[quote=siew;191867]Hi!

Trying to factor C143 number,

it it possible to split working between 4 PC?
8x4core Q6600 think can improve time consumption, but how to do this correctly?

I set Threads=4 on each PC, but what about dividini relation' space?

tanx[/quote]
It would be much more efficient to use the GNFS or SNFS (General or Special Number Field Sieve) instead of QS for such a large number.
[url]http://gilchrist.ca/jeff/factoring/nfs_beginners_guide.html[/url]

[quote=siew;191867]Hi!

Trying to factor C143 number,

it it possible to split working between 4 PC?
8x4core Q6600 think can improve time consumption, but how to do this correctly?

I set Threads=4 on each PC, but what about dividini relation' space?

tanx[/quote]

As mini-geek says, GNFS would be much more suitable for this size number. My biggest job using yafu was 130 digits, and I definately wasted a bunch of cpu time doing that. It took 50 CPUs about 4 days to complete the factorization. You're number is about 16 times harder so with 16 cores you're looking at about 200 days, assuming everything works fine (for numbers this size, you are definately in uncharted territory for yafu).

[quote=rekcahx;191857][CODE]
Bsquared, I just factored 120024621595864025218910923338006229301425391924739107060397489618540260099717126712633721435874689
to 2925384400770993343049563645260155764299463512577 and 41028666716152310208039084132300909697472499651457
[/CODE][COLOR=#000000]in few hours with Yafu 1.12. Thank you.[/COLOR]
[/quote]

You're very welcome! :smile: Nice one!

I can't compile yafu 1.12 under 32bit linux, I guess it's because the assembly code has to be fixed for 32bit linux, am I right?

[quote=BigBrother;193733]I can't compile yafu 1.12 under 32bit linux, I guess it's because the assembly code has to be fixed for 32bit linux, am I right?[/quote]

Yeah, that's right. I have an older version which compiles under 32bit linux, but ever since google ported my page over to google sites, I haven't been able to edit it. If I can't get it fixed soon I can PM you the older version... maybe in the next version I can get it working for 32bit systems again.

I finally put the 64bit windows version of YAFU online here:
[url]http://gilchrist.ca/jeff/factoring/[/url]

I also sent Ben a copy so it should be appearing on the official site soon I would imagine.

[FONT=Arial Narrow]echo "siqs(196647532279496299)" | yafu[/FONT]
reproducibly blows up on 64-bit Linux:
[CODE]10/28/09 01:28:12 [B]v1.12[/B] @ ya,
Initializing with Tom's Fast Math (x86-64 asm)...
cached 78504 primes. pmax = 1000099
detected cpu 9, with L1 = 65536 bytes, L2 = 6291456 bytes

starting pQS on c(18): 196647532279496299
s_init = 0, pmax = 337, cutoff = 10110(30 * pmax)
inner_block_size = 32768, blocksize = 65536
n = 18 digits, 58 bits
B = 40, T = 1.10, k = 1
base closnuf = 30
232(40) relations found: 78 full + 154 from 459 partial, using 1 blocks
matrix exhausted
Total pQS time = 0.0100 seconds.
*** glibc detected *** yafu: corrupted double-linked list: 0x000000000057cfb0 ***
======= Backtrace: =========
/lib64/libc.so.6[0x2aeb05fd234e]
/lib64/libc.so.6[0x2aeb05fd351d]
/lib64/libc.so.6(__libc_free+0x6c)[0x2aeb05fd395c]
yafu[0x412e4d]...
[/CODE]
1.10 doesn't.

[QUOTE=Batalov;194049][FONT=Arial Narrow]echo "siqs(196647532279496299)" | yafu[/FONT]
reproducibly blows up on 64-bit Linux:[/QUOTE]

Same thing happens with the 64k version of both 32bit and 64bit Windows binaries.

The 32k version of 32bit and 64bit Windows works fine.

Jeff.

Thanks guys, I'll look into it.

[Edit]
Fixed now. I have some other changes cooking, so unless there is violent opposition, I'll delay releasing the fixed version until those are done (hopefully a week or so at most).

I frequently see a inconsistent digit count, like in this example:
[CODE]10/29/09 10:27:52 v1.12 @ LAPTOP,
Initializing with Tom's Fast Math (x86-32 asm)...
cached 78504 primes. pmax = 1000099
detected cpu 6, with L1 = 32768 bytes, L2 = 3145728 bytes


starting SIQS on [B][U]c81[/U][/B]: 244319173756229060069606235012489045211851322430046929300276036276182797256294079

==== sieve params ====
n = [B][U]82 digits[/U][/B], 270 bits[/CODE]
Very often the difference between these two is 1, sometimes it is 2. I never noticed this in previous versions of yafu.

Here's an example with difference 2:

[CODE]10/29/09 11:08:25 v1.12 @ LAPTOP,
Initializing with Tom's Fast Math (x86-32 asm)...
cached 78504 primes. pmax = 1000099
detected cpu 6, with L1 = 32768 bytes, L2 = 3145728 bytes


starting SIQS on [B][U]c82[/U][/B]: 7223116501689193712141700949217785000864313256971575304576231537274155910783453251

==== sieve params ====
n = [B][U]84 digits[/U][/B], 278 bits[/CODE]