Need help with a virus tales-from-the-crypt 2.0 - Msieve

[darkskysofrenia]

hello,

i factor keys for a friend and i'm surprised, my computer factor the first one ( 154 integers digits ) in 58seconds, my friend send me another one ( always 154 interger digits ) and the same computer( dual Xeon X5670 --> 12cores/24Threads@3.06Ghz ) is working on it scince 3 days ...

I thought that what is making the difficulty/time to compute was the number of digits, could you explain me what's make the difficulty , how to estimate the time needed, etc ...

is there a way to estimate the time remaring on factmsieve.py in verbose mode actualy lines are 3 sequences of digits like 8520 5464315465 324684635164, i see the first series goes 12 by 12 , the next one should be 8532 XXXXXXX XXXXXXXXXXXX

thanks for help

[wombatman]

Originally Posted by dearnizam

Hi Jux,

Thanks alot for help and your factor number has helped me to recover my files of D drive and when i tried to recover files of C drive this is coming with different public key.

Im trying to get factor for this.

Im very much thankfull to Wombatman,Jux and all the members of this group.

Regards
Nizamuddin

Glad to hear it's working! If you run into more issues, don't hesitate to ask!

[wombatman]

This one: http://www.factordb.com/index.php?id...00000811949725

is now fully factored. I forget who it belonged to.

[wombatman]

Originally Posted by darkskysofrenia

hello,

i factor keys for a friend and i'm surprised, my computer factor the first one ( 154 integers digits ) in 58seconds, my friend send me another one ( always 154 interger digits ) and the same computer( dual Xeon X5670 --> 12cores/24Threads@3.06Ghz ) is working on it scince 3 days ...

I thought that what is making the difficulty/time to compute was the number of digits, could you explain me what's make the difficulty , how to estimate the time needed, etc ...

is there a way to estimate the time remaring on factmsieve.py in verbose mode actualy lines are 3 sequences of digits like 8520 5464315465 324684635164, i see the first series goes 12 by 12 , the next one should be 8532 XXXXXXX XXXXXXXXXXXX

thanks for help

The rule of thumb I know of is every 5 digit increase in the length of the composite (assuming it's only got 2 factors) roughly doubles the time to factor it.

For factmsieve.py, which step are you on? It seems like it is the polynomial selection step. If you have a CUDA-capable graphics card (Nvidia), that will vastly speed that part up.

[chris2be8]

Originally Posted by darkskysofrenia

hello,

i factor keys for a friend and i'm surprised, my computer factor the first one ( 154 integers digits ) in 58seconds, my friend send me another one ( always 154 interger digits ) and the same computer( dual Xeon X5670 --> 12cores/24Threads@3.06Ghz ) is working on it scince 3 days ...

It would help if you told us what the two numbers are. But my best guess is that the first number had several small factors and one large one. The small factors were quickly found by ECM, which was lucky. The second number presumably has two large factors, which yafu is now using gnfs on. It should succeed but will take several more days.

Chris

[VictordeHolland]

Originally Posted by darkskysofrenia

hello,

i factor keys for a friend and i'm surprised, my computer factor the first one ( 154 integers digits ) in 58seconds, my friend send me another one ( always 154 interger digits ) and the same computer( dual Xeon X5670 --> 12cores/24Threads@3.06Ghz ) is working on it scince 3 days ...

The first one you factored probably had a lot of small factors, which were found with ECM. The second one probably has no or little small factors and the (larger) remaining composite has to be factored by GNFS.

One of the weaknesses of Teslacrypt is the useage of these easily factorable composites.

[darkskysofrenia]

Originally Posted by darkskysofrenia View Post
the first one ( easy factored ) was this one :


the second one still running on my computer scince approximatively thuersday 31.12.2015 20h00 is :

0x0BD5B4F3CB4D761648C9BD9AA4A03E570D14294E8499A5F9F05D7574278A1D5A79D5A19D1BEB34B196D0040B9E25EB253C9EB45FDAC4EFB6AFB9C10EEBFE8FD2

actualy ( 48h later ) ..... http://www.hostingpics.net/viewer.ph...319encours.jpg

[VictordeHolland]

Originally Posted by darkskysofrenia

the second one still running on my computer scince approximatively thuersday 31.12.2015 20h00 is :

0x0BD5B4F3CB4D761648C9BD9AA4A03E570D14294E8499A5F9F05D7574278A1D5A79D5A19D1BEB34B196D0040B9E25EB253C9EB45FDAC4EFB6AFB9C10EEBFE8FD2

actualy ( 48h later ) ..... http://www.hostingpics.net/viewer.ph...319encours.jpg

Are you sieving with all cores? The CPU load on most cores is non-existent??

[darkskysofrenia]

Yes all cores are enabled in msieve, but the 100% load on 24 cores was only during the first step ( i don't remember the name of the step ). i've try other smaller factorisation for compare how does it works , and i've try other ways ( YAFU and others ) even if i'm totaly newbie in factorissation i'm curious to see how it's work and how the computer "handle" that

for now i found YAFU is a lot faster , but on this one he failed , probaby a configuration problem

i had launch another msieve but directly from the .exe not by the factmsieve.py on the same number but he failed too :


so on this famous KEY the factmsieve.py was the longer one but the only one with no fail for now....

supens... hope factorisation done soon

[Dubslow]

The last error is saying that Msieve failed to allocate 1GB of memory.

[darkskysofrenia]

strange ... the computer had 24Gb of ram and last time i check at least 19Gb free ... and it's ECC registerd memory ( normaly more efficent for computaion )