mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Software (https://www.mersenneforum.org/forumdisplay.php?f=10)
-   -   Prime95 v30.4/30.5/30.6 (https://www.mersenneforum.org/showthread.php?t=26376)

kriesel 2021-04-23 17:34

That showed 6 factors found, in 1104 attempts. (1 per 184 attempts) All factors found were from start 2/7.
Start etc. breakdown is:
[CODE]nth_run start-value count percentage
1 2/7 1068 96.74%
2 6/5 22 1.99%
3 other ("random") 14 1.27%[/CODE] I think a more balanced / optimal mix would have counts for nth_run=1 and nth_run=2 about equal, and count for nth_run=3 somewhat less than nth_run=1 count because of random start values' lower probability of finding factors than the special 2/7 or 6/5 start values. ([URL]https://mersenneforum.org/showpost.php?p=575865&postcount=7[/URL] [URL]https://mersenneforum.org/showpost.php?p=575892&postcount=15[/URL])

chalsall 2021-04-23 17:46

[QUOTE=kriesel;576642]That showed 6 factors found, in 1104 attempts. (1 per 184 attempts) All factors found were from start 2/7. Start breakdown is ...[/QUOTE]

And my understanding that there are only two "slices" through this P+1 space, so "other" really means either 2/7 or 6/5. Do I understand that correctly?

A question just hit me... Is the space explored with P+1 different as a function of the B1 and B2 values?

James Heinrich 2021-04-23 17:50

[QUOTE=chalsall;576646]And my understanding that there are only two "slices" through this P+1 space, so "other" really means either 2/7 or 6/5. Do I understand that correctly?[/QUOTE]My understanding (blind leading the blind here) was that the start values can be pretty much anything, but those two particular combinations have a [i]slightly[/i] higher chance of finding a factor than any other two start values.
The [c]nth_run[/c] parameter sets the the start values to 1=(2/7), 2=(6/5), 3=(random/random)

masser 2021-04-23 18:30

[QUOTE=James Heinrich;576647]My understanding (blind leading the blind here) was that the start values can be pretty much anything, but those two particular combinations have a [i]slightly[/i] higher chance of finding a factor than any other two start values.
The [c]nth_run[/c] parameter sets the the start values to 1=(2/7), 2=(6/5), 3=(random/random)[/QUOTE]

That is my understanding of P+1 too.

ATH 2021-04-23 18:31

2021-06-15 14:00 UTC

36 factors out of 10,950 curves: 19 P+1 smooth, 17 P-1 smooth

[CODE]
2021-04-22 06:48:35.610 [M]287873[/M] B1=10000000 B2=1000000000 start=2/7 167460871862758047584571103871 98.080 bits P+1 smooth P+1=2^7*3*7*149*31153*137737*1326817*73440613 P-1=2*5*287873*7162339*55539811*146235611
2021-04-23 03:15:32.123 [M]1600967[/M] B1=2000000 B2=146000000 start=2/7 20960634373165261374400215271 94.082 bits P-1 smooth P+1=2^3*11*13*16047334103*1141761785856421 P-1=2*3^4*5*219889*864733*1600967*85006373
2021-04-23 10:25:31.597 [M]1601783[/M] B1=2000000 B2=146000000 start=2/7 5920048152491971978196689064497 102.223 bits P-1 smooth P+1=2*7*1034951899*408579937608148822093 P-1=2^4*3*191*6637*10103*53017*1601783*113399107
2021-04-23 11:04:21.267 [M]296663[/M] B1=10000000 B2=500000000 start=2/7 33472736879943425145991507583 94.757 bits P+1 smooth P+1=2^7*3^2*7*79*3361*41897*9979727*37388971 P-1=2*23*149*34963*296663*470841841030657
2021-04-23 11:51:30.253 [M]1602143[/M] B1=2000000 B2=146000000 start=2/7 43496906104874570142712609 85.169 bits P-1 smooth P+1=2*5*7*11*56489488447889052133393 P-1=2^5*3^3*15241*17977*1602143*114686597
2021-04-23 16:33:29.030 [M]298307[/M] B1=10000000 B2=500000000 start=2/7 14600351802304915562080191901748143 113.492 bits P-1 smooth P+1=2^4*17*53677763979062189566471293756427 P-1=2*3*59*119591*298307*1512431*2126617*359445677
2021-04-24 10:32:21.530 [M]92821[/M] B1=40000000 B2=5240000000 start=2/7 6042060144547476324897523777583 102.253 bits P+1 smooth P+1=2^4*3*23*131*331*1039*2791*9151*19069*249427781 P-1=2*7*92821*3105342439*1497273943641227
2021-04-24 19:44:44.873 [M]14033251[/M] B1=500000 B2=10000000 start=6/5 105040117068459648578879 76.475 bits P+1 smooth P+1=2^6*3*5*13*47*2377*5743*5953*2203631 P-1=2*19^2*14033251*10367157737549
2021-04-27 01:25:29.000 [M]2300563[/M] B1=2000000 B2=146000000 start=2/7 5089496112298754218607 72.108 bits P+1 smooth P+1=2^4*3*7*13*23*67*1399*17077*31649017 P-1=2*31*2300563*35681981746451
2021-04-28 21:04:47.190 [M]19000027[/M] B1=1000000 B2=57000000 start=6/5 67480853305125346435319 75.837 bits P+1 smooth P+1=2^3*3*5*61*389*19553*152027*7972339 P-1=2*4021*19000027*441633774077
2021-04-29 19:08:41.640 [M]97613[/M] B1=40000000 B2=2000000000 start=6/5 64164573431546940521972630767 95.696 bits P+1 smooth P+1=2^4*19*3761*6569*1183279*128173*56329439 P-1=2*3^2*1877*97613*19455879744956054287
2021-04-29 20:38:32.100 [M]92419[/M] B1=40000000 B2=5240000000 start=6/5 26566754812550500792150275766183 104.389 bits P+1 smooth P+1=2^3*41*1013*26309*26267*17137*157061*42987011 P-1=2*3^3*92419*332567*1031519539*15517682839
2021-04-29 22:04:03.763 [M]92459[/M] B1=40000000 B2=40000000 start=6/5 543259182333011804411673092327891767 118.709 bits P+1 smooth P+1=2^3*11*6133*78173*411667*758189*1997101*20657183 P-1=2*3^3*311*92459*7965173*43924671467181023777
2021-04-30 04:37:22.343 [M]19000103[/M] B1=1000000 B2=54000000 start=6/5 38637801535161640335089 75.032 bits P-1 smooth P+1=2*3*5*7*1097*167720630008949257 P-1=2^4*53*163151*14698427*19000103
2021-05-03 [M]8413609[/M] B1=600000 B2=18000000 start=80/13 1707392059494028275527 70.532 bits P+1 smooth P+1=2^3*3^3*103*367*1153*23293*7786127 P-1=2*137*8413609*740628413011
2021-05-06 [M]8805827[/M] B1=500000 B2=26500000 start=2/7 4175293398605056163377 71.822 bits P-1 smooth P+1=2*29*37*239*236917*34360791211 P-1=2^4*3^2*479*479*19559*351457*8805827
2021-05-07 [M]1924751[/M] B1=2500000 B2=2500000 start=2/7 4106464775643002298257 71.798 bits P+1 smooth P+1=2*3*23*109*1621*4919*15241*2246411 P-1=2^4*19*43*263*1924751*620577121
2021-05-07 [M]14034199[/M] B1=1260000 B2=69300000 start=2/7 91301911349511932534929 76.273 bits P-1 smooth P+1=2*5*61*149675264507396610713 P-1=2^4*3*23*734017*8028179*14034199
2021-05-10 [M]8676163[/M] B1=550000 B2=550000 start=2/7 538825884739595715103 68.868 bits P-1 smooth P+1=2^5*19*43*15973*568091*2271287 P-1=2*3*2953*12163*288181*8676163
2021-05-10 [M]14037587[/M] B1=500000 B2=23000000 start=2/7 4451469611034755351513689 81.881 bits P-1 smooth P+1=2*5*59*179*191*199*769*1442067492049 P-1=2*23*3803*5639*75037*49411*1217077
2021-05-12 [M]7974041[/M] B1=500000 B2=26500000 start=6/5 68707931929349790615353 75.863 bits P-1 smooth P+1=2*3^2*621347*7357529*834964831 P-1=2^3*233*252913*7974041*18277271
2021-05-12 [M]8815291[/M] B1=500000 B2=26000000 start=6/5 384543458187807031642369 78.347 bits P-1 smooth P+1=2*5*137*280688655611537979301 P-1=2^8*3^2*7^2*17*3037*7484047*8815291
2021-05-19 [M]505877[/M] B1=5000000 B2=365000000 start=2/7 44167102513715082525497417 85.191 bits P+1 smooth P+1=2*3*79*38729*41413*724729*80162629 P-1=2^3*421*17417*505877*1488361886393
2021-05-21 [M]317557[/M] B1=7500000 B2=592500000 start=2/7 19609089470736235875780917849 93.986 bits P+1 smooth P+1=2*3^2*5^2*7*11*1613*1774667*3473221*56920739 P-1=2^3*317557*7718728240416773947583
2021-05-21 [M]10436197[/M] B1=750000 B2=750000 start=6/5 107236434531058304052209 76.505 bits P-1 smooth P+1=2*3*5*7^2*59*521*2153299*1102125263 P-1=2^4*31*71*506479*576101*10436197
2021-05-22 [M]10436197[/M] B1=1500000 B2=40000000 start=2/7 30961714233434346754568163721 94.644 bits P-1 smooth P+1=2*7^2*212840241931*1484380287573919 P-1=2^3*3^2*5*211*6257*6395981*975943*10436197
2021-05-24 [M]318007[/M] B1=7500000 B2=592500000 start=2/7 42491473899696171527294085337 95.101 bits P-1 smooth P+1=2*109*728831*267435136387605011711 P-1=2^3*3^2*19*337*2131*288577*318007*471307069
2021-05-25 [M]4006333[/M] B1=2500000 B2=2500000 start=2/7 720987230884185534870238439 89.220 bits P+1 smooth P+1=2^3*3*5*11*317*19433*55663*881743*1806533 P-1=2*53*273787*4006333*6201001559113
2021-05-25 [M]4006423[/M] B1=2500000 B2=2500000 start=2/7 2993775828088355588457143 81.308 bits P+1 smooth P+1=2^3*3^2*7*193*593*2927*8123*47279*46171 P-1=2*19*6661*4006423*2952156977603
2021-05-30 [M]9903857[/M] B1=1200000 B2=67200000 start=2/7 112297389502182125412007 76.572 bits P-1 smooth P+1=2^3*43*7949*41067543051408443 P-1=2*3^3*11*53*57193*6297383*9903857
2021-06-01 [M]1891277[/M] B1=5000000 B2=365000000 start=2/7 9032797260237960505049348164007 102.833 bits P+1 smooth P+1=2^3*3*29^2*139*197*263*661*2591*159179*227941607 P-1=2*37*1891277*18189535877*3548246508311
2021-06-06 [M]4021001[/M] B1=2500000 B2=172500000 start=2/7 85759933205284860056473 76.183 bits P-1 smooth P+1=2*43*47*269*189407*416427900659 P-1=2^3*3^2*7*268439*4021001*157642787
2021-06-07 [M]1717043[/M] B1=5000000 B2=365000000 start=2/7 36120234091485938570203343 84.901 bits P+1 smooth P+1=2^4*3*379*5861*39829*1209811*7030423 P-1=2*7*151*1717043*9950946743408221
2021-06-09 [M]4030357[/M] B1=2500000 B2=177500000 start=2/7 307255907312376282607943 78.024 bits P+1 smooth P+1=2^3*3*43*733*70709*1804129*3184009 P-1=2*7*109*272381*4030357*183411001
2021-06-13 [M]4034971[/M] B1=2500000 B2=172500000 start=2/7 35646482986018490883974514241 94.848 bits P-1 smooth P+1=2*317*56224736571007083413209013 P-1=2^6*3*5*13*18119*1333411*4034971*29299717
2021-06-14 [M]4032491[/M] B1=2500000 B2=172500000 start=2/7 7131831162329673363401 72.595 bits P+1 smooth P+1=2*3^2*1839949*1975823*108987007 P-1=2^3*5^2*463157*4032491*19092791
[/CODE]

