02:25:05 2009 relations with 6 large ideals: 65624
Tue Mar 17 02:25:05 2009 relations with 7+ large ideals: 1956458
Tue Mar 17 02:25:05 2009 48321110 relations and about 36157107 large ideals
Tue Mar 17 02:25:05 2009 commencing singleton removal, pass 2
Tue Mar 17 02:40:31 2009 found 9155556 singletons
Tue Mar 17 02:40:31 2009 current dataset: 39165554 relations and about 26059891 large ideals
Tue Mar 17 02:40:31 2009 commencing singleton removal, pass 3
Tue Mar 17 02:53:35 2009 found 2277363 singletons
Tue Mar 17 02:53:35 2009 current dataset: 36888191 relations and about 23730296 large ideals
Tue Mar 17 02:53:35 2009 commencing singleton removal, pass 4
Tue Mar 17 03:06:31 2009 found 457039 singletons
Tue Mar 17 03:06:31 2009 current dataset: 36431152 relations and about 23270935 large ideals
Tue Mar 17 03:06:31 2009 commencing singleton removal, final pass
Tue Mar 17 03:22:04 2009 memory use: 625.1 MB
Tue Mar 17 03:22:04 2009 commencing in-memory singleton removal
Tue Mar 17 03:22:07 2009 begin with 36431152 relations and 28311956 unique ideals
Tue Mar 17 03:22:57 2009 reduce to 29751154 relations and 21354874 ideals in 15 passes
Tue Mar 17 03:22:57 2009 max relations containing the same ideal: 28
Tue Mar 17 03:23:12 2009 reading rational ideals above 720000
Tue Mar 17 03:22:57 2009 max relations containing the same ideal: 28
Tue Mar 17 03:23:12 2009 reading rational ideals above 720000
Tue Mar 17 03:23:12 2009 reading algebraic ideals above 720000
Tue Mar 17 03:23:12 2009 commencing singleton removal, final pass
Tue Mar 17 03:41:43 2009 keeping 25051358 ideals with weight <= 20, new excess is 2111890
Tue Mar 17 03:42:20 2009 memory use: 950.6 MB
Tue Mar 17 03:42:21 2009 commencing in-memory singleton removal
Tue Mar 17 03:42:24 2009 begin with 29751154 relations and 25051358 unique ideals
Tue Mar 17 03:43:04 2009 reduce to 29741126 relations and 25041329 ideals in 9 passes
Tue Mar 17 03:43:04 2009 max relations containing the same ideal: 20
Tue Mar 17 03:43:23 2009 removing 2954431 relations and 2554431 ideals in 400000 cliques
Tue Mar 17 03:43:25 2009 commencing in-memory singleton removal
Tue Mar 17 03:43:28 2009 begin with 26786695 relations and 25041329 unique ideals
Tue Mar 17 03:44:12 2009 reduce to 26529422 relations and 22224804 ideals in 11 passes
Tue Mar 17 03:44:12 2009 max relations containing the same ideal: 20
Tue Mar 17 03:44:28 2009 removing 2184109 relations and 1784109 ideals in 400000 cliques
Tue Mar 17 03:44:30 2009 commencing in-memory singleton removal
Tue Mar 17 03:44:33 2009 begin with 24345313 relations and 22224804 unique ideals
Tue Mar 17 03:45:05 2009 reduce to 24197120 relations and 20290182 ideals in 9 passes
Tue Mar 17 03:45:05 2009 max relations containing the same ideal: 20
Tue Mar 17 03:45:20 2009 removing 1943955 relations and 1543956 ideals in 400000 cliques
Tue Mar 17 03:45:22 2009 commencing in-memory singleton removal
Tue Mar 17 03:45:24 2009 begin with 22253165 relations and 20290182 unique ideals
Tue Mar 17 03:45:47 2009 reduce to 22128147 relations and 18619161 ideals in 7 passes
Tue Mar 17 03:45:47 2009 max relations containing the same ideal: 20
Tue Mar 17 03:46:01 2009 removing 1815035 relations and 1415035 ideals in 400000 cliques
Tue Mar 17 03:46:02 2009 commencing in-memory singleton removal
