Featured request

[bsquared]

Originally Posted by xilman

The true error rate is actually very much smaller than that.

Which is why I said "upper bounded by" earlier in the post ;)

Do we (i.e., anyone) know what the true "average" error rate is? I've only ever heard it referred to as "much much less than 4^-k", but never more concretely than that. Is it 8^-k, 50^-k, something else?

[jasonp]

See here for a more advanced treatment of the error probability. It depends on the number of tests that have passed so far and also on the size of the input.

[bsquared]

Originally Posted by jasonp

See here for a more advanced treatment of the error probability. It depends on the number of tests that have passed so far and also on the size of the input.

Great, thanks!

So, if I understand correctly, the equation at the end of section 5 says that for a random 300 bit number, 20 MR tests gives a probability of failure less than:

 
300^(3/2) * 2^20 / sqrt(20) * 4 ^ (2 - sqrt(20 * 300)) = 4.5e-37

Indeed, that is much much less than 1 in a trillion.

[Dubslow]

How about this: An option to factor/nfs to suppress msieve/polyselect output? There's so much... or, say, print every third line (coeff 35280 specialq 1 - 73245 other 10340 - 24816) or just

nfs: commencing polynomial search over range: 35251 - 35501
deadline: 100 CPU-seconds per coefficient

. Any of the above would significantly unclutter things.
Thanks!

[bsquared]

Originally Posted by Dubslow

How about this: An option to factor/nfs to suppress msieve/polyselect output? There's so much... or, say, print every third line (coeff 35280 specialq 1 - 73245 other 10340 - 24816) or just

nfs: commencing polynomial search over range: 35251 - 35501
deadline: 100 CPU-seconds per coefficient

. Any of the above would significantly unclutter things.
Thanks!

Already on the list, but I'm afraid there isn't much I can do about it, yet... see posts #2 and #10.

[jasonp]

Yeah, the tons of output have been in there for a while, and the Msieve source needs to be modified to remove them. Maybe leave them in only for degree 6 poly selection, which still gives a lot of trouble.

[Dubslow]

Is it possible to print total rels and/or unique rels in the middle of sieving/filtering? I realize I could just use msieve, but the nfs() provides the automated sieve/filter/sieve/f/s/f/s... that's so nice . (Regardless, I do plan on playing with msieve sometime soon, and learning lasieve for which there's no -h :P.)

P.S. Is it possible to print est. req. rels before/during sieving?

(P.P.S. Maybe I should just start my own thread so I don't clutter the "actual" thread :P)

[bsquared]

Originally Posted by Dubslow

Is it possible to print total rels and/or unique rels in the middle of sieving/filtering? I realize I could just use msieve, but the nfs() provides the automated sieve/filter/sieve/f/s/f/s... that's so nice . (Regardless, I do plan on playing with msieve sometime soon, and learning lasieve for which there's no -h :P.)

P.S. Is it possible to print est. req. rels before/during sieving?

(P.P.S. Maybe I should just start my own thread so I don't clutter the "actual" thread :P)

Have you run with -v? It will do most of what you are asking (all except uniq rels, which it can only know if it runs filtering, which we don't want to do often).

[Mathew]

Very low priority:

Add nfs capability to mac osx.
The poly creation works great, but yafu segfaults when attempting to use the lasieve binaries.

Note:The binaries work fine with factmsieve.py

[Dubslow]

Whenever I run something like "yafu -v factor(asdfasdfasljkf)", I get an error like "no switch detected". When I do "yafu factor(asdfasdfadsf) -v", that works just fine; the problem is I can't do the latter with aliqueit. Is it possible to get the former to work so I can put "~/yafu/yafu -v" as my yafu cmd in aliqueit.ini?

[schickel]

Originally Posted by Dubslow

Whenever I run something like "yafu -v factor(asdfasdfasljkf)", I get an error like "no switch detected". When I do "yafu factor(asdfasdfadsf) -v", that works just fine; the problem is I can't do the latter with aliqueit. Is it possible to get the former to work so I can put "~/yafu/yafu -v" as my yafu cmd in aliqueit.ini?

As a workaround on Windows, you could try making a batch file with the switches you desire and then put the batch file as the yafu command.

Example:

yafu_cmd = YAFUBAT

YAFUBAT.bat:

yafu_path\yafu -v %1

(Of course, I haven't looked at exactly how aliqueit.exe passes the arguments to the particular factoring program......but above I'm assuming that the argument is passed on the command line, therefore the batch file passes its first parameter to YAFU.)