James Heinrich 2021-04-23 18:44

[QUOTE=kriesel;576642]Start etc. breakdown is[/QUOTE]That breakdown is now also available at the bottom of the [url=https://www.mersenne.org/report_pplus1/]report page[/url].

chalsall 2021-04-23 20:01

[QUOTE=chalsall;576646]A question just hit me... Is the space explored with P+1 different as a function of the B1 and B2 values?[/QUOTE]

Thanks for the edification wrt my first paragraph.

My second question stands. (Please forgive me. I'm a coder / sys-admin. I barely understand half of the maths conversations held around these-here-parts... :smile:)

chalsall 2021-04-23 21:11

[QUOTE=chalsall;576661]Thanks for the edification wrt my first paragraph.[/QUOTE]

Sorry... Juggling too many things at the moment...

That was meant to read thanks to James and masser for the knowledge. What you wrote makes sense.

It's kinda cool how fast stuff happens here! :tu:

It might also be a good example to point to demonstrate how deterministic code can quickly enter rarified domains where empirical data and statistics have to be brought to bear to properly understand what's to be expected. A bit like the halting problem.

I'm serious when I say coding up and then running the Mandelbrot set on my C64 is when I became agnostic (from atheist). :chalsall:

LaurV 2021-04-24 04:19

When you open the pipe for P+1 auto assignments call me. I will put ~30 cores to it for a week or more.

LaurV 2021-04-24 07:24

[QUOTE=kriesel;576626]Or P+1 relevant columns, added to that page, that already has TF and P-1 combined concisely?
[/QUOTE]
+1 :tu:

Miszka 2021-04-24 14:24

I tried:
Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,1000000000,"6000199"
Result:
P+1 found a factor in stage #1, B1=10000000.
UID: mikr/Nitron, M1000033 has a factor: 6000199 (P+1, B1=10000000)
The programme didn't continue.
What did I do wrong that stage #2 wasn't executing?

axn 2021-04-24 14:36

[QUOTE=Miszka;576735]I tried:
Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,1000000000,"6000199"
Result:
P+1 found a factor in stage #1, B1=10000000.
UID: mikr/Nitron, M1000033 has a factor: 6000199 (P+1, B1=10000000)
The programme didn't continue.
What did I do wrong that stage #2 wasn't executing?[/QUOTE]

Could be a bug. However, I am confused about the bolded parameter: Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,[B]1000000000[/B],"6000199"

It is supposed to be the depth of TF in bits. The value you've given (1000000000) is way too high. Don't know if that has anything to do with the bug.

masser 2021-04-24 14:56

[QUOTE=Miszka;576735]I tried:
Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,1000000000,"6000199"
Result:
P+1 found a factor in stage #1, B1=10000000.
UID: mikr/Nitron, M1000033 has a factor: 6000199 (P+1, B1=10000000)
The programme didn't continue.
What did I do wrong that stage #2 wasn't executing?[/QUOTE]

How much memory (RAM) does your machine have?

The worktodo line below requires 3.6 Gb of RAM on my machine; I would guess that a larger B1 would imply a larger B2 and require more RAM.
[CODE]
Pplus1=N/A,1,2,1000033,-1,[B]1000000[/B],250000000,2,[B]66[/B],"6000199"[/CODE]

Here's some of the relevant output:
[CODE]
[Work thread Apr 24 08:50] M1000033 stage 1 is 99.59% complete. Time: 1.689 sec.
[Work thread Apr 24 08:50] M1000033 stage 1 complete. 3903098 transforms. Time: 331.811 sec.
[Work thread Apr 24 08:50] Stage 1 GCD complete. Time: 0.090 sec.
[Work thread Apr 24 08:50] With trial factoring done to 2^66, optimal B2 is 64*B1 = 64000000.
[Work thread Apr 24 08:50] Chance of a new factor assuming no ECM has been done is 0.7%
[Work thread Apr 24 08:50] D: 2730, relative primes: 9507, stage 2 primes: 3706588, pair%=97.25
[Work thread Apr 24 08:50] Using 3603MB of memory.
[Work thread Apr 24 08:50] Stage 2 init complete. 30085 transforms. Time: 6.130 sec.
[Work thread Apr 24 08:50] M1000033 stage 2 is 0.51% complete. Time: 1.682 sec.
[Work thread Apr 24 08:50] M1000033 stage 2 is 1.02% complete. Time: 1.694 sec.
[Work thread Apr 24 08:50] M1000033 stage 2 is 1.53% complete. Time: 1.675 sec.[/CODE]

Miszka 2021-04-24 14:59

[QUOTE=axn;576736]Could be a bug. However, I am confused about the bolded parameter: Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,[B]1000000000[/B],"6000199"

It is supposed to be the depth of TF in bits. The value you've given (1000000000) is way too high. Don't know if that has anything to do with the bug.[/QUOTE]

Clearly I misunderstood the parameter you indicated. I thought it was about the B2 increase.

