Archived sieve reservations and discussions

[gribozavr]

n_480000-483000__k_100000_200000__19apr2011.txt
http://www.sendspace.com/file/rvuanx

n_486300-489999__k_1_100000__19apr2011.txt
http://www.sendspace.com/file/xegj3t

Combined sieve file is being uploaded.

[cipher]

oddball i tried combining them and try sieving... it won't work. :( let me know if you get different results.

[Blackwood]

Originally Posted by cipher

i tried combining them and try sieving... it won't work. :(

Try removing the header from the second file.

Or, you can wait for gribozavr to finish uploading the combined file.

[gribozavr]

480000-490000_19apr2011.zip
http://www.sendspace.com/file/d2o6kr

[cipher]

Output

C:\CUDA\binfull>tps -i 490f.txt -p 6010e12 -P 6020e12 -m 270336 -Q 10e6
tpsieve version-0.2.3 (testing)
Found K's from 3 to 9999999.
Found N's from 480000 to 490000.
nstart=480000, nstep=30
Read 32086513 terms from NewPGen format input file `490f.txt'
Changed nstep to 29
tpsieve initialized: 3 <= k <= 9999999, 480000 <= n < 490000
malloc: Not enough space

[Lennart]

Originally Posted by cipher

Output

I got the same problem here.

I am now testing CPU version.


Lennart

[Oddball]

Originally Posted by cipher

oddball i tried combining them and try sieving... it won't work. :( let me know if you get different results.

I tried it on a CPU and couldn't get it to work either:

tpsieve version 0.3.4 (testing)
Compiled Sep 3 2009 with GCC 4.4.0
nstart=489990, nstep=30
Read 32086513 terms from NewPGen format input file `48.txt'
tpsieve initialized: 3 <= k <= 9999999, 480000 <= n < 489999
Sieve started: 6002000000000000 <= p < 6003000000000000
malloc: No such file or directory

The sieve runs fine if the n=485K-490K file is used. FYI, this was done using a Core 2 Duo at 3GHz with 3.25 GB of RAM using Windows XP. The sieve speed for the n=485K-490K file using one core is 29M p/sec.

I'll try the combined file on a Phenom II later today and see how it goes.

[Lennart]

I am running on a i7 970 with 12GB Ram. No problem on CPU but I get the same error as chiper on my GPU GTX 570

btw I am using ~4.2GB ram when I sieve

Lennart

[Oddball]

Originally Posted by Oddball

I'll try the combined file on a Phenom II later today and see how it goes.

Didn't work. I had 8GB of RAM and used the 64 bit version of Windows 7, and I got:

tpsieve version 0.3.4 (testing)
Compiled Sep 3 2009 with GCC 4.4.0
nstart=489990, nstep=30
malloc: Not enough space

Both the 480K-485K file and the 485K-490K file work perfectly on this machine.

[mdettweiler]

3136T-3168T (range 2) complete, 9911 factors: http://www.sendspace.com/file/n95c1u

This one missed the rally cutoff by a long shot. I seriously overestimated the speed of this machine when picking out my original range, and since my only access to it is remote, it wasn't convenient to upload a portion of the range to get credit for the rally.

Gary, if you're reading this, now I see exactly what you mean about your first-generation AMD Phenoms being complete and utter stinkers at sieving! (And at LLRing, too... )

[gribozavr]

Originally Posted by Oddball

Didn't work. I had 8GB of RAM and used the 64 bit version of Windows 7, and I got:

Did you use 64 bit version of tpsieve?

1505T-1700T finished.

Taking 3685T-3900T.