Tue Mar 17 03:46:04 2009 begin with 20313112 relations and 18619161 unique ideals
Tue Mar 17 03:46:25 2009 reduce to 20195841 relations and 17084920 ideals in 7 passes
Tue Mar 17 03:46:25 2009 max relations containing the same ideal: 20
Tue Mar 17 03:46:37 2009 removing 1738310 relations and 1338310 ideals in 400000 cliques
Tue Mar 17 03:46:38 2009 commencing in-memory singleton removal
Tue Mar 17 03:46:40 2009 begin with 18457531 relations and 17084920 unique ideals
Tue Mar 17 03:47:03 2009 reduce to 18341656 relations and 15628732 ideals in 9 passes
Tue Mar 17 03:47:03 2009 max relations containing the same ideal: 20
Tue Mar 17 03:47:14 2009 removing 1194919 relations and 931788 ideals in 263131 cliques
Tue Mar 17 03:47:16 2009 commencing in-memory singleton removal
Tue Mar 17 03:47:17 2009 begin with 17146737 relations and 15628732 unique ideals
Tue Mar 17 03:47:34 2009 reduce to 17087462 relations and 14636860 ideals in 7 passes
Tue Mar 17 03:47:34 2009 max relations containing the same ideal: 20
Tue Mar 17 03:47:49 2009 relations with 0 large ideals: 225914
Tue Mar 17 03:47:49 2009 relations with 1 large ideals: 828225
Tue Mar 17 03:47:49 2009 relations with 2 large ideals: 2657347
Tue Mar 17 03:47:49 2009 relations with 3 large ideals: 4650894
Tue Mar 17 03:47:49 2009 relations with 4 large ideals: 4678615
Tue Mar 17 03:47:49 2009 relations with 5 large ideals: 2680257
Tue Mar 17 03:47:49 2009 relations with 6 large ideals: 820877
Tue Mar 17 03:47:49 2009 relations with 7+ large ideals: 545333
Tue Mar 17 03:47:49 2009 commencing 2-way merge
Tue Mar 17 03:48:04 2009 reduce to 11561801 relation sets and 9111199 unique ideals
Tue Mar 17 03:48:04 2009 commencing full merge
Tue Mar 17 03:51:29 2009 memory use: 920.3 MB
Tue Mar 17 03:51:31 2009 found 5890292 cycles, need 5555399
Tue Mar 17 03:51:35 2009 weight of 5555399 cycles is about 389138759 (70.05/cycle)
Tue Mar 17 03:51:35 2009 distribution of cycle lengths:
Tue Mar 17 03:51:35 2009 1 relations: 706060
Tue Mar 17 03:51:35 2009 2 relations: 644338
Tue Mar 17 03:51:35 2009 3 relations: 653166
Tue Mar 17 03:51:35 2009 4 relations: 610995
Tue Mar 17 03:51:35 2009 5 relations: 558915
Tue Mar 17 03:51:35 2009 6 relations: 495446
Tue Mar 17 03:51:35 2009 7 relations: 429284
Tue Mar 17 03:51:35 2009 8 relations: 366327
Tue Mar 17 03:51:35 2009 9 relations: 308453
Tue Mar 17 03:51:35 2009 10+ relations: 782415
Tue Mar 17 03:51:35 2009 heaviest cycle: 18 relations
Tue Mar 17 03:51:36 2009 commencing cycle optimization
Tue Mar 17 03:51:51 2009 start with 29803717 relations
Tue Mar 17 03:53:09 2009 pruned 863376 relations
Tue Mar 17 03:53:09 2009 memory use: 963.5 MB
Tue Mar 17 03:53:09 2009 distribution of cycle lengths:
Tue Mar 17 03:53:09 2009 1 relations: 706060
Tue Mar 17 03:53:09 2009 2 relations: 659583
Tue Mar 17 03:53:09 2009 3 relations: 679958
Tue Mar 17 03:53:09 2009 4 relations: 631349
Tue Mar 17 03:53:09 2009 5 relations: 579578
Tue Mar 17 03:53:09 2009 6 relations: 507995
Tue Mar 17 03:53:09 2009 7 relations: 438949
Tue Mar 17 03:53:09 2009 8 relations: 368345
Tue Mar 17 03:53:09 2009 9 relations: 304130
Tue Mar 17 03:53:09 2009 10+ relations: 679452
Tue Mar 17 03:53:09 2009 heaviest cycle: 18 relations
Tue Mar 17 03:53:32 2009
Tue Mar 17 03:53:32 2009 commencing linear algebra