Miszka 2021-04-24 15:01

[QUOTE=masser;576738]How much memory (RAM) does your machine have?
...
[/QUOTE]

32GB

masser 2021-04-24 15:06

[QUOTE=Miszka;576741]32GB[/QUOTE]

I would try the worktodo line that I posted above; if that works, maybe try subsequent doublings of B1 to see if there's a value where the code stops working. Report back here. It might be a bug, if the code doesn't gracefully adapt to constrained memory conditions yet.

OR maybe the code doesn't behave if the TF bit level is set to a value that is much too high.

ATH 2021-04-24 16:04

30.6b2:
"KeepPplus1SaveFiles=0" is default and you have to use "KeepPplus1SaveFiles=1" to turn it on, which is opposite of what undoc.txt states.



When you do P-1/P+1 on Fermat numbers: Pminus1=1,2,<2[SUP]n[/SUP]>,1 the server automatically notices and calls it Fn and fills out the known factors in the P-1/P+1 line, which is awesome.

Could this be possible for Mersenne numbers that it checks for known factors on the server, and fills them in? or would it strain the server too much?
If possible we could use some notation like empty quotation marks "" in the factors spot or zero: "0" to have it fetch automatically, just in case we want to retain the option to run curves on the full Mersenne number.

Prime95 2021-04-24 17:00

[QUOTE=axn;576736]Could be a bug. However, I am confused about the bolded parameter: Pplus1=N/A,1,2,1000033,-1,10000000,250000000,2,[B]1000000000[/B],"6000199"

It is supposed to be the depth of TF in bits. The value you've given (1000000000) is way too high. Don't know if that has anything to do with the bug.[/QUOTE]

The sieve depth must be less than 100. For reasons unknown (code was cut/pasted from parsing P-1 lines) the parsing stops with a sieve depth >= 100.

Miszka 2021-04-24 18:31

[QUOTE=masser;576742]I would try the worktodo line that I posted above; if that works, maybe try subsequent doublings of B1 to see if there's a value where the code stops working. Report back here. It might be a bug, if the code doesn't gracefully adapt to constrained memory conditions yet.

OR maybe the code doesn't behave if the TF bit level is set to a value that is much too high.[/QUOTE]

For Pplus1=N/A,1,2,1000033,-1,1000000,250000000,2,[B]66[/B],"6000199"

[Apr 24 18:59] Worker starting
[Apr 24 18:59] Setting affinity to run worker on CPU core #2
[Apr 24 18:59]
[Apr 24 18:59] P+1 on M1000033 with B1=1000000, B2=TBD
[Apr 24 18:59] Setting affinity to run helper thread 1 on CPU core #3
[Apr 24 18:59] Setting affinity to run helper thread 2 on CPU core #4
[Apr 24 18:59] Setting affinity to run helper thread 3 on CPU core #5
[Apr 24 18:59] Using FMA3 FFT length 48K, Pass1=768, Pass2=64, clm=2, 4 threads
[Apr 24 19:06] M1000033 stage 1 complete. 4357352 transforms. Time: 407.104 sec.
[Apr 24 19:06] Stage 1 GCD complete. Time: 0.083 sec.
[Apr 24 19:06] With trial factoring done to 2^66, optimal B2 is 64*B1 = 64000000.
[Apr 24 19:06] Chance of a new factor assuming no ECM has been done is 0.7%
[Apr 24 19:06] D: 2730, relative primes: 9507, stage 2 primes: 3706588, pair%=97.25
[Apr 24 19:06] Using 3603MB of memory.
[Apr 24 19:06] Stage 2 init complete. 30085 transforms. Time: 5.316 sec.
[Apr 24 19:12] M1000033 stage 2 complete. 3851030 transforms. Time: 329.027 sec.
[Apr 24 19:12] Stage 2 GCD complete. Time: 0.083 sec.
[Apr 24 19:12] M1000033 completed P+1, B1=1000000, B2=64000000, Wi4: 1A33F4F5

As you can see with correct parameters it works fine. My mistake was that I set the [I]how_far_factored[/I] parameter too high (I misinterpreted it)

masser 2021-04-24 19:59

[QUOTE=ATH;576652]7 factors out of 1599 curves: 2 P+1 smooth, 4 P-1 smooth, 1 both P+1/P-1 smooth


List sorted by time found, I might update this list for a bit with new factors

[CODE]
2021-04-22 06:48:35.610 [M]287873[/M] B1=10000000 B2=1000000000 start=2/7 167460871862758047584571103871 97 bits P+1/P-1 smooth P+1=2^7*3*7*149*31153*137737*1326817*73440613 P-1=2*5*287873*7162339*55539811*146235611
2021-04-23 03:15:32.123 [M]1600967[/M] B1=2000000 B2=146000000 start=2/7 20960634373165261374400215271 94 bits P-1 smooth P+1=2^3*11*13*16047334103*1141761785856421 P-1=2*3^4*5*219889*864733*1600967*85006373
2021-04-23 10:25:31.597 [M]1601783[/M] B1=2000000 B2=146000000 start=2/7 5920048152491971978196689064497 102 bits P-1 smooth P+1=2*7*1034951899*408579937608148822093 P-1=2^4*3*191*6637*10103*53017*1601783*113399107
2021-04-23 11:04:21.267 [M]296663[/M] B1=10000000 B2=500000000 start=2/7 33472736879943425145991507583 95 bits P+1 smooth P+1=2^7*3^2*7*79*3361*41897*9979727*37388971 P-1=2*23*149*34963*296663*470841841030657
2021-04-23 11:51:30.253 [M]1602143[/M] B1=2000000 B2=146000000 start=2/7 43496906104874570142712609 85 bits P-1 smooth P+1=2*5*7*11*56489488447889052133393 P-1=2^5*3^3*15241*17977*1602143*114686597
2021-04-23 16:33:29.030 [M]298307[/M] B1=10000000 B2=500000000 start=2/7 14600351802304915562080191901748143 113 bits P-1 smooth P+1=2^4*17*53677763979062189566471293756427 P-1=2*3*59*119591*298307*1512431*2126617*359445677
2021-04-24 10:32:21.530 [M]92821[/M] B1=40000000 B2=5240000000 start=2/7 6042060144547476324897523777583 102 bits P+1 smooth P+1=2^4*3*23*131*331*1039*2791*9151*19069*249427781 P-1=2*7*92821*3105342439*1497273943641227
[/CODE][/QUOTE]

Got a 6/5:

[CODE]{"status":"F", "exponent":14033251, "worktype":"P+1", "factors":["105040117068459648578879"], "b1":500000, "b2":10000000, "start":"6/5", "fft-length":737280, "security-code":"6B8F2965", "program":{"name":"Prime95", "version":"30.6", "build":1, "port":8}, "timestamp":"2021-04-24 18:17:38"}[/CODE]

ATH 2021-04-24 20:40

[QUOTE=ATH;576747]30.6b2:
"KeepPplus1SaveFiles=0" is default and you have to use "KeepPplus1SaveFiles=1" to turn it on, which is opposite of what undoc.txt states.
[/QUOTE]

When turning this on doing P+1 in the 53K range, some files are 7 KB while others are 200+ MB, I think because they contain stage2 data right?
Is it possible to leave the small files just after stage1 completed?

Prime95 2021-04-24 20:46

[QUOTE=ATH;576768]When turning this on doing P+1 in the 53K range, some files are 7 KB while others are 200+ MB, I think because they contain stage2 data right?
Is it possible to leave the small files just after stage1 completed?[/QUOTE]

If any big files are left behind, that is a bug.

James Heinrich 2021-04-24 21:02

[QUOTE=ATH;576652]List sorted by time found, I might update this list for a bit with new factors[/QUOTE]Or you could just look at the live report:
[url]https://www.mersenne.org/report_pplus1/?f=1[/url]

firejuggler 2021-04-24 21:29

The factor fpund for M14033251 is really NOT smooth,
k= 19^2 *10 367157 737 549


I do not really understand how pp1 work

masser 2021-04-24 21:36

[QUOTE=firejuggler;576772]The factor fpund for M14033251 is really NOT smooth,
k= 19^2 *10 367157 737 549
[/QUOTE]

You're stuck on thinking of factors as f = 2kp+1.

The factor is: f = 105040117068459648578879
The factor+1 is: f+1 = 105040117068459648578880 = 2^6 × 3 × 5 × 13 × 47 × 2377 × 5743 × 5953 × 2 203631
So, if I had chosen B1 = 6000 and B2 = 2500000, with enough P+1 runs, I would have found this factor.

