fond of a factor? Burn yourself to become remains

[Happy5214]

Originally Posted by Viliam Furik

It depends on the allocated RAM, and other things, so allowing ~2.5 GB of RAM per worker probably makes the difference.

Meanwhile...
Oh, you generated your B2s yourself. I let Prime95 generate them. So that's probably the main difference.

You can simply use this tool on mersenne.ca here. You can set your exponent range and B1, and it will generate worktodo lines for you. You only need to delete those which already had a bigger B1 done. Or use a script for them...

No, I used https://github.com/keisentraut/prime95-optimal-worktodo, which leaves B2 blank (letting mprime determine it). However, it also adjusts the TF field in the worktodo based on the ECM level, so I think I was getting lower B2's since there was more ECM work done (and thus a higher implicit TF level). My max RAM allocation for that instance is 32 GB, and it never got close to that while I was doing that range.

[Happy5214]

M5210099 has a factor: 2944892728119316210127 (P-1, B1=2500000, B2=207500000) (71.319 bits)

It's puny, I know, but I'm fond of it because it's my first non-ECM stage 2 Mersenne factor. My previous P-1 and P+1 factors were both found in stage 1.

[Miszka]

M23243107 has three factors in the TF71 range. It's kind of a curiosity.
I happened to find 2 factors in one TF range several times, but 3 for the first time!

[Viliam Furik]

Originally Posted by Miszka

M23243107 has three factors in the TF71 range. It's kind of a curiosity.
I happened to find 2 factors in one TF range several times, but 3 for the first time!

Not so long ago I had a brief discussion with user tuckerkao about the possibility of this happening... Nice luck!

[James Heinrich]

Originally Posted by Miszka

M23243107 has three factors in the TF71 range. It's kind of a curiosity.

But not as rare as you might think. There are >2500 such examples in GIMPS range (but only 6 new examples have been found so far in 2021).
There are (so far) only 7 examples where there are 4 known factors in the same bitlevel:
33218089, 128757283, 143570279, 234616103, 563950589, 578590853, 946287211

[Miszka]

Originally Posted by James Heinrich

But not as rare as you might think. There are >2500 such examples in GIMPS range (but only 6 new examples have been found so far in 2021).
There are (so far) only 7 examples where there are 4 known factors in the same bitlevel:
33218089, 128757283, 143570279, 234616103, 563950589, 578590853, 946287211

These 7 examples have a fairly low bitlevel and it will be rather very difficult to find similar ones with a bitlevel above 70-some

[Jwb52z]

P-1 found a factor in stage #2, B1=760000, B2=20835000.
UID: Jwb52z/Clay, M105530011 has a factor: 700781945616999543935075224151 (P-1, B1=760000, B2=20835000)

99.145 bits.

[James Heinrich]

Not mine, but a pretty triple factor:

M105587893 has a 236.753-bit (72-digit) composite (P23+P24+P26) factor: 186125867355522179214500290117704070927566484401081129960541990698683479 (P-1,B1=823000)

[masser]

Originally Posted by James Heinrich

Not mine, but a pretty triple factor:

M105587893 has a 236.753-bit (72-digit) composite (P23+P24+P26) factor: 186125867355522179214500290117704070927566484401081129960541990698683479 (P-1,B1=823000)

That's pretty sublime.

[LaurV]

Nice! That is a beauty! I didn't found a triple yet.
Congrats to the founder! (Kriesel)

[ET_]

The smaller with 76.191 bits is close to the 76 bit range for such exponents