Tue Mar 17 03:53:36 2009 read 5555399 cycles
Tue Mar 17 03:54:17 2009 cycles contain 15793938 unique relations
Tue Mar 17 04:06:27 2009 read 15793938 relations
Tue Mar 17 04:07:03 2009 using 20 quadratic characters above 536870658
Tue Mar 17 04:09:23 2009 building initial matrix
Tue Mar 17 04:16:45 2009 memory use: 1932.5 MB
Tue Mar 17 04:17:01 2009 read 5555399 cycles
Tue Mar 17 04:19:24 2009 matrix is 5555005 x 5555399 (1661.6 MB) with weight 477636769 (85.98/\
col)
Tue Mar 17 04:19:24 2009 sparse part has weight 374475569 (67.41/col)
Tue Mar 17 04:22:09 2009 filtering completed in 3 passes
Tue Mar 17 04:22:12 2009 matrix is 5537507 x 5537707 (1659.2 MB) with weight 476861544 (86.11/\
col)
Tue Mar 17 04:22:12 2009 sparse part has weight 374026443 (67.54/col)
Tue Mar 17 04:24:56 2009 read 5537707 cycles
Tue Mar 17 04:25:01 2009 matrix is 5537507 x 5537707 (1659.2 MB) with weight 476861544 (86.11/\
col)
Tue Mar 17 04:25:01 2009 sparse part has weight 374026443 (67.54/col)
Tue Mar 17 04:25:01 2009 saving the first 48 matrix rows for later
Tue Mar 17 04:25:06 2009 matrix is 5537459 x 5537707 (1579.4 MB) with weight 386205899 (69.74/\
col)
Tue Mar 17 04:25:06 2009 sparse part has weight 358655749 (64.77/col)
Tue Mar 17 04:25:06 2009 matrix includes 64 packed rows
Tue Mar 17 04:25:06 2009 using block size 10922 for processor cache size 256 kB
Tue Mar 17 04:26:33 2009 commencing Lanczos iteration (4 threads)
Tue Mar 17 04:26:33 2009 memory use: 1684.9 MB
Fri Mar 20 13:16:43 2009 lanczos halted after 87572 iterations (dim = 5537459)
Fri Mar 20 13:17:00 2009 recovered 38 nontrivial dependencies
Fri Mar 20 13:17:05 2009
Fri Mar 20 13:17:05 2009 commencing square root phase
Fri Mar 20 13:17:05 2009 reading relations for dependency 1
Fri Mar 20 13:17:07 2009 read 2768513 cycles
Fri Mar 20 13:17:24 2009 cycles contain 9738988 unique relations
Fri Mar 20 13:31:40 2009 read 9738988 relations
Fri Mar 20 13:33:18 2009 multiplying 7890728 relations
Fri Mar 20 14:06:40 2009 multiply complete, coefficients have about 190.65 million bits
Fri Mar 20 14:06:50 2009 initial square root is modulo 6948443
Fri Mar 20 14:47:43 2009 reading relations for dependency 2
Fri Mar 20 14:47:49 2009 read 2769545 cycles
Fri Mar 20 14:48:06 2009 cycles contain 9742205 unique relations
Fri Mar 20 15:02:46 2009 read 9742205 relations
Fri Mar 20 15:04:24 2009 multiplying 7893884 relations
Fri Mar 20 15:37:40 2009 multiply complete, coefficients have about 190.72 million bits
Fri Mar 20 15:37:50 2009 initial square root is modulo 6989111
Fri Mar 20 16:18:38 2009 reading relations for dependency 3
Fri Mar 20 16:18:54 2009 read 2767726 cycles
Fri Mar 20 16:19:12 2009 cycles contain 9740170 unique relations
Fri Mar 20 16:34:03 2009 read 9740170 relations
Fri Mar 20 16:35:41 2009 multiplying 7891052 relations
Fri Mar 20 17:08:50 2009 multiply complete, coefficients have about 190.65 million bits
Fri Mar 20 17:09:00 2009 initial square root is modulo 6948847
Fri Mar 20 17:49:41 2009 prp61 factor: 2546288496284990213389598514399701583273729812805688674719861
Fri Mar 20 17:49:41 2009 prp166 factor: 1709580153944291167278033766583519088276840877200520825653946078713332910510582579693073239474491548002030379354944737248786446095996982486052301124094778854470292609
Fri Mar 20 17:49:41 2009 elapsed time 88:15:54