With P+1 we're looking for f+1 to be smooth, or f+1 = 2kp+2 = 2(kp+1) => kp+1 has to be smooth.

firejuggler 2021-04-24 21:54

Ah, thank you for your explanation.

James Heinrich 2021-04-24 22:05

[strike]duplicate post[/strike]

R. Gerbicz 2021-04-24 23:29

[QUOTE=Prime95;576305]Prime95 version 30.6 build 2 is available.

I consider the new features to be lightly tested. No need to upgrade unless you are doing ECM or care to try out the new P+1 factoring work.
[/QUOTE]

Have you included p as a factor in the p+1 (!!) method? Looks like not...
If really not then you will find no p-1 factors where p>B1 (unless with some luck), and all other factors except at most one are smaller than B1, the largest one is <=B2 as usual.
My small example:

M86743 has a factor q=5867520663913, here q-1=2^3*3*19*86743*148339

With the P+1 method, B1=50000,B2=2000000, nthrun=1, so with:
Pplus1=N/A,1,2,86743,-1,50000,2000000,1,40

[CODE]
[Work thread Apr 25 00:55] P+1 on M86743 with B1=50000, B2=2000000
[Work thread Apr 25 00:55] Chance of finding a factor assuming no ECM has been done is an estimated 7.78%
[Work thread Apr 25 00:55] Using FMA3 FFT length 4K
[Work thread Apr 25 00:55] M86743 stage 1 is 8.92% complete. Time: 0.103 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 18.25% complete. Time: 0.105 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 27.45% complete. Time: 0.103 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 36.68% complete. Time: 0.103 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 45.98% complete. Time: 0.102 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 55.37% complete. Time: 0.103 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 64.64% complete. Time: 0.102 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 73.64% complete. Time: 0.102 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 83.02% complete. Time: 0.102 sec.
[Work thread Apr 25 00:55] M86743 stage 1 is 92.29% complete. Time: 0.103 sec.
[Work thread Apr 25 00:55] M86743 stage 1 complete. 217103 transforms. Time: 1.115 sec.
[Work thread Apr 25 00:55] Stage 1 GCD complete. Time: 0.003 sec.
[Work thread Apr 25 00:55] D: 840, relative primes: 1744, stage 2 primes: 143800, pair%=95.24
[Work thread Apr 25 00:55] Using 57MB of memory.
[Work thread Apr 25 00:55] Stage 2 init complete. 5119 transforms. Time: 0.095 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 12.63% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 25.68% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 38.86% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 52.03% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 65.02% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 77.73% complete. Time: 0.112 sec.
[Work thread Apr 25 00:55] M86743 stage 2 is 90.49% complete. Time: 0.111 sec.
[Work thread Apr 25 00:55] M86743 stage 2 complete. 154981 transforms. Time: 0.862 sec.
[Work thread Apr 25 00:55] Stage 2 GCD complete. Time: 0.003 sec.
[Work thread Apr 25 00:55] M86743 completed P+1, B1=50000, B2=2000000, Wi8: 024544B3
[/CODE]

But with B1=100000,B2=2000000 we can get:
Pplus1=N/A,1,2,86743,-1,100000,2000000,1,40
[CODE]
[Work thread Apr 25 00:52] P+1 on M86743 with B1=100000, B2=2000000
[Work thread Apr 25 00:52] Chance of finding a factor assuming no ECM has been done is an estimated 8.51%
[Work thread Apr 25 00:52] Using FMA3 FFT length 4K
[Work thread Apr 25 00:52] M86743 stage 1 is 4.33% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 9.00% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 13.64% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 18.23% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 22.86% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 27.53% complete. Time: 0.105 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 32.23% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 36.73% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 41.38% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 46.04% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 50.62% complete. Time: 0.106 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 55.24% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 59.66% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 64.43% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 69.11% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 73.69% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 78.30% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 82.81% complete. Time: 0.104 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 87.54% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 92.17% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 is 96.58% complete. Time: 0.103 sec.
[Work thread Apr 25 00:52] M86743 stage 1 complete. 435081 transforms. Time: 2.253 sec.
[Work thread Apr 25 00:52] Stage 1 GCD complete. Time: 0.003 sec.
[Work thread Apr 25 00:52] D: 840, relative primes: 1793, stage 2 primes: 139341, pair%=93.26
[Work thread Apr 25 00:52] Using 59MB of memory.
[Work thread Apr 25 00:52] Stage 2 init complete. 5261 transforms. Time: 0.081 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 12.65% complete. Time: 0.112 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 25.76% complete. Time: 0.113 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 39.18% complete. Time: 0.112 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 52.88% complete. Time: 0.112 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 65.95% complete. Time: 0.112 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 78.77% complete. Time: 0.115 sec.
[Work thread Apr 25 00:52] M86743 stage 2 is 91.69% complete. Time: 0.113 sec.
[Work thread Apr 25 00:52] M86743 stage 2 complete. 153070 transforms. Time: 0.863 sec.
[Work thread Apr 25 00:52] Stage 2 GCD complete. Time: 0.003 sec.
[Work thread Apr 25 00:52] P+1 found a factor in stage #2, B1=100000, B2=2000000.
[Work thread Apr 25 00:52] M86743 has a factor: 5867520663913 (P+1, B1=100000, B2=2000000)
[Work thread Apr 25 00:52] No work to do at the present time. Waiting.
[/CODE]

Actually it will discover q only when B1>=p=86743, since for this q the p factor is the 2nd largest factor.

henryzz 2021-04-25 10:55

Is there any logic behind trying to identify which starting values actually do P+1? For example, if we multiply the original composite by a small prime that will be found with very small P+1 bounds or known P+1 factors but won't be found with P-1 then we will know that it is actually doing P+1. The downside of this is that we will have to run P+1 on this larger composite which may mess up the efficiency of the arithmetic for Mersenne numbers. This could be better than running 2 or 3 times though.

ATH 2021-04-25 13:04

[QUOTE=henryzz;576824]Is there any logic behind trying to identify which starting values actually do P+1? For example, if we multiply the original composite by a small prime that will be found with very small P+1 bounds or known P+1 factors but won't be found with P-1 then we will know that it is actually doing P+1. The downside of this is that we will have to run P+1 on this larger composite which may mess up the efficiency of the arithmetic for Mersenne numbers. This could be better than running 2 or 3 times though.[/QUOTE]

The bounds depends on the smoothness of "the unknown factor we are trying to find +1" not on the original composite, and the same with the seeds x0. If they work or not depends on the unknown factor:
"[I]However not all seeds will succeed: only half of the seeds 'x0' work for P+1
(namely those where the Jacobi symbol of x0^2-4 and P is -1.) Unfortunately,
since P is usually not known in advance, there is no way to ensure that this
holds.[/I]"

henryzz 2021-04-25 15:51

[QUOTE=ATH;576829]The bounds depends on the smoothness of "the unknown factor we are trying to find +1" not on the original composite, and the same with the seeds x0. If they work or not depends on the unknown factor:
"[I]However not all seeds will succeed: only half of the seeds 'x0' work for P+1
(namely those where the Jacobi symbol of x0^2-4 and P is -1.) Unfortunately,
since P is usually not known in advance, there is no way to ensure that this
holds.[/I]"[/QUOTE]

I forgot it was the factor and not the composite. :down:

ATH 2021-04-25 22:25

Moved the posts regarding missing P-1 / P+1 work to the "<20M unfactored" thread:
[url]https://mersenneforum.org/showthread.php?t=22476[/url]

storm5510 2021-04-26 00:41

[QUOTE=Prime95;576305]Prime95 version 30.6 build 2 is available.

I consider the new features to be lightly tested. No need to upgrade unless you are doing ECM or care to try out the new P+1 factoring work...
[/QUOTE]

I like to run ECM's of varying sizes at time. I saw something different in the Stage 2 output after I started running this version. Something called "D-Block" in each output line. Is there a special significance to this?

Prime95 2021-04-26 01:43

[QUOTE=storm5510;576882]Something called "D-Block" in each output line. Is there a special significance to this?[/QUOTE]

No special significance. It is simply a measure of how far along you are in stage 2.

Prime95 2021-04-27 03:04

[QUOTE=gLauss;576509]Not sure if it is stupid what I'm doing (it most probably is), but I'm getting stuck for an unreasonable amount of time in the "initializing step" for stage 2 and the CPU is only used a few percent there.
Why did it take 6 minutes to initiliaze after calculating stage 1 GCD? Something strange must be going on here. might be a bug...[/QUOTE]

Fixed in 30.6 build 3

Prime95 2021-04-27 03:20

30.6 build 3 is ready

Prime pairing is ever so slightly better. Lengthy stage 2 init problem on small exponents fixed. Several other minor bug fixes and improvements.

