[QUOTE=Xyzzy;323101]We are unable to tell when we are doing work. It affects our system in no way. But even the slightest lag would end our involvement, since we have but one computer now that is used for everything. Maybe we have been lucky so far?[/QUOTE]

No issues here running a matrix using up to 6 GB of memory out of 8GB, I can even burn DVD's.

[quote]Hey, wait, and before I forgot, Lionel had implemented server side remdups before the sun had set on his server. Wouldn't it be a good thing to restore?[/quote]
I had performed [i]manual[/i] server side remdups + recompression with pbzip2, so as to save some space on the poor little server's HDD, and thereby reduce the need for client starvation :wink:
For that task, the server needed some help from another, much more powerful server.

[QUOTE=Xyzzy;323101]We are unable to tell when we are doing work. It affects our system in no way. But even the slightest lag would end our involvement, since we have but one computer now that is used for everything. Maybe we have been lucky so far?[/QUOTE]

[QUOTE=pinhodecarlos;323103]No issues here running a matrix using up to 6 GB of memory out of 8GB, I can even burn DVD's.[/QUOTE]

Honestly I'm not quite sure what he's talking about when he says that running a matrix lags out anything. Since it doesn't quite scale with all the cores, when I run a matrix with more than two cores a significant chunk of the cpu time doesn't get used, so my system doesn't notice. (On the other hand, when I have ggnfs running on all 8 threads (4 cores) and Msieve crunching on the GPU, then the system sometimes gets a bit laggy :razz:)

May I officially reserve 3617523089023m19 ?

I'm a little confused about the discrepancy between the crunching_e.php page which suggests 225.6M relations in and 13.9M to go, and the .projection file which ends

[code]
20121231075302 size_before=14591969541 size_after=14597657159 size_projected=19931528690.522 rels_projected=308023958.37219[/code]

Should I believe that it's 6% from completion or 27% ?

The website site pending relations includes only the pending wu's, not the unsent ones. Trust the projection file.

Made it before the bell
 

L2605B:

[code]
prp78 factor: 326312963122859945773965691756113369362672409304215716326093460828942516760151
prp113 factor: 24952794953038334198693944758466070268026541031735878238803227663365543663601321624944847257002234583119777331571
[/code]

525 minutes to download 13.8GB of relations. 320 minutes on 1 CPU to convert 223382555 relations (189259297 unique) into an 8543247 x 8543424 (3701.5 MB) matrix with weight 1064391185 (124.59/col) using target_density=110. 37 hours on 24 CPUs to solve the matrix, 70 min x8 for square roots

Filtering is now taking a reasonable amount of time.
24*37 is 888 and filtering took 320. What is your efficiency running on 24 cpus? If you ran on one would it take longer than the filtering?

[QUOTE=henryzz;323244]Filtering is now taking a reasonable amount of time.
24*37 is 888 and filtering took 320. What is your efficiency running on 24 cpus? If you ran on one would it take longer than the filtering?[/QUOTE]

Filtering took 320 minutes, not 320 hours!

[QUOTE=fivemack;323264]Filtering took 320 minutes, not 320 hours![/QUOTE]

:blush:

[QUOTE=Batalov;322437]Anyone postprocessing 853_83_minus1?[/QUOTE]
Done.
[CODE]p95 factor: 27779063012872420716125396618815439819446735659065155051488605598176567403468417928581085278719
p98 factor: 80411248691669309986541006535173080167262195334009677659750550111739597533416651286215368163950371
[/CODE]

I would like to reserve 11411_108