[QUOTE=Batalov;286301]Don't know if this was calculated before.
[/QUOTE]
Wow! That is nice! I like [URL="http://factorization.ath.cx/aliquot.php?type=1&aq=453798"]that graphic[/URL]!

[QUOTE=Batalov;286301]Don't know if this was calculated before.
[/QUOTE]I had not seen that anywhere. Very interesting stats.....

Wow, unconnected is having a tremendous run with 11040. The downdriver was captured at i6488 at a size of 152 digits. Right now it's down to 115 digits and is at i6662, so it's been 175 lines and 37 digits so far!

Here's hoping it goes a lot lower.....

[QUOTE=schickel;286922]Wow, unconnected is having a tremendous run with 11040. The downdriver was captured at i6488 at a size of 152 digits. Right now it's down to 115 digits and is at i6662, so it's been 175 lines and 37 digits so far!

Here's hoping it goes a lot lower.....[/QUOTE]

just looked it up it's now down to C106 with a C90 cofactor. at i6704

I saw the beauty went as low as 20 digits and it almost reached ten thousands terms!! Nice-nice. Is that a record of highest digitlength when catching a downdriver?

I hope I didn't jinx it by talking about it....[QUOTE=LaurV;287007]I saw the beauty went as low as 20 digits and it almost reached ten thousands terms!! Nice-nice. Is that a record of highest digit length when catching a downdriver?[/QUOTE]Actually, it's in fourth place. If you read starting from [URL="http://www.mersenneforum.org/showthread.php?p=256630#post256630"]here[/URL] and go through #1147, you'll see that we've set the bar pretty high. The top spots are occupied by 4788 at 175-digits, 1134 at 164-digits, 3906 at 158-digits, and 11040 at 152-digits.

And on another note....here's a blast from the past from May '09:[QUOTE=mklasson;172844]Ooh, this morning greeted me with a new downdriver record. [url=http://factorization.ath.cx/search.php?aq=139314&action=last&fr=&to=]139314[/url]'s downdriver run started with a c131 at index 1724 and ended with a c23 at index 2288 for a total length of 565 lines and a reduction of 108 digits! :toot: Scary [url=http://factorization.ath.cx/aliquot.php?aq=139314]fall[/url], isn't it?[/QUOTE]That was indeed a record, but guess what. Records are made to be broken, and we have a new #1!

The downdriver for 11040 ran from i6488, 152-digits to i7087, 25-digits, or [B][COLOR="Red"]600 lines[/COLOR][/B] and [B][COLOR="red"]127-digits[/COLOR][/B]! In fact, this puts the record for 1578 listed on Wolfgang's page to shame.

Great work, unconnected!

[QUOTE=schickel;287015] The top spots are occupied by 4788 at 175-digits, 1134 at 164-digits, 3906 at 158-digits, and 11040 at 152-digits.
[/QUOTE]
Is there any place where all these records are noted? (beside of schickel's copybook :D). If not, we should make a thread with "Aliquot records". That would be pretty interesting for guys like me, and quite motivating for [strike]newbies[/strike] everyone else (at least, is more fun to try breaking aliquot records than splitting aliquot terms, isn't it? :P)

[QUOTE=LaurV;287021]Is there any place where all these records are noted? (beside of schickel's copybook :D). If not, we should make a thread with "Aliquot records". That would be pretty interesting for guys like me, and quite motivating for [strike]newbies[/strike] everyone else (at least, is more fun to try breaking aliquot records than splitting aliquot terms, isn't it? :P)[/QUOTE]The premiere place (at least according to Wikipedia!) is Wolfgang's [URL="http://www.aliquot.de/aliquote.htm"]page[/URL]; another place is Clifford's [URL="http://www.lafn.org/~ax810/aliquot.htm"]site[/URL] (which has fallen out of sync by >1 year, unfortunately).