Download links:
[B][COLOR="Red"]
Before downloading, make sure prime.spl is uploaded to the server. There was a bug in spool file format for versions 30.6b1 and 30.6b2. Version 30.6b3 will have no trouble creating and reading spool files that are compatible with version 30.5 and earlier.
[/COLOR][/B]
Windows 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b3.win64.zip[/URL]
Linux 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b3.linux64.tar.gz[/URL]

LaurV 2021-04-27 04:28

b2 was reasonably working in a win7 computer, 10 cores, 5 workers x2 cores each, therefore I decided to install it in a second computer, 18 cores, 6 workers, x3 cores each. Now, there are two differences, first this is a win10, second, b3 was issued meantime, so I installed b3.

I am stuck with "error larger than 0.5, trying to resume from before", in all 6 workers, infinite loop. There is no "before", this is fresh start. When I "stop" all workers, worker 2 and 3 won't stop, but continue to flood the screen. If I exit, it hangs, I see the green icon, but it does nothing and it needs to be stopped from task man. In all this time (screen flooding) it takes about 4% cpu resources.

edit: b2 is the same in this computer. Both b2 and b3 will run normal when non-P+1 assignments provided in worktodo. The "flooding" is only if P+1 lines present. Lines are copied from the other computer, where it worked.

LaurV 2021-04-27 05:02

1 Attachment(s)
OTOH, he uses 672k, which should be right for this range (if you look to my last reported results on P+1, I am tickling the 13M expos, took the first 30 or so, for a spin)
[ATTACH]24774[/ATTACH]

Prime95 2021-04-27 05:59

@Laurv: can you supply the worktodo.txt?

LaurV 2021-04-27 07:21

Yep, luckily I put them in comments instead of deleting them, to be able to continue my work. Here are the first ones of each worker. You can split them in workers any way, the effect is the same, this only happens on the win10/x299 computer, no idea if connected.



[CODE]
Pplus1=N/A,1,2,13001147,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001173,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001237,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001239,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001269,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001291,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001323,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001341,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001347,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001507,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001539,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001683,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001731,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001897,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13001993,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13002083,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13002149,-1,2000000,100000000,2,69
Pplus1=N/A,1,2,13002173,-1,2000000,100000000,2,69
[/CODE]


Please note that the same assignments work on the other computer, with both b2 and b3.

tha 2021-04-27 08:32

[QUOTE=Prime95;576565]Fixed in next build. Thanks.[/QUOTE]

Downloaded the new version. It now does not report anything about the composite factor reported by the client to the server. I would expect a list of factors already known to the server and, if present, a new factor as a new factor.

I must admit I have not found a new factor with the new version so I continue to do P-1 factoring on exponents with three factors or more already found and no or little previous P-1 done before. Will report when I have found one.

R. Gerbicz 2021-04-27 13:57

[QUOTE=R. Gerbicz;576783]With the P+1 method, B1=50000,B2=2000000, nthrun=1, so with:
Pplus1=N/A,1,2,86743,-1,50000,2000000,1,40
[/QUOTE]

[QUOTE=Prime95;576974]30.6 build 3 is ready
[/QUOTE]

This is still doesn't find q in the first test.
More greedy run, try:
Pplus1=N/A,1,2,86743,-1,86742,2000000,1,40
Pplus1=N/A,1,2,86743,-1,86743,2000000,1,40

Only the 2nd one finds the q=5867520663913=2^3*3*19*86743*148339+1=2^3*3*19*148339*p+1, when B1 reached p, but including p in all runs you would find it much earlier [as in the normal P-1 test when you include p as a factor in the beginning].

OK, overworking on smallish p values it is still not a problem if B1>=p, but
in the optimal setup B1<<p for the wavefront exponents [obviously since the product of primes<x is roughly exp(x)>2^x], so you miss all of these factors, unless p is the biggest factor of q-1 and p<=B2.

axn 2021-04-27 14:26

[QUOTE=R. Gerbicz;576996]This is still doesn't find q in the first test.[/QUOTE]

I second this request. There is virtually no cost in throwing in a 'p' in the B1 product. If it helps in finding an odd P-1 factor here or there, that is always a win.

kriesel 2021-04-27 14:26

Pminus1 of 660M fails to start on AVX512F Xeon Phi
 
After a previous assignment completed, I got something like the following in worker 1 of a 4-worker instance of prime95 v30.5b1. The issue was repeatable. The related worktodo line is "Pminus1=N/A,1,2,660000031,-1,3600000,180000000" and seems to be correctly interpreted, but not running for some reason. (Perhaps the Pminus1 code path was not updated for AVX512 limits? I dunno.)
[CODE][Apr 27 00:27] P-1 on M660000031 with B1=3600000, B2=180000000
[Apr 27 00:27] Cannot initialize FFT code, errcode=1002
[Apr 27 00:27] Worker stopped.[/CODE]After updated to v30.6b3, the issue is still repeatable.
It is correctly identifying the cpu.[CODE][Main thread Apr 27 08:27] Mersenne number primality test program version 30.6
[Main thread Apr 27 08:27] Optimizing for CPU architecture: Xeon Phi, L2 cache size: 34x1 MB
[Main thread Apr 27 08:27] Starting workers.[/CODE][CODE][Apr 27 08:27] Worker starting
[Apr 27 08:27] Setting affinity to run worker on CPU core #1
[Apr 27 08:27]
[Apr 27 08:27] P-1 on M660000031 with B1=3600000, B2=180000000
[Apr 27 08:27] Cannot initialize FFT code, errcode=1002
[Apr 27 08:27] Worker stopped.[/CODE]This is occurring on a Xeon Phi 7250, 68 core, on Windows 10 build 20H2 19042.928. It has been running in 4-worker, 17 core per worker configuration for months.
Prime95 CPU options show it is recognizing it as both FMA3 and AVX512F capable, which should handle exponents up to 920M and ~1169M. [URL]https://www.mersenneforum.org/showpost.php?p=546374&postcount=8[/URL]
I have previously benchmarked up to 64M fft length on this system. [URL]https://www.mersenneforum.org/showpost.php?p=563304&postcount=11[/URL]

The string "1002" does not appear in the documentation files distributed with the prime95 program.
An advanced search in the mersenne forum for the string "errcode" turned up nothing recent.

It should not be hitting in AVX512F P-1, the sse2 or similar fft limit as in [URL]https://mersenneforum.org/showthread.php?p=400845&highlight=errcode#post400845[/URL]

The advice in [URL="https://mersenneforum.org/showthread.php?p=447931&highlight=errcode#post447931"]this 2016 thread[/URL] does not seem to apply here.
There is no "CpuSupport" string in local.txt
Options/CPU reports the correct cpu features for the cpu.
The string "FFT2" is not present in worktodo.txt.

Since the issue occurred immediately upon encountering the worktodo entry, before any progress was made and logged in the worker window, and there is no file in the working directory identifiable by name as relating to the 660000031 exponent, it seems unlikely to be an issue with invalid or corrupt save files.
It is not an issue of disk space, with over 200 GB free.

gLauss 2021-04-27 14:32

[QUOTE=R. Gerbicz;576996]This is still doesn't find q in the first test.
[/QUOTE]
Yes, this should be fixed. I have mentioned this in post [url=https://mersenneforum.org/showpost.php?p=576609&postcount=188]#188[/url], too.

I am wondering what the optimal P-1/P+1/ECM strategy is and what P+1 bounds I should choose for my exponents. My current strategy is:
1) Get current ECM level, e.g. B1=3M.
2) George said that the P+1 should be done to two stages above the current ECM level, e.g. B1=50M.
3) Before the P+1 is done, one should do a P-1 run with at least twice the bound of the P+1, e.g. B1=100M.
4) If there is no ECM done, then I think it is optimal to do choose B1 for the P+1 as the current P-1 B1 divided by a factor of 2 to 5.
However, this is only my gutfeeling. Is there someone who can calculate the odds?

axn 2021-04-27 14:34

[QUOTE=kriesel;577002]It should not be hitting in AVX512F P-1, the sse2 or similar fft limit as in [URL]https://mersenneforum.org/showthread.php?p=400845&highlight=errcode#post400845[/URL][/QUOTE]

Not sure what you mean, but... AVX512 has worse FFT limits compared to FMA3. The error message looks like the same one in that thread, implying that it exceeds largest available FFT size. IIUC, the fft size selection is done by gwnum code and not any of the testing code.

How do you know that AVX512 can work correctly on this number?

