| Author |
Message |
|
ShadowPhoenix
Newbie
Joined: 19 Aug 2009
Posts: 20
|
 Posted: 20 Aug 2009 11:55:56 pm Post subject: |
|
|
FloppusMaximus wrote: Small ideals aren't better or worse than large ones (well, OK, they're better if they make it easier to find relations.) My point was just that Msieve is expecting a different distribution of ideals than what we have. I don't think this is a major problem, but it is a rather complicated business, and I can't claim to understand it all.
The main goal is simply for us to have more relations than ideals (once that happens, we can build a matrix, solve for linear dependencies, and hopefully, derive a factorization from one of those dependencies.) A secondary goal is for the matrix to be as small as possible (so that it's easier to solve.) The singleton removal phase serves both of these purposes, by removing unhelpful relations. You'll notice that Msieve performs singleton removal in two steps; in the first step, it only looks at "large" ideals above some limit (e.g., 40 million); in the second step, it adds smaller (but still rather large) ideals into the mix. But it always ignores ideals smaller than 7.2 million, and it also ignores ideals that occur very frequently; these ideals figure into the "target excess". So as we add more relations, the target excess will continue to increase (though, hopefully, the actual excess will increase faster.) At any rate, that's the best I can gather from reading bits of the Msieve source. 
Very helpful!
I am also reading various documentation...
http://mersenneforum.org/showthread.php?t=10495
I wonder if we can attract the author to our cause |
|
| Back to top |
|
|
squalyl
Advanced Newbie
Joined: 04 Aug 2009
Posts: 57
|
| Posted: 21 Aug 2009 01:45:38 am Post subject: |
|
|
(lost my quote)
The 46M-50M range is still in the works, do you think you will need the 50-55M range? |
|
| Back to top |
|
|
dga
Newbie
Joined: 20 Aug 2009
Posts: 7
|
| Posted: 21 Aug 2009 08:06:31 am Post subject: |
|
|
ShadowPhoenix wrote: @dga:
Thank you so much for your work! I can do everything needed for final computation on yours (if you don't mind) if you can provide me with a user/IP. All that's needed is download some data from HTTP, uncompress, cat it together, download msieve, make msieve (no admin rights needed, it gcc's it in place), and then run msieve.
Or I can make a python script to do all that so you can inspect source code . Whichever way you prefer.
Unfortunately, I'll have to go the python script route. The cluster is behind a wee bit of a firewall, and I don't have a great way of giving VPN access to just the cluster nodes. |
|
| Back to top |
|
|
squalyl
Advanced Newbie
Joined: 04 Aug 2009
Posts: 57
|
|
| Back to top |
|
|
FloppusMaximus
Advanced Member
Joined: 22 Aug 2008
Posts: 472
|
| Posted: 21 Aug 2009 04:59:52 pm Post subject: |
|
|
Great! Another 2 million might be enough, but if they aren't, at least we should get a better idea of how close we are.
Just to be clear, the "ti73" column in your chart is the 0102 (Flash app) key, correct? I'm still searching for better polynomials for the TI-73 OS key (nothing great yet, I found one with score 2.48e-12.) |
|
| Back to top |
|
|
jasonp
Newbie
Joined: 21 Aug 2009
Posts: 4
|
| Posted: 21 Aug 2009 06:37:02 pm Post subject: |
|
|
ShadowPhoenix wrote: I wonder if we can attract the author to our cause
You guys don't need my help at this point :)
Happy factoring. |
|
| Back to top |
|
|
Graphmastur
Advanced Member
Joined: 25 Mar 2009
Posts: 360
|
| Posted: 21 Aug 2009 07:00:14 pm Post subject: |
|
|
jasonp wrote: ShadowPhoenix wrote: I wonder if we can attract the author to our cause
You guys don't need my help at this point :)
Happy factoring.
Cool!!! It's the author!!! I am guessing the author of the boinc project? Or msieve? |
|
| Back to top |
|
|
FloppusMaximus
Advanced Member
Joined: 22 Aug 2008
Posts: 472
|
| Posted: 21 Aug 2009 07:43:07 pm Post subject: |
|
|
jasonp wrote: You guys don't need my help at this point :)
Happy factoring.
Thanks! :)
Graphmastur: jasonp is the author of Msieve.
ShadowPhoenix: when you have a chance, could you give the filtering another try? |
|
| Back to top |
|
|
dga
Newbie
Joined: 20 Aug 2009
Posts: 7
|
| Posted: 21 Aug 2009 09:03:18 pm Post subject: |
|
|
jasonp wrote: ShadowPhoenix wrote: I wonder if we can attract the author to our cause
You guys don't need my help at this point :)
Happy factoring.
Oh goodness. Now you've got both a pi generating maniac and a pi searching maniac reading your boards...  |
|
| Back to top |
|
|
FloppusMaximus
Advanced Member
Joined: 22 Aug 2008
Posts: 472
|
| Posted: 21 Aug 2009 09:45:42 pm Post subject: |
|
|
I've just noticed that the 22M-25M range (sieved by Skittle) is missing from squalyl's server. I'm uploading it now. This will probably make a significant difference!
On a side note, I've had at least 3 power outages in the past few days. This makes it a little bit difficult to run extended computations. |
|
| Back to top |
|
|
jasonp
Newbie
Joined: 21 Aug 2009
Posts: 4
|
| Posted: 21 Aug 2009 10:01:40 pm Post subject: |
|
|
dga wrote: Oh goodness. Now you've got both a pi generating maniac and a pi searching maniac reading your boards...
That feels like a lifetime ago now...