Other than that, what gets posted here is what we notice, either as we're working it, or afterward upon review.... Unfortunately, as far as downdriver captures go, as I said, we've set the bar pretty high. Most everyone will not be able to go up to 175 digits.....

So, in that spirit, here are some current records:

Longest sequences (>8000):[code]842592 8003. sz 172 2^3 * 3 * 5 * 13 * 587 * 823 * 1627
195528 8017. sz 144 2 * 3^4 * 24110979082363
552150 8197. sz 147 2^2 * 3 * 7^2 * 17 * 953 * 16493
453798 8565. sz 148 2^2 * 7
11040 9020. sz 109 2^4 * 431 * 2527823
933436 12392. sz 161 2^2 * 7 * 13[/code]The shortest ones are at 400 lines...)

Largest sequences (>= 170 digits):[code] 276 1720. sz 170 2 * 3^3 * 5^2 * 17 * 17863
19560 486. sz 170 2^3 * 3 * 5 * 7 * 59 * 4520183
5748 1470. sz 172 2^2 * 7 * 13
842592 8003. sz 172 2^3 * 3 * 5 * 13 * 587 * 823 * 1627
564 3373. sz 175 2^2 * 3^2 * 7 * 13 * 71 * 4292236942619
2340 696. sz 177 2^3 * 3^2 * 5 * 13 * 107 * 142501 * 31734053699405310179
3270 647. sz 177 2^5 * 3 * 7 * 31 * 3739 * 497867 * 1045129066769
966 893. sz 178 2^2 * 3^2 * 5 * 83 * 2099
162126 4283. sz 178 2^3 * 3 * 5 * 461
552 1057. sz 179 2^2 * 3 * 71 * 145633
8352 1739. sz 181 2^2 * 3 * 7 * 53 * 2101111 * 38938746337737993077599
660 895. sz 183 2^3 * 3^2 * 5 * 3163 * 14159 * 32070039222359 * 81741146396847353333[/code]The smallest are 110 digits, except during downdriver runs.

The highest powers of 2 (>=10):[code]947208 1648. sz 120 2^10 * 3 * 11 * 23 * 433
589212 1036. sz 118 2^10 * 3 * 23 * 337
536904 1775. sz 124 2^10 * 3 * 5^2 * 389 * 28541
552876 2040. sz 123 2^10 * 3^2 * 5^2 * 13 * 2377097
622830 871. sz 116 2^10 * 3^3 * 7
715620 773. sz 120 2^12 * 3 * 13 * 769
583800 449. sz 119 2^12 * 3 * 7 * 673
116712 1585. sz 117 2^12 * 3^2[/code]

There is only one merge (so far) that can't be found if you check using a UInt (up to 2^32). 237552 merging with 70740 has a value of about 4.7 * 10^9 which is a bit over.

I tried to save space by using a uint32 for my hash table rather than a ULongLong and found I missed this one merge when I did so.

[QUOTE=Greebley;293652]There is only one merge (so far) that can't be found if you check using a UInt (up to 2^32). 237552 merging with 70740 has a value of about 4.7 * 10^9 which is a bit over.

I tried to save space by using a uint32 for my hash table rather than a ULongLong and found I missed this one merge when I did so.[/QUOTE]IIRC, that should be the biggest size you have to worry about. There was one merge last year that required updating the UB merge program from Clifford's site from 8 or 9 digits to 10 digits to properly identify....I'll have to dig back through my email archives to recall which one.

Oooh, so close and yet so far!
 

Check out this factorization from 660:[code] 894 . 65051760592868542076211649810607577485597730105437293588816311361066076233410142765489920117370807000836308186259069153189865018491113647173837923678665633360181424192261587457433560 = 2^3 * 3^2 * 5 * 180699334980190394656143471696132159682215916959548037746711976002961322870583729904138666992696686113434189406275192092194069495808649019927327565774071203778281733867393298492871[/code]Missed escaping the driver by [i]that[/i] much.....