EDIT:- Just noticed the reference to the 64M fft benchmark. Hmm...

kriesel 2021-04-27 15:49

[QUOTE=axn;577004]AVX512 has worse FFT limits compared to FMA3..[/QUOTE]Not sure what you mean by worse (maybe one implementation has worse bits/word limit for the same fft length), but FMA3 is documented to go to 50M fft length / 920M exponent capability, AVX512 to 64M fft, 1169M exponent limit, plus I've already successfully benchmarked up to 64M fft length on that same system as refused to run 660M Pminus1. Seems inconsistent.

Thanks for looking at it. It's puzzling. If I knew a supported syntax to force specific bounds via PFactor worktodo entry, I would give that a try too. But unnecessary; without specifying bounds, it has problems with the PFactor worktodo form also:[CODE]PFactor=N/A,1,2,660000031,-1,84,2[/CODE][CODE][Apr 27 09:58] Worker starting
[Apr 27 09:58] Setting affinity to run worker on CPU core #1
[Apr 27 09:58] Optimal P-1 factoring of M660000031 using up to 12288MB of memory.
[Apr 27 09:58] Assuming no factors below 2^84 and 2 primality tests saved if a factor is found.
[Apr 27 09:58] Optimal bounds are B1=4172000, B2=165474000
[Apr 27 09:58] Chance of finding a factor is an estimated 4.11%
[Apr 27 09:58]
[Apr 27 09:59] Cannot initialize FFT code, errcode=1002
[Apr 27 09:59] Worker stopped.[/CODE]Also fails in P-1 attempt if given a PRP line requiring it. (PRP=N/A,1,2,660000031,-1,84,2)
Also fails to launch a PRP for it; with worktodo entry [CODE]PRP=N/A,1,2,660000031,-1,84,0[/CODE][CODE][Apr 27 10:13] Worker starting
[Apr 27 10:13] Setting affinity to run worker on CPU core #1
[Apr 27 10:13] PRP cannot initialize FFT code for M660000031, errcode=1002
[Apr 27 10:13] Number sent to gwsetup is too large for the FFTs to handle.
[Apr 27 10:13] Worker stopped.[/CODE]PRP=N/A,1,2,630132049,-1,84,0
also reproduces the issue. It accepts 600M as a PRP assignment, requiring >32M fft ok (estimated run time >15 months);[CODE]PRP=N/A,1,2,600023209,-1,84,0[/CODE][CODE][Apr 27 10:36] Setting affinity to run worker on CPU core #1
...
[Apr 27 10:36] Setting affinity to run helper thread 16 on CPU core #17
[Apr 27 10:37] Starting Gerbicz error-checking PRP test of M600023209 using AVX-512 FFT length 34560K, Pass1=2304, Pass2=15K, clm=1, 17 threads
[Apr 27 10:37] Preallocating disk space for the proof interim residues file p600023209.residues
[Apr 27 10:41] PRP proof using power=7x2 and 64-bit hash size.
[Apr 27 10:41] Proof requires 9.6GB of temporary disk space and uploading a 1200MB proof file.[/CODE]Also accepts 613112609 and higher that use the same 34560K length.

I've previously run P-1 at least up to 852.xM on FMA3 (different system, prime95 v29.8b6, Windows 10.)
[CODE][Fri Oct 11 13:38:45 2019]
{"status":"NF", "exponent":852348659, "worktype":"P-1", "b1":5950000, "b2":117512500, "fft-length":52428800, "security-code":"redacted", "program":{"name":"Prime95", "version":"29.8", "build":6, "port":4}, "timestamp":"2019-10-11 18:38:45", "user":"Kriesel", "computer":"peregrine", "aid":"redacted"}[/CODE]Now here is the stranger part; same 660M exponent that initially was a problem, can launch ok in prime95 v30.5b1 on a Xeon Phi 7210 (64 core, 4 worker, Windows 10 same build) for PRP or Pminus1. Comparing local.txt, prime.txt between systems and dropping the 7250 to 16 cores/worker did not resolve it.

kriesel 2021-04-27 17:29

But an even larger exponent launches ok in Pminus1 worktodo form on the 7250 on a 42M fft
[CODE][Apr 27 12:15] P-1 on M754036583 with B1=4500000, B2=200000000
[Apr 27 12:15] Setting affinity to run helper thread 1 on CPU core #2
...
[Apr 27 12:15] Using AVX-512 FFT length 42M, Pass1=2K, Pass2=21K, clm=1, 17 threads
[Apr 27 12:15] Setting affinity to run helper thread 15 on CPU core #16
[Apr 27 12:15] Setting affinity to run helper thread 16 on CPU core #17
[/CODE]

Prime95 2021-04-28 00:42

[QUOTE=LaurV;576978]OTOH, he uses 672k, which should be right for this range (if you look to my last reported results on P+1, I am tickling the 13M expos, took the first 30 or so, for a spin)
[ATTACH]24774[/ATTACH][/QUOTE]

Fixed in build 4.

Until I upload that version, shy away from AVX-512 FFT lengths that are a multiple of 7.

Prime95 2021-04-28 00:48

[QUOTE=R. Gerbicz;576996]This is still doesn't find q in the first test.
More greedy run, try:
Pplus1=N/A,1,2,86743,-1,86742,2000000,1,40
Pplus1=N/A,1,2,86743,-1,86743,2000000,1,40.[/QUOTE]

Including the Mersenne exponent in the P+1 bounds is cheap, but also encourages using the wrong tool to find P-1 factors.

1) P-1 factoring is faster than P+1 factoring.
2) P-1 does not miss any B1/B2 smooth factor-1 values, whereas P+1 factoring misses half of them.

Unless someone points to a good reason to use P+1 factoring to find P-1 factors, I'm not inclined to add the Mersenne exponent to the B1 bound.

Prime95 2021-04-28 00:57

[QUOTE=tha;576982]Downloaded the new version. It now does not report anything about the composite factor reported by the client to the server. I would expect a list of factors already known to the server and, if present, a new factor as a new factor.

I must admit I have not found a new factor with the new version so I continue to do P-1 factoring on exponents with three factors or more already found and no or little previous P-1 done before. Will report when I have found one.[/QUOTE]

The fix I made was to no longer truncate the message you were getting:

[CODE][Comm thread Apr 22 18:59] Already have factor 72324517147464481 for M9994027
[Comm thread Apr 22 18:59] Already [/CODE]

When you say "now does not report anything", what message did you get and what were you expecting?

Prime95 2021-04-28 01:52

[QUOTE=kriesel;577012]But an even larger exponent launches ok in Pminus1 worktodo form on the 7250 on a 42M fft[/QUOTE]

I tried your problem exponent on my laptop with version 30.6b3. It launched OK using a 36M FFT.

I don't know what is going on.

Prime95 2021-04-28 01:59

[QUOTE=gLauss;577003]
I am wondering what the optimal P-1/P+1/ECM strategy is and what P+1 bounds I should choose for my exponents. My current strategy is:
1) Get current ECM level, e.g. B1=3M.
2) George said that the P+1 should be done to two stages above the current ECM level, e.g. B1=50M.
3) Before the P+1 is done, one should do a P-1 run with at least twice the bound of the P+1, e.g. B1=100M.
4) If there is no ECM done, then I think it is optimal to do choose B1 for the P+1 as the current P-1 B1 divided by a factor of 2 to 5.
However, this is only my gutfeeling. Is there someone who can calculate the odds?[/QUOTE]

As to "George said", more accurate would be "George suggested". Like you, I'm going on a gut feeling.

I do not think there is an optimal strategy. I can say without a doubt that the right order is 1) try P-1, 2) try P+1, 3) try ECM. You want to choose what are seemingly excessive P-1 and P+1 bounds because if you ever decide to increase the bounds you have to re-do all the work that was originally done.

axn 2021-04-28 02:19

[QUOTE=Prime95;577064]Including the Mersenne exponent in the P+1 bounds is cheap, but also encourages using the wrong tool to find P-1 factors.

1) P-1 factoring is faster than P+1 factoring.
2) P-1 does not miss any B1/B2 smooth factor-1 values, whereas P+1 factoring misses half of them.

Unless someone points to a good reason to use P+1 factoring to find P-1 factors, I'm not inclined to add the Mersenne exponent to the B1 bound.[/QUOTE]

Noone is asking about using P+1 runs [B]instead of[/B] P-1 runs to find P-1 factors. But, sometimes machines makes mistakes and P-1 runs could miss factors. If we're running P+1 anyways, and if it is capable of finding some P-1 factors, and sometimes P-1 factors are missed by prior runs, it would be good to let P+1 have the capability to find some of them.