Factoring the first key generated a *lot* of online press, and I don't think anyone else has hooked up the sieving tools to BOINC. If you run out of calculator keys, mersenneforum has a thriving factoring community. |
|
| Back to top |
|
|
frmky
Newbie
Joined: 22 Aug 2009
Posts: 5
|
| Posted: 22 Aug 2009 12:52:43 pm Post subject: |
|
|
| You've got a lot of sieving in progress here. Let me know if you need help with the postprocessing needed to obtain the factors. |
|
| Back to top |
|
|
ShadowPhoenix
Newbie
Joined: 19 Aug 2009
Posts: 20
|
| Posted: 22 Aug 2009 02:59:40 pm Post subject: |
|
|
FloppusMaximus wrote: jasonp wrote: You guys don't need my help at this point :)
Happy factoring.
Thanks! :)
Graphmastur: jasonp is the author of Msieve.
ShadowPhoenix: when you have a chance, could you give the filtering another try?
Yep, going to do that about now. |
|
| Back to top |
|
|
ShadowPhoenix
Newbie
Joined: 19 Aug 2009
Posts: 20
|
| Posted: 22 Aug 2009 04:24:56 pm Post subject: |
|
|
oh shit. Good news is that we got enough.
Bad news:
mmap(size=37150720) failed (error code=12)
37 GIGABYTES?
Code:
start with 24230789 relations
pruned 608079 relations
memory use: 654.8 MB
distribution of cycle lengths:
1 relations: 573832
2 relations: 600349
3 relations: 615117
4 relations: 546922
5 relations: 494071
6 relations: 416825
7 relations: 359775
8 relations: 299232
9 relations: 244356
10+ relations: 515615
heaviest cycle: 16 relations
RelProcTime: 1635
commencing linear algebra
read 4666094 cycles
cycles contain 13843274 unique relations
read 13843274 relations
using 20 quadratic characters above 536870730
building initial matrix
memory use: 1655.0 MB
read 4666094 cycles
matrix is 4665679 x 4666094 (1333.9 MB) with weight 439682261 (94.23/col)
sparse part has weight 317015877 (67.94/col)
filtering completed in 3 passes
matrix is 4643382 x 4643582 (1331.2 MB) with weight 438529982 (94.44/col)
sparse part has weight 316470372 (68.15/col)
read 4643582 cycles
matrix is 4643382 x 4643582 (1331.2 MB) with weight 438529982 (94.44/col)
sparse part has weight 316470372 (68.15/col)
saving the first 48 matrix rows for later
matrix is 4643334 x 4643582 (1289.3 MB) with weight 350031019 (75.38/col)
sparse part has weight 310116423 (66.78/col)
matrix includes 64 packed rows
using block size 65536 for processor cache size 6144 kB
commencing Lanczos iteration (8 threads)
msieve(34226,0xa02c0720) malloc: *** mmap(size=37150720) failed (error code=12)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
failed to allocate 37148656 bytes
|
|
| Back to top |
|
|
frmky
Newbie
Joined: 22 Aug 2009
Posts: 5
|
| Posted: 22 Aug 2009 04:30:00 pm Post subject: |
|
|
ShadowPhoenix wrote: 37 GIGABYTES?
Code: failed to allocate 37148656 bytes
That's 37 megabytes. |
|
| Back to top |
|
|
ShadowPhoenix
Newbie
Joined: 19 Aug 2009
Posts: 20
|
| Posted: 22 Aug 2009 04:38:40 pm Post subject: |
|
|
frmky wrote: ShadowPhoenix wrote: 37 GIGABYTES?
Code: failed to allocate 37148656 bytes
That's 37 megabytes.
You are right. My fault! *thinks*
It shouldn't run out... 12 gigs were free before start....
Any way I could resume it? |
|
| Back to top |
|
|
frmky
Newbie
Joined: 22 Aug 2009
Posts: 5
|
| Posted: 22 Aug 2009 04:42:24 pm Post subject: |
|
|
Yes. Replace -nc with -ncr -nc3
Edit: If that doesn't work, use -nc2 -nc3
Last edited by Guest on 22 Aug 2009 04:53:43 pm; edited 1 time in total |
|
| Back to top |
|
|
ShadowPhoenix
Newbie
Joined: 19 Aug 2009
Posts: 20
|
| Posted: 22 Aug 2009 04:59:43 pm Post subject: |
|
|
frmky wrote: Yes. Replace -nc with -ncr -nc3
Edit: If that doesn't work, use -nc2 -nc3
Roger, will let you know how it goes soon.
EDIT>:
-ncr -nc3 returns "cannot open matrix checkpoint file."
EDIT2>:
-nc2 -nc3 now cannot allocate memory either :(
Maybe it's because this is 32 land? gah. I'll have the python script done then, unles someone has an idea.
Last edited by Guest on 22 Aug 2009 05:33:14 pm; edited 1 time in total |
|
| Back to top |
|
|
FloppusMaximus
Advanced Member
Joined: 22 Aug 2008
Posts: 472
|
| Posted: 22 Aug 2009 10:13:50 pm Post subject: |
|
|
That's a shame. I'm surprised - I know the linear algebra requires a lot of memory, but from what I remember last time, it didn't require nearly as much as the filtering. And the matrix in this case is smaller.
Anyway, it's great to see that we have enough relations now.
dga, would you be willing to lend us one of your machines for this? Or Skittle, is your machine still available? |
|
| Back to top |
|
|
frmky
Newbie
Joined: 22 Aug 2009
Posts: 5
|
| Posted: 23 Aug 2009 12:40:13 am Post subject: |
|
|
If no one else wishes to do it, I can finish it.
Edit: I've started it.
linear algebra completed 19613 of 4577051 dimensions (0.4%, ETA 85h30m)
Last edited by Guest on 23 Aug 2009 03:31:37 am; edited 1 time in total |
|
| Back to top |
|
|
|
Open Topic & United-TI Talk => General Open Topic
