mersenneforum.org Getting others to do the work on exponents I like (was: Trial Factoring Progress)
 Register FAQ Search Today's Posts Mark Forums Read

 2021-02-16, 06:39 #1 tuckerkao   "Tucker Kao" Jan 2020 Head Base M168202123 10111010002 Posts Getting others to do the work on exponents I like (was: Trial Factoring Progress) I found a factor between 2^74 and 2^75 for M168,377,329 several minutes ago - https://www.mersenne.org/report_expo...exp_hi=&full=1 M82,589,939 has a known factor too - https://www.mersenne.org/report_expo...exp_hi=&full=1 Any similarities can be observed between these 2 exponents? Last fiddled with by tuckerkao on 2021-02-16 at 06:41
 2021-02-16, 06:51 #2 Uncwilly 6809 > 6502     """"""""""""""""""" Aug 2003 101×103 Posts 2×5,309 Posts They are both known composites. As are the bulk of all Mersenne Number.
 2021-03-03, 03:43 #3 tuckerkao   "Tucker Kao" Jan 2020 Head Base M168202123 23×3×31 Posts I believe I have figured out the assignment system by the trial factoring within Category 2 and 3. After I manually submit the results between 2^74 to 2^76, someone like curtisc will start to test the exponent I want with a faster PRP result - https://www.mersenne.org/report_expo...exp_hi=&full=1 Last fiddled with by tuckerkao on 2021-03-03 at 03:47
 2021-03-03, 13:19 #4 kriesel     "TF79LL86GIMPS96gpu17" Mar 2017 US midwest 3×133 Posts The P-1 factoring done on https://www.mersenne.ca/exponent/108377323 is quite inadequate. It looks like curtisc has systems with prime95 default memory settings preventing stage 2 P-1. Doing adequate bounds P-1 the first time is efficient; doing stage1 only first or probability-of-factor-per-cpu-hour first, with or without a followup second factoring to adequate bounds to retire the P-1 task, is not efficient. I'm running a cleanup P-1 on that exponent now, which will complete in about an hour. If you want to primality test yourself those exponents you begin with TF, immediately after reporting the TF to adequate bounds, request a manual PRP assignment for the same exponent, then adequately P-1 factor it before beginning the primality test (PRP/GEC/proof). Last fiddled with by kriesel on 2021-03-03 at 13:28
 2021-03-06, 03:10 #5 tuckerkao   "Tucker Kao" Jan 2020 Head Base M168202123 23·3·31 Posts It seemed like that the chance of finding a factor only dropped by 0.0794% for not doing the trial factoring from 2^77 to 2^78 after the adequate P-1 factoring conducted - https://www.mersenne.ca/exponent/168374303 I finished the 2^78 bit on another exponent and didn't find a factor anyway, so unless someone really finds at least a factor between 2^77 to 2^80, I probably won't do the trial factoring further, just aiming for the direct PRP test.
2021-03-06, 03:55   #6
axn

Jun 2003

150116 Posts

Quote:
 Originally Posted by tuckerkao It seemed like that the chance of finding a factor only dropped by 0.0794% for not doing the trial factoring from 2^77 to 2^78 after the adequate P-1 factoring conducted - https://www.mersenne.ca/exponent/168374303
No, it is 0.3868%, if we go by the calculation. However, I don't trust either of those numbers. I think empirically, it is more like 1%.

Quote:
 Originally Posted by tuckerkao I finished the 2^78 bit on another exponent and didn't find a factor anyway, so unless someone really finds at least a factor between 2^77 to 2^80, I probably won't do the trial factoring further, just aiming for the direct PRP test.
It seems like you're trying to rationalize not running deeper TF? I mean, it is your hardware, so do what you want, but you'll be better off in the long run by doing the recommended TF.

Last fiddled with by axn on 2021-03-06 at 03:55 Reason: quote

2021-03-06, 09:19   #7
tuckerkao

"Tucker Kao"
Jan 2020

13508 Posts

Quote:
 Originally Posted by axn No, it is 0.3868%, if we go by the calculation. However, I don't trust either of those numbers. I think empirically, it is more like 1%.
Your % is the chance that the trial factoring may find a factor between 2^77 to 2^78, but Viliam F found a factor in M107,373,143 which can also be found from the P-1 factoring -
https://www.mersenne.org/report_expo...exp_hi=&full=1

So how likely will a factor between 2^77 to 2^78 skip the P-1 factoring check?

I ran the trial factoring from 2^77 to 2^78 for another exponent and that got my video card torched hot. I'd rather Viliam F perform this action if must needed.

Last fiddled with by tuckerkao on 2021-03-06 at 09:21

2021-03-06, 10:15   #8
axn

Jun 2003

150116 Posts

Quote:
 Originally Posted by tuckerkao Your % is the chance that the trial factoring may find a factor between 2^77 to 2^78,
No, it isn't. The chance of a factor if you TF from 2^77 to 2^78 is 1/78 or 1.28%. The chance will be reduced if we're doing the TF after P-1, but it will still be closer to 1% rather than 0.07% which is just nonsense.

2021-03-06, 10:32   #9
axn

Jun 2003

537710 Posts

Quote:
 Originally Posted by tuckerkao So how likely will a factor between 2^77 to 2^78 skip the P-1 factoring check?
The P-1 probability at B1=1M, B2=40M with TF done to 2^77 is 4.3848. Same probability with TF done to 2^78 is 3.9980. Hence, the overlap of probability is 0.3868 (i.e. the probability that there is a factor that could be found by either P-1 or TF).

So probability that TF will find a factor after P-1 has completed is 1.28 (normal TF prob) - 0.3868 (overlap prob) = 0.895% (or about 70% of the normal TF prob). This is the correct probability.

 2021-03-06, 11:11 #10 tuckerkao   "Tucker Kao" Jan 2020 Head Base M168202123 23×3×31 Posts Looks like this is a personal preference ratio and may depend on each individual PC machines. It takes me 16 straight hours to run a trial factoring from 2^77 to 2^78, finish a PRP of the exponent that size will take me 28 days nonstop. I have the liquid cooling for CPU but not GPU, the GPU couldn't run at its full speed if it's overheating. I don't know about the CPU and GPU speeds of Kriesel and Viliam's computers, but the ratio doesn't appear to save me time in the long run. Last fiddled with by tuckerkao on 2021-03-06 at 11:13
2021-03-06, 12:32   #11
axn

Jun 2003

537710 Posts

Quote:
 Originally Posted by tuckerkao It takes me 16 straight hours to run a trial factoring from 2^77 to 2^78, finish a PRP of the exponent that size will take me 28 days nonstop.
On what hardware?

 Similar Threads Thread Thread Starter Forum Replies Last Post LaurV GPU to 72 81 2020-12-02 05:17 jfamestad PrimeNet 3 2016-11-06 20:32 lidocorc PrimeNet 4 2008-11-06 18:48 jocelynl 15k Search 0 2003-07-11 14:23 ThomRuley Software 5 2003-05-30 20:34

All times are UTC. The time now is 21:13.

Sun Jul 3 21:13:59 UTC 2022 up 80 days, 19:15, 0 users, load averages: 2.11, 1.40, 1.31