Falkentyne 2021-04-28 04:10

1 Attachment(s)
[QUOTE=Prime95;577072]As to "George said", more accurate would be "George suggested". Like you, I'm going on a gut feeling.

I do not think there is an optimal strategy. I can say without a doubt that the right order is 1) try P-1, 2) try P+1, 3) try ECM. You want to choose what are seemingly excessive P-1 and P+1 bounds because if you ever decide to increase the bounds you have to re-do all the work that was originally done.[/QUOTE]

Hi, I didn't test 30.6 beta 4 but beta 2 and 3 are broken on Rocket Lake (11900k).
testing large FFT with AVX disabled causes all threads to crash at the same time at the end of the first iteration.

30.5 beta 2 works fine.

tha 2021-04-28 08:22

[QUOTE=Prime95;577066]The fix I made was to no longer truncate the message you were getting:

[CODE][Comm thread Apr 22 18:59] Already have factor 72324517147464481 for M9994027
[Comm thread Apr 22 18:59] Already [/CODE]

When you say "now does not report anything", what message did you get and what were you expecting?[/QUOTE]

I did find a new factor this morning, and indeed it does no longer truncate the message:

[CODE]
PrimeNet success code with additional info:
[Comm thread Apr 28 05:52] Composite factor 1433220254355182327305060291296816494851568454069468578543 = 78687058243775872759827977 * 3027792433 * 5563081984303 * 1081354441
[Comm thread Apr 28 05:52] Already have factor 3027792433 for M9011287
[Comm thread Apr 28 05:52] Already have factor 5563081984303 for M9011287
[Comm thread Apr 28 05:52] Already have factor 1081354441 for M9011287
[/CODE]

What I was expecting was that each composite factor reported to the server would still be factorized in the return message. But indeed that does not serve any real purpose. The message 'factoring result was not needed' is fine.

One more request though, for mprime, if one chooses option 4, 'Test/Continue' the first output to the screen is a re-display of the menu. That does not serve any purpose as it is the last on display and the second one only clutters the screen. Since nowadays all terminal screens are scroll-able it would make more sense to have the re-display of the menu following a ^C.

R. Gerbicz 2021-04-28 11:09

[QUOTE=axn;577075]Noone is asking about using P+1 runs [B]instead of[/B] P-1 runs to find P-1 factors. But, sometimes machines makes mistakes and P-1 runs could miss factors. If we're running P+1 anyways, and if it is capable of finding some P-1 factors, and sometimes P-1 factors are missed by prior runs, it would be good to let P+1 have the capability to find some of them.[/QUOTE]

Exactly, add to that there is no error check in P-1. You could get a partial error check in stage one if you are doing a mixed prp/pm1 test, what gpuowl is doing. I could be wrong but you could also get error in gpuowl's P-1 first stage run [this is an error in P-1, not in the prp part].

And what would be the optimal run for a mixed prp/pm1 test?
If a full P+1 test before this [obviously not with gpuowl, this is not supported there] then you lost many factors, need to halt the prp test after the p-1 run. It is possible that the optimal run is using increasing B1,B2 bounds for P+-1 runs.

kriesel 2021-04-28 16:14

Issue reproduced eventually on second system, and a feature request
 
[QUOTE=Prime95;577069]I tried your problem exponent on my laptop with version 30.6b3. It launched OK using a 36M FFT.

I don't know what is going on.[/QUOTE]The plot thickens. The Xeon Phi 7210 could run the M660M P-1, until it couldn't, for unknown reason. Excerpted from the worker window:[CODE][Apr 27 11:12] Worker starting
[Apr 27 11:12] Setting affinity to run worker on CPU core #1
[Apr 27 11:12]
[Apr 27 11:12] P-1 on M660000031 with B1=3600000, B2=180000000
[Apr 27 11:12] Setting affinity to run helper thread 1 on CPU core #2
[Apr 27 11:12] Setting affinity to run helper thread 2 on CPU core #3
[Apr 27 11:12] Setting affinity to run helper thread 7 on CPU core #8
[Apr 27 11:12] Setting affinity to run helper thread 4 on CPU core #5
[Apr 27 11:12] Setting affinity to run helper thread 5 on CPU core #6
[Apr 27 11:12] Using AVX-512 FFT length 36M, Pass1=1536, Pass2=24K, clm=2, 16 threads
[Apr 27 11:12] Setting affinity to run helper thread 3 on CPU core #4
[Apr 27 11:12] Setting affinity to run helper thread 8 on CPU core #9
[Apr 27 11:12] Setting affinity to run helper thread 9 on CPU core #10
[Apr 27 11:12] Setting affinity to run helper thread 10 on CPU core #11
[Apr 27 11:12] Setting affinity to run helper thread 11 on CPU core #12
[Apr 27 11:12] Setting affinity to run helper thread 12 on CPU core #13
[Apr 27 11:12] Setting affinity to run helper thread 13 on CPU core #14
[Apr 27 11:12] Setting affinity to run helper thread 14 on CPU core #15
[Apr 27 11:12] Setting affinity to run helper thread 6 on CPU core #7
[Apr 27 11:12] Setting affinity to run helper thread 15 on CPU core #16
[Apr 27 11:30] M660000031 stage 1 is 0.19% complete. Time: 1109.194 sec.
...
[Apr 27 22:59] M660000031 stage 1 is 7.70% complete. Time: 1049.200 sec.
[Apr 27 23:17] M660000031 stage 1 is 7.89% complete. Time: 1058.078 sec.
[Apr 27 23:19] [B]Worker stopped while running needed benchmarks.[/B]
[Apr 27 23:26] [B]Benchmarks complete, restarting worker.[/B]
[Apr 27 23:26]
[Apr 27 23:26] P-1 on M660000031 with B1=3600000, B2=180000000
[Apr 27 23:26] [B]Cannot initialize FFT code, errcode=1002[/B]
[Apr 27 23:26] [B]Worker stopped.[/B][/CODE]There's nothing about the issue in prime.log. It gets logged in results.txt but there is no additional information[CODE][Tue Apr 27 23:26:58 2021]
Cannot initialize FFT code, errcode=1002[/CODE][B]Is there some debugging I could turn on for this?[/B] Searching for "debug" in the documentation files revealed only the unrelated following in whatsnew.txt (which maybe ought be edited to remove Scott Kurowski's name if he's no longer actively participating):[CODE]New features in Version 16.5 of prime95.exe
-------------------------------------------

1) A new httpnet.dll can provide Scott Kurowski with debugging information.
Simply create a primenet.ini file with these lines:
[Debug]
PacketLog=1
Output file is pnHttp.txt.
[/CODE]Similarly searching for "verbos" yielded only in undoc.txt,[CODE]You can change the amount of information the program outputs while setting affinities.
AffinityVerbosity=0 or 1 (default is 1)
AffinityVerbosityTorture=0 or 1 (default is 0)
AffinityVerbosityTime=0 or 1 (default is 0)
AffinityVerbosityBench=0 or 1 (default is 0)
These settings are made in prime.txt. Zero for less output, one for more output.
The first setting is for all normal work (LL, TF, ECM, P-1, P+1, PRP). The others
are for torture testing, Advanced/Time, and benchmarking.
[/CODE]Also, although there was other work in the worker's section of the worktodo, it stopped dead and left 16 cores idle for hours, until found.
It would be good if there was an option to have it comment out a problem worktodo line, and attempt the next in line, to keep it productive. In this case that worktodo section was (with AID redaction):
[CODE][Worker #1]
Pminus1=N/A,1,2,660000031,-1,3600000,180000000
PRP=(aid),1,2,350021701,-1,79,0[/CODE]so it could have progressed to a different work type, different exponent, different fft length and probably remained productive, by modifying the relevant worktodo section to something like the following and then restarting the worker:
[CODE][Worker #1]
[COLOR=Red][B];[/B][/COLOR]Pminus1=N/A,1,2,660000031,-1,3600000,180000000
[COLOR=red][B];preceding commented out because [Apr 27 23:26] Cannot initialize FFT code, errcode=1002[/B][/COLOR]
PRP=(aid),1,2,350021701,-1,79,0[/CODE]To recap:
It interrupts work it is doing, to do more benchmarks;
it indicates there's an issue to resume the interrupted work in progress;
it doesn't tell WHY it can no longer do what it had previously been doing before interrupted by benchmarking;
there does not appear to be any option for having it tell more about the problem it encountered preventing resumption;
it has other work it could switch to, to avoid loss of throughput, but does not support switching to that automatically.
I had hoped to do P-1 of large exponents on FMA3 and AVX512, but have now found that 2 of the 4 AVX512 (both Xeon Phi, different models) have some undetermined issues with it, unpredictably. Attempting them on the FMA3 i7-8750H or AVX512 i5-1035G1 laptops would take weeks. I may try continuing the ~8% complete M660000031 there anyway, rather than throwing away the progress and running it on gpuowl / Radeon VII (which are deeply queued already).