Update: Sieving for 6,343+ is 14% complete.

Tue Mar 17 04:25:06 2009 using block size 10922 for processor cache size 256 kB

How is your processor L2 cache size as low as 256 kB?
My laptop (Core 2 Duo) cache size is 1 MB and that of Core 2 Quad desktop is 4 MB.
Each of the two systems has 2 GB RAM.
My laptop is 1.73 GHz and that of desktop is 2.4 GHz.

@ fivemack: Can you give details about:
-> Your processor speed, and processor family/configuration (?) Core 2 Quad/Duo, Pentium 4, like that.
-> Your physical memory (RAM) size (I assume that you use Windows)
-> Your true processor cache size, if it is not as low as 256 kB.

@ bdodson: How many computers do you have? In 2007, you found out an ECM factor almost everyday, how much computers do you run ECM upon? Plus, you know that you contribute some machines to the NFSNET project, right? How many systems for that case?
Finally, how many computers do you use so for the other big Number Field Sieve Polynomial Selection/Sieving/Linear Algebra jobs?

[fivemack]

Hi Raman.

That calculation was done on a Core i7 920, running probably at 2800MHz (the Linux cpu-speed-control mechanism and the speed-control mechanism built into the i7 interact in ways I don't quite understand yet); the four cores have 256k L2 cache each and share an 8192k L3 cache, but msieve 1.39 doesn't detect the L3. The machine has 12GB of DDR3/800 memory and runs Ubuntu Linux 8.10.

[Raman]

Thanks for doing so the Linear Algebra.

Originally Posted by Raman

I am still not confident of whether the DVDs have reached safely in the form that I sent... so just for verification purposes... here are the details...

I sent two DVDs
Data on one is 4.35 GB, and that of other is 4.34 GB.

The first disk contains two folders, Core2 and Core3, and the second disk contains the folders, Core1 and Core4, along with some files related to GGNFS and msieve.

The spairs.out are the relation files in the GGNFS/msieve format.
The folder Core1 contains 97 such files.
Core2 contains 104 files.
Core3 contains 81 files.
Core4 contains 77 files.

So, you start the matrix job, you would have to merge all these 359 files into some 6_343M.dat file, along with some file msieve.n containing
N c226
at the first line, beginning of the file.

I assume that this is what you did so... May I know of the fact that...
How large and dense is the matrix?

Oh, okay. But my relation set had 80687051 relations (see above), but yours had only 78058702? Almost 2.6 million lesser?

Did you receive all the 359 spairs.out files safely, or some files got corrupted? How many files got corrupted and how big are they? You told only one small file, but that contributes to as much as 2.6 million relations? I am curious to know about it up, only so, thus.

[jasonp]

Originally Posted by fivemack

the four cores have 256k L2 cache each and share an 8192k L3 cache, but msieve 1.39 doesn't detect the L3.

Intel's recommended method for detecting cache sizes guarantees that your code will guess wrong for processors that are very recent. Msieve v1.40 should detect the L3 size correctly.

[fivemack]

Raman: core3/spairs38.out was corrupted.

Your 80.6M figure includes free relations (IE you must have run msieve once and then added more relations to the end of the msieve.dat file), mine doesn't. Since there were 2.4 million free relations, that makes the numbers add up correctly.

[Raman]

Actually, I thought that I could entirely be proud of myself if I was able to do so the matrix job by myself. But, I couldn't. My laptop has been heating too much, and I needed continuous power supply for a period of atleast 10 days, to do so the linear algebra continuously, starting up right from the scratch of filtering.

Also that once I restarted it up, the matrix didn't fit up right within 2 GB RAM + 3 GB virtual memory size, so I had to give up. I may try to increase the available memory size later on, say so for the next number 6,343+, and then thus try it up by myself, by using my resources, finally.

I decided to send the relations by using the post service, because of the fact that you were curious for solving the matrix, and that I wanted to check how the post service works, for sending the DVD by using the service of post, for the verification purposes.