[QUOTE=KingKurly;258727]I'll be finishing up a P-1 on M332205149 a little later this evening,[/QUOTE]
Cool! We will have 97 that have been P-1'ed.

I have not included P-1 into the factoring effort number. If you can report on the GHz-days that you get, I will try to calibrate it.

Well, it looks like it [url=http://v5www.mersenne.org/report_exponent/?exp_lo=332214409]worked[/url] :)
P-1 with the specified bounds should give [url=http://mersenne-aries.sili.net/prob.php?exponent=332214409&b1=3365000&b2=95902500]210.272541[/url] GHz-days credit. (P-1 credit doesn't vary depending on the TF done beforehand, but the factor probability does). This exponent is a very good example of why a little more GPU-TF is an excellent idea before running large P-1s -- the factor is 77.7 bits; another 138GHz-days of work would have taken it from 2^76 to 2^78 and found the factor. With a good GPU that's only around a 1-day assignment.

[QUOTE=James Heinrich;258730]Well, it looks like it [URL="http://v5www.mersenne.org/report_exponent/?exp_lo=332214409"]worked[/URL] :)
P-1 with the specified bounds should give [URL="http://mersenne-aries.sili.net/prob.php?exponent=332214409&b1=3365000&b2=95902500"]210.272541[/URL] GHz-days credit. (P-1 credit doesn't vary depending on the TF done beforehand, but the factor probability does). This exponent is a very good example of why a little more GPU-TF is an excellent idea before running large P-1s -- the factor is 77.7 bits; another 138GHz-days of work would have taken it from 2^76 to 2^78 and found the factor. With a good GPU that's only around a 1-day assignment.[/QUOTE]
Someone else did a P-1 that finished the same day as mine? That's quite the coincidence! Mine is at 99.17% of stage 2 and is projected to finish in a little over four hours from now.

Much more likely someone reported a TF result. Few of us are crazy enough to do such large P-1s :smile:
But don't be surprised if PrimeNet tells you "result not needed". Although (I think?) you should still get the ~210GHz-days credit.

[QUOTE=James Heinrich;258733]Much more likely someone reported a TF result. Few of us are crazy enough to do such large P-1s :smile:
But don't be surprised if PrimeNet tells you "result not needed". Although (I think?) you should still get the ~210GHz-days credit.[/QUOTE]Your link referenced M332214409 but my exponent is M332205149. I think I'll be okay. And yes I am crazy. :loco:

Back when I started the assignment, I had 8GB of RAM. Now I have 16GB. I don't know if I'll do another 100M digit P-1, but I will certainly continue doing P-1! When the result for M332205149 posts, I should jump from #37 to #30 in the last-365-day rankings.

[QUOTE=KingKurly;258734]Your link referenced M332214409 but my exponent is M332205149[/QUOTE]Oh. Well then. That would be my fault. Sorry! :blush:
You're right, [b]your[/b] exponent is [url=http://v5www.mersenne.org/report_exponent/?exp_lo=332205149]currently[/url] TF'd to 2^76, and assigned to you for P-1. May you find a wonderful factor in 3 hours or so. :smile:

I've done 3 such P-1s (332203901, 332217439, 333000091) over the last few years. But I agree it's a long-term commitment of a large chunk of resources, and I'm not sure that I'll do any more in that range, in the near term at least. I'm having a lot more fun (and success!) working in smaller ranges; in the [url=http://v5www.mersenne.org/report_top_500_P-1/]past year[/url] I've just about caught up to ANONYMOUS for P-1 factors found (should pass within a week or so), and I've found more than anyone else except [i]markr[/i].

I ran the factor for the exponent M332214409 (that you linked, not the one I'm working on) through my little Ruby script that analyzes factors -- it's about as smooth as sandpaper. I think we definitely would [I][B]not [/B][/I]have found this one in P-1! :smile:

M332214409 has factor 244701274525241425768649 (77.69 bits)
k = 2^2 * 92072042894609

[QUOTE=KingKurly;258738]I think we definitely would [I][B]not [/B][/I]have found this one in P-1! :smile:[/QUOTE]That's pretty impressive... Finding that with P-1 would take about [url=http://mersenne-aries.sili.net/prob.php?exponent=332214409&b1=1000&b2=92072042894609]335[b][i]T[/i][/b]Hz-[i]years[/i][/url] :surrender

[QUOTE=KingKurly;258727]I'll be finishing up a P-1 on M332205149 a little later this evening, assuming my power doesn't go out with these crazy storms. Current time is 19:05 and current estimate is 00:20. Using B1=3365000 and B2=95902500. Stage 1 finished back on March 5 without finding a factor. Obviously I'll report back with my findings. It's finishing earlier than I had originally projected because I added an extra core to the assignment.

I previously reported that I'd be doing this work at [URL]http://mersenneforum.org/showthread.php?p=254426#post254426[/URL] and perhaps other places... not sure why I didn't think to post in this thread until now.[/QUOTE]
Well, somewhat unsurprisingly, P-1 was unable to find a factor for my exponent [URL="http://mersenne.org/report_exponent/?exp_lo=332205149"]M332205149[/URL]. The server assigned credit of 181.3220 GHz-days, which seems a little low according to James' estimate. Strange. Oh well, that makes 97 exponents tested, right?

Credit aside, I do not plan to do further work on this exponent: if anyone else is interested in working on it, feel free to reserve the assignment with the server and have at it! :)

Here is a status report for the range from 332192831 to 332399999:

[code]Date of data 4/20/2011
Average bit depth for first 100 expos 78.23
Average bit depth for first 1000 expos 76.20
100th active expo (no factor found) 332198357
1000th active expo (no factor found) 332245261
Unitless total effort number 131,459,072 estimated GHz/day equiv = 184,787
Number of first 100 expos to 2^76 92
Number left in range 4290

Estimated expos in range to be removed 503
(by taking all expos to 2^82)[/code]
[code]Bit # at bit level
70 1128
71 1758
72 192
73 66
74 108
75 130
76 303
77 248
78 9
79 324
80 23
81 1
P-1 98 ≈8,718,710 effort[/code]
~1,672,614 effort equivalent in factors found by TF. There are 3 factors certain to be from P-1.

The attached graph shows a second inflection point. More GPU's seem to have joined the effort.

[QUOTE=KingKurly;258746]The server assigned credit of 181.3220 GHz-days, which seems a little low according to James' estimate. Strange.[/QUOTE]Since v26.x, new FFT sizes are available, and when smaller FFT sizes can be used, they should take less time; you therefore get less credit. I finally [url=http://www.mersenneforum.org/showpost.php?p=258851&postcount=525]figured it out[/url] and [url=http://mersenne-aries.sili.net]mersenne-aries.sili.net[/url] P-1 (and L-L) credit calculations now match PrimeNet once again.