Some Somewhat Easier n^i Sequences Available for Termination

[gd_barnes]

I'm done with 14316^38.

I'm now working on 42^100 and 113^89 until termination.

Ed,

I ran my own counts on sequences remaining by base in my spreadsheet to make sure I was up to date. I found some problems with your 1 and 2 remaining lists:

Should be on 1 remaining list:
33550336^24 only remains for base 33550336.

Should be on 2 remaining list:
564^58 and 60 only remain for base 564.
14288^38 and 40 only remain for base 14288.

I took into account all recent work. We have not worked on any of these bases recently.

[EdH]

Originally Posted by VBCurtis

I've had polyselect hang on occasional small ad-values, which is what caused me to add "admin = " lines to my polyselect params. It seems that when I skip 0-5000 or 0-10k, these hangs don't happen. The evidence for me is one range that isn't returned, status is stuck at 99%, and "top" shows one thread still going on polyselect (or maybe timed out, resubmitted, timed out again, and stuck).
CADO isn't smart enough to send another WU out for a higher range to reach 100%, sadly.

In other news, I came home to a 103F (~40C) day, and my house took 6+ hours for AC to cool to some reasonably tolerable temp. I haven't fired up my desktop yet, so C14x jobs will wait a day or so for this heat to ebb a little.

Bummer on the hot house. I have a hot room that isn't insulated well enough from the nearby ones. No A/C, just a fan for the rest of the upstairs, but our climate isn't quite so warm. My shed is where the bulk of the bigger machines sit. Unfortunately, the warmth doesn't translate to comfort in the winter.

I've experienced the instances you described, sitting at 98/99% until timeout/resubmit, but oddly, this was sitting at 16%, had not timed out and all the clients were disengaged and waiting for new work. I do have them using --single to process WUs, so I can control them all via scripts, but none showed that type of closure. The params should have been the latest ones from the harvest thread, which I intended to add to with the c165 results.

[EdH]

I must have missed those in my visual search - one of the reasons I try to use so many scripts when I can.

I've queued up 75^87, 1210^50, 47^95, 510510^28 and 75^87. Let's see how far I get. . .

[gd_barnes]

42^100 and 113^89 terminate

Base 42 is complete.

75^87 and 1210^50 terminate

Base 1210 now has 2 remaining.

Thank you to Ed for the last two. :-)

[gd_barnes]

47^95 and 89^93 terminate

I'm working on 510510^28.

Base 47 now has 2 sequences remaining. But this is a strange one. One of the remaining is 47^97. It merges with 1134 of the main project. It's the only same-parity merge that I'm aware of on this project. In the main project, if a sequence merges with another, it is no longer listed anywhere...like it's no longer remaining. So using that logic, 47^97 no longer remains and hence base 47 only has 1 sequence remaining: 47^99.

[EdH]

Anything that merges with the main Aliqueit project is still tracked by my scripts, but is out of our reach due to size. It's reservation is then tracked by the main Aliqueit reservation system so it is left alone by the n^i project.

All of my last run has finished. I need to review where that leaves us and see where I'll work next. 47^99 will probably be left for the moment. It looks like it's going to fight all the way.

[gd_barnes]

Later today I'll post some updated lists.

If you don't see anything reasonable on your updated 1/2 remaining lists, you could refer back to the last current size < 150 list or the largest small factor list.

[EdH]

Originally Posted by gd_barnes

Later today I'll post some updated lists.

If you don't see anything reasonable on your updated 1/2 remaining lists, you could refer back to the last current size < 150 list or the largest small factor list.

Thanks. I'm looking into params for possibly running the c173 for 92^90.* I'm interested in comparing GPU LA to 40-thread CPU LA for that size, anyway. My best time for a c160 was just shy of 25 hours total. That means the c173 would probably be looking at one week+, at best. But, that may very well take 92^90 into easy termination. ATM, I'm playing elsewhere, though.

* I'm running a little more ECM first, but I'd like to work on params in case I need them.

[gd_barnes]

510510^28 terminates

Base 510510 has 1 sequence remaining.

Originally Posted by EdH

Thanks. I'm looking into params for possibly running the c173 for 92^90.* I'm interested in comparing GPU LA to 40-thread CPU LA for that size, anyway. My best time for a c160 was just shy of 25 hours total. That means the c173 would probably be looking at one week+, at best. But, that may very well take 92^90 into easy termination. ATM, I'm playing elsewhere, though.

* I'm running a little more ECM first, but I'd like to work on params in case I need them.

That's a lot of patience. Good luck!

[gd_barnes]

137^85 and 179^81 terminate

I have expanded my search area up to a starting size (index=0) of 185 digits from 180 digits. This means that all same-parity sequences starting at <= 185 digits have all been ECM'd to at least t35. There were originally ~80 sequences with a starting size range of 181 to 185 digits including ones for recently initialized bases.

This has resulted in 4 terminations recently: the 2 above plus 89^93 and 113^89 reported previously. There may be 1-2 more in the next 1-2 days. It has also resulted in some additional sequences being added to the updated lists that I will post next, some of which are a bit interesting.

Same-parity starting sizes > 180 digits are only present in bases 66 thru 239, 31704, 33550336, 6469693230, 8589869056, and 7420738134810.

[gd_barnes]

Here are updated lists of bases with 1-2 same-parity sequences remaining as of this post. Most are ECM'd to t35. Some may be ECM'd to t40.

Bases with 1 remaining:

47^99: 161/126/3
277^69: 165/165/3
510510^30: 172/151/31
9699690^24: 168/146/7
33550336^24: 181/135/3
200560490130^14: 160/131/3
7420738134810^14: 182/161/3

Bases 47, 9699690, and 200560490130 are likely the easiest. Base 9699690 higher size being offset by not having a factor 3.

For base 47, 47^97 is no longer considered remaining since it merged with a smaller sequence in the main project.

Bases with 2 remaining:

Base exponent: size/cofac/small factor
37    105: 163/138/3, 107: 155/137/7
55    93: 157/152/5, 99: 172/126/5
241    67: 154/138/5, 69: 160/119/3
257    65: 150/139/7, 69: 161/128/3
288    64: 158/132/17, 65: 158/137/3
338    64: 151/132/3, 65: 165/148/8317
385    61: 158/140/37, 65: 161/155/43
552    58: 154/139/7, 60: 165/157/5
564    58: 160/155/5, 60: 166/145/5
660    56: 152/150/79, 60: 170/140/7
1184    52: 160/152/5, 54: 166/154/3
1210    52: 161/118/3, 54: 167/135/3
2310    48: 161/140/5, 50: 168/140/3
12496    38: 156/145/3, 40: 164/158/3
14264    38: 157/145/53, 40: 167/159/5
14288    38: 152/146/570919, 40: 167/159/5
14316    38: 157/136/229, 40: 167/152/5
131071    33: 161/138/5, 35: 172/171/5

Bases 241, 257, and 288 look the easiest here and appear to be about equal difficulty. Bases 338 and 385 could be considered too. The large smallest factor for 338^65 will drop it quickly. Base 385 smallest factors are both > 35.