Pascal's OPN roadblock files

[hyramgraff]

Originally Posted by RichD

If you haven’t noticed the factored numbers are either removed from or the composites reduced in the tXXX file within 24-48 hours. This shows the work done by you and others. I would suggest you go to higher levels if you have the resources to do so.

I didn't know that the tXXX files were being updated that often. The date next to the link on Pascal's page makes it look like the mwrb file was last updated in September and the tXXX files were last updated in July.

I do plan on running ECM with larger values of B1 but I think I'll need to come up with a better system of tracking how much ECM work has been done to keep up with the updates to the t2100 file. I'll probably post whatever I come up with on GitLab.

Finally, I found one more full factorization: C785 = P27 * PRP759 http://factordb.com/index.php?id=1100000000685515782

[RichD]

Originally Posted by hyramgraff

The date next to the link on Pascal's page makes it look like the mwrb file was last updated in September and the tXXX files were last updated in July.

Those are the dates new numbers are added to the files. To be more exact, that is the date the files are recreated from a massive batch run (several weeks) Pascal performs once or twice a year. My understanding is, as road blocks are broken down (factors found) it opens up more paths/branches down the tree. Sometimes “breaking” a road block will circumvent others. Even small factors could be important since they add to the abundancy which may reduce the proof tree.

Again, thank you for your work and Happy Factoring!!!

[lavalamp]

Originally Posted by hyramgraff

Finally, I found one more full factorization: C785 = P27 * PRP759 http://factordb.com/index.php?id=1100000000685515782

You seem to be having excellent luck with these!

[RichD]

Originally Posted by RichD

Taking:
41341^43-1
41953^43-1
33403^47-1

Done.
From the MWRB file - taking:
45953^41-1
46183^41-1

[hyramgraff]

I got my first full factorization at B1=50e3: C481 = P32 * PRP450 http://factordb.com/index.php?id=1100000000438620915

[RichD]

Originally Posted by RichD

Done.
From the MWRB file - taking:
45953^41-1
46183^41-1

Done - taking:
45289^41-1
33589^47-1

[hyramgraff]

I just reported another batch of factors. Here are a couple of the most interesting finds:

C301 = P37 * C265 http://factordb.com/index.php?id=1100000000689613648

C823 = P27 * PRP796 http://factordb.com/index.php?id=1100000000636616901

[lavalamp]

Originally Posted by hyramgraff

C301 = P37 * C265 http://factordb.com/index.php?id=1100000000689613648

This one is curious, as a 37 digit factor was already known, since 2014.

It seems some of these numbers have already been ECM'd to 35 or more digits. Perhaps it would be worth checking already known factors before attempting ECM on numbers which may already have had such work performed (as evidenced by large non-trivial factors).

[henryzz]

I am currently doing one curve at 1e6 on all composites in t2100

103779088383719485547086941953^17-1 has the factor 87361225775398245856726392536519 32 digits
135105887072055320983346846089448958367^13-1 has the factor 6566270747483867453428245964829 31 digits
3233171590041989829482313542253604454158216062752807487501139486248304967498923^7-1 has the factor 739060607606588874242540105797 30 digits
11640556502202899100578852298084809614840987775009786321821295976353896391558464273076472556919001080302691697720747993525529^5-1 has the factor 1246642145079216378883035471989951 34 digits
230329301^67-1 has the factor 154591295093408859570464900949557 33 digits
25154726394283160032573146394377212297310413904839520861470441148180550393725989891870564567817789316845728547629675431315820226067658875516998684045614488702461600154326807178415424024718969904341409829134186549718640099900121572317471812267468632120902724767403^3-1 has the composite factor 10623306852997607030642207221 29 digits

Done upto 555 digits

[hyramgraff]

Originally Posted by lavalamp

This one is curious, as a 37 digit factor was already known, since 2014.

It seems some of these numbers have already been ECM'd to 35 or more digits. Perhaps it would be worth checking already known factors before attempting ECM on numbers which may already have had such work performed (as evidenced by large non-trivial factors).

Some of these numbers have certainly been ECM'd very thoroughly but since I don't have a reliable way of determining which numbers have been tested, I'm testing them all myself. My current mean time between factors is somewhere around 3-5 hours so I think testing all 65k composites in the t2100 file has been worthwhile.

Also, I have two new full factorization:

C792 = P31 * PRP761 http://www.factordb.com/index.php?id...00000685518656

C666 = P26 * PRP640 http://www.factordb.com/index.php?id...00000685524925

[lavalamp]

Originally Posted by hyramgraff

I don't have a reliable way of determining which numbers have been tested

But that's just it, you do.

If a number has no known factors (or only small or algebraic factors) then there's nothing you can do.

But if a number has a 35 digit non-trivial factor, then you can relatively safely assume ECM to 35 digits has been run, and skip to 40 digits, B1=3M.