mersenneforum.org > Data Thinking out loud about getting under 20M unfactored exponents
 Register FAQ Search Today's Posts Mark Forums Read

 2021-07-07, 07:42 #540 lycorn     "GIMFS" Sep 2002 Oeiras, Portugal 27648 Posts I´m just about to finish taking 15M range from 69 to 70 bits. The range has so far yield 189 factors, since the 11th of May. Been using a GTX1160Ti plus some help from a few Colab instances. I expect maybe half a dozen more till I call it a day. Nothing to write home about, but not too bad. So I intend to start 14.8 and 14.9 soon, if there is no impact on anybody´s work. Is anybody working or intending to start working on said ranges? If not, I´ll start in one or two days (TFing 69-> 70). Last fiddled with by lycorn on 2021-07-07 at 07:43
 2021-07-07, 14:44 #541 ZacHFX     Mar 2017 Halifax, NS 618 Posts I found 45 factors taking 12.1 up to 71 bits. Took about as long as expected. Interestingly, one of the very first, 12100063, yielded a factor in both ranges, so 44 total exponents have been factored. Unfortunately, that is the extent of my contribution until the temperature starts to drop in the fall.
2021-07-07, 16:42   #542
lycorn

"GIMFS"
Sep 2002
Oeiras, Portugal

22·3·127 Posts

Quote:
 Originally Posted by ZacHFX Unfortunately, that is the extent of my contribution until the temperature starts to drop in the fall.

Have you also been affected by the heat wave(s) that hit Canada lately? I believe it was cooler at your place.

Gosh, British Columbia hit temperature values that were never seen even in Southern Portugal. The highest there was, to the best of my knowledge, 47.6 C some twenty years ago or so, and it was something astonishing. But it was in a place where temps around 42/43 C are expected every summer. Now 49.6 C in Canada??!!...

 2021-07-07, 18:50 #543 ZacHFX     Mar 2017 Halifax, NS 72 Posts No, if anything it was slightly below seasonal temperatures here. Our temperatures are heavily moderated by ocean temperatures, so above 30C or below -15C happens only a handful of times each year. I am trying to be more energy conscious, so if the heat isn't warming my home, I try not to produce it at all. Last fiddled with by ZacHFX on 2021-07-07 at 18:51
 2021-07-10, 07:52 #544 LaurV Romulan Interpreter     "name field" Jun 2011 Thailand 986910 Posts We have seen 11M didn't move in the last 30 days. Therefore, we are doing some P-1 there. Mind your toes... Edit: exactly 1001 factors needed. Last fiddled with by LaurV on 2021-07-10 at 08:14
2021-07-10, 08:55   #545
axn

Jun 2003

149816 Posts

Quote:
 Originally Posted by LaurV We have seen 11M didn't move in the last 30 days. Therefore, we are doing some P-1 there. Mind your toes... Edit: exactly 1001 factors needed.
Whole of that range ?!

What bounds are you planning to use for each 0.1m subrange?

 2021-07-10, 10:07 #546 LaurV Romulan Interpreter     "name field" Jun 2011 Thailand 268D16 Posts Not sure yet, still testing, I am thinking about 5-6 minutes max, per candidate per GPU. Which means somewhere at B1=3M, B2=100M. See the end of the range, results are coming in. The round numbers, I mean, like 107127521 This is again borderline, and I don't really beleive I will get lucky again and again forever, like in the past (but I hope ), so for sure there will be at least 2 bits more TF needed. We don't need to finish everything, but have to start from somewhere. @Chris: Is it possible that you pull the 11.xM (from the end towards beginning, x$$\ne$$5) to 71 and pass them to me, as we did with 35.1? (some are already to 71++, but most are 69 or 70). If so, then pretty-pretty-pleaaaseee.. (strike that out, see my next post - maybe we will return to this later, if deeper P-1 fails) Last fiddled with by LaurV on 2021-07-10 at 14:10
 2021-07-10, 11:47 #547 lycorn     "GIMFS" Sep 2002 Oeiras, Portugal 22·3·127 Posts Speaking of toes, I just finished 15M to 70 bits. 196 factors in exactly 2 months. Whoever feels like it, please help him(her)self in that range. As posted above, I am now crunching 14.8 and 14.9 M ranges. I plan to move downwards to 14.1 M, but will keep you posted so as no "accident" happens.
2021-07-10, 11:48   #548
axn

Jun 2003

149816 Posts

Quote:
 Originally Posted by LaurV Not sure yet, still testing, I am thinking about 5-6 minutes max, per candidate per GPU. Which means somewhere at B1=3M, B2=100M. See the end of the range, results are coming in. The round numbers, I mean, like 107127521
Sounds good. But I suspect 11.0 & 11.7 would need much higher bounds seeing as they need many more factors than the rest.

 2021-07-10, 14:04 #549 LaurV Romulan Interpreter     "name field" Jun 2011 Thailand 986910 Posts Ok, for 11M (and probably the neighbors too), P-1 is definitively the way to go. One TF from 69 to 70 takes close to 3 minutes in a GPU that spits ~5100GHzDays/Day, and going to 71 takes close to 6 minutes, and both of them have a chance of finding a factor in anything between 70 and 100 trials, considering the amount of P-1 already done in the range, while about 6 minutes of P-1 in a good GPU (like V100 or RVii) has a chance of 1 in 12 (ignoring the already-done P-1) or 1 in 25-30 (considering the P-1 already done, and all candidates factored to 72, which they are not - therefore a guess is the chances are about one in 23, because we like the number 23 ). So, yeah, P-1 till the hammer breaks... Edit: I added snip and edited the post above to cancel the request for Chris, we stay at P-1 for now... Last fiddled with by LaurV on 2021-07-10 at 14:23
 2021-07-10, 15:57 #550 petrw1 1976 Toyota Corona years forever!     "Wayne" Nov 2006 Saskatchewan, Canada 3·1,663 Posts I agree As we get to these lower ranges TF becomes more expensive and P1 cheaper. With the end goal of completely clearing a range my strategy is to find the best balance. 1. TF as deep as is reasonable/affordable to get enough factors 2. Pick P1 bounds that should find the rest of the factors. The order of the above doesn't matter and can be adjusted by how much of each help I'm getting. Last fiddled with by petrw1 on 2021-07-10 at 16:03

 Similar Threads Thread Thread Starter Forum Replies Last Post jschwar313 GPU to 72 3 2016-01-31 00:50 Batalov Factoring 6 2011-12-27 22:40 jasong jasong 1 2008-11-11 09:43 devarajkandadai Math 4 2007-07-25 03:01 WraithX GMP-ECM 1 2006-03-19 22:16

All times are UTC. The time now is 20:07.

Sat Jan 22 20:07:01 UTC 2022 up 183 days, 14:36, 0 users, load averages: 1.66, 1.92, 1.84