If a prerelease prime95 version with greater P-1 or fft startup debugging output became available (in forum thread, or by dropbox link in PM?) I would be happy to try it out.


update: An AVX512 i5-1035g1 laptop prime95 v30.6b2, Windows 10 build 20H2 19042.928, 16gb ram) was able to resume the M660M P-1 run in progress, with ETA 3.5 weeks. We'll see how that goes.

Prime95 2021-04-28 16:27

[QUOTE=kriesel;577125]The plot thickens. [/QUOTE]

Try renaming gwnum.txt. Restart prime95, will the exponent now run? If so, email/PM the gwnum.txt file to me. Thanks.

kriesel 2021-04-28 17:37

[QUOTE=Prime95;577127]Try renaming gwnum.txt. Restart prime95, will the exponent now run? If so, email/PM the gwnum.txt file to me. Thanks.[/QUOTE]
On Xeon Phi 7210 that ran M660M P-1 until a prime95-initiated benchmark interrupted progress, then won't resume:

Stop all prime95 workers
rename gwnum.txt wasgwnum.txt
reintstate M660M P-1 as first line of worker 1 worktodo by notepad edit
continue all workers

Issue with fft for M660M P-1 reoccured. Maybe it only reads gwnum.txt on startup. Try again, more thoroughly.

Stop prime95 workers
verify there's no gwnum.txt by name sort in the working directory in explorer
Exit prime95 completely
Use task manager to verify prime95.exe is not running at all
Relaunch prime95 v30.5b1 with a double click

Prime95 resumes M660M P-1, ETA ~1.5 weeks

Check for a gwnum.txt email attachment, George, & thanks!
(at 122KB, much too big for a PM. FYI, Xeon Phi 7250 system that could not even start it, has a 133KB gwnum.txt. Laptop that could continue it has a 9KB gwnum.txt. Stealth size limit?)

Prime95 2021-04-28 19:06

[QUOTE=kriesel;577137]
Prime95 resumes M660M P-1, ETA ~1.5 weeks

Check for a gwnum.txt email attachment,[/QUOTE]

Fixed in 30.6 build 4

Prime95 2021-04-28 21:50

30.6 build 4 is ready

Several bug fixes: Torture tests, AVX-512 P+1 on FFT lengths that are a multiple of 7, best 36M FFT selection

Download links:
[B][COLOR="Red"]
Before downloading, make sure prime.spl is uploaded to the server when upgrading from early versions of 30.6. There was a bug in spool file format for versions 30.6b1 and 30.6b2. Version 30.6b4 will have no trouble creating and reading spool files that are compatible with version 30.5 and earlier plus versions 30.6b3 and later.

[/COLOR][/B]
Windows 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.win64.zip[/URL]
Linux 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.linux64.tar.gz[/URL]
Source code: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.source.zip[/URL]

Falkentyne 2021-04-28 22:10

[QUOTE=Prime95;577159]30.6 build 4 is ready

Several bug fixes: Torture tests, AVX-512 P+1 on FFT lengths that are a multiple of 7, best 36M FFT selection

Download links:
[B][COLOR="Red"]
Before downloading, make sure prime.spl is uploaded to the server when upgrading from early versions of 30.6. There was a bug in spool file format for versions 30.6b1 and 30.6b2. Version 30.6b4 will have no trouble creating and reading spool files that are compatible with version 30.5 and earlier plus versions 30.6b3 and later.

[/COLOR][/B]
Windows 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.win64.zip[/URL]
Linux 64-bit: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.linux64.tar.gz[/URL]
Source code: [URL]https://mersenne.org/ftp_root/gimps/p95v306b4.source.tar.gz[/URL][/QUOTE]

Super fast test on 11900k. Large FFT AVX disabled no longer instacrashes all threads after one iteration. Thank you (can't do more conclusive testing because I'm gaming).

ATH 2021-04-29 01:29

Weird error in mprime 30.5b1 (or primenet server error).

Factors for these 2 exponents got reported twice each: [M]3021587[/M] and [M]3310193[/M]

and for some reason they got added twice to the PRP-CF worktodo line by the server:
PRP=0C618B35F23DD4EB9B3E5D7094A1____,1,2,3021587,-1,99,0,3,1,"93415939664009,19611913284769456132423,19611913284769456132423"
PRP=165545938287D7EBB30C87F04923____,1,2,3310193,-1,99,0,3,1,"2755537060921,555881755075443360919,555881755075443360919"

which of course resulted in:
19611913284769456132423 does not divide M3021587
555881755075443360919 does not divide M3310193

and the 2 assignments were skipped.

Prime95 2021-04-29 03:04

[QUOTE=ATH;577177]Weird error in mprime 30.5b1 (or primenet server error).

Factors for these 2 exponents got reported twice each: [M]3021587[/M] and [M]3310193[/M].[/QUOTE]

I've seen this before -- no clue as to why it happens.
The database has been repaired.

LaurV 2021-04-29 04:33

Got 306.b4, it works fine for P+1 in 13M. Work resumed.

Good job Gorge! :tu:

rainchill 2021-04-29 14:02

I still think the main title bar should show the version and build so it is in screenshots here and anywhere they show up around the web.

LaurV 2021-04-30 05:14

[QUOTE=LaurV;577181]Got 306.b4, it works fine for P+1 in 13M. Work resumed.

Good job Gorge! :tu:[/QUOTE]
Update: b4 still has the issue with hanging when exiting (it was doing P+1, it may be related or not). It needs to be killed from taskman on both Win7 and Win10. When you "stop" it (from the menu, or right-clicking the icon), two or three workers (from 6) still won't stop, but also do nothing ("stopped" or "exited" P95 takes about 4% CPU). "Stopped" P95 is then not responsive (you can not "stop" it, because it is obviously stopped, and you can not "continue", because some tasks are running :razz:).

Hopefully as long as you don't exit it, and don't stop it, it works fine.

kruoli 2021-04-30 14:25

On my Windows 7 laptop, I now had it multiple times that Prime95 "forgot" a lot of assignments. I set [C]MaxExponents=1024[/C] and [C]UnreserveDays=1024[/C] in [C]prime.txt[/C]. Using [C]WellBehavedWork=1[/C] makes no difference. This happened with every 30.6 build so far. I put lots of (but far less than 1,000) assignments of P+1 in the worktodo file (4 workers).

nordi 2021-04-30 17:00

[QUOTE=kruoli;577288]I now had it multiple times that Prime95 "forgot" a lot of assignments.[/QUOTE]
Maybe you should try "UnreserveExponents=999", which does the opposite of what the name implies: It protects 999 exponents per worker from bein unreserved. I set it on my system because mprime kept unreserving assignments.

Prime95 2021-04-30 17:55

[QUOTE=LaurV;577269]Update: b4 still has the issue with hanging when exiting (it was doing P+1, it may be related or not). It needs to be killed from taskman on both Win7 and Win10. When you "stop" it (from the menu, or right-clicking the icon), two or three workers (from 6) still won't stop, .[/QUOTE]

In what P+1 stage were the hung workers? Are the workers mult-threaded?

ATH 2021-04-30 22:04

[QUOTE=nordi;577299]Maybe you should try "UnreserveExponents=999", which does the opposite of what the name implies: It protects 999 exponents per worker from bein unreserved. I set it on my system because mprime kept unreserving assignments.[/QUOTE]

I have UnreserveDays=9999. Not sure if it actually uses that full value, but it does not complain and does not unreserve anything.

ixfd64 2021-04-30 23:26

George, any chance you could update [url=https://mersenneforum.org/showpost.php?p=568252&postcount=2]post #2[/url] with the version 30.6 bug fixes?

LaurV 2021-05-01 08:22

[QUOTE=Prime95;577310]In what P+1 stage were the hung workers? Are the workers mult-threaded?[/QUOTE]
Don't know about the stage, sorry. I [URL="https://www.mersenneforum.org/showpost.php?p=577353&postcount=43"]finished it[/URL], and I won't be back to it too soon, say until they will be offered for automatic assignments. Kinda too much of a headache to make the exponent list for assignments. I also finished my work in 13M (for now). Be back later.

About the threading, yes, there were 6 workers, each using 3 cores, in a 18-cores CPU (i9-10980xe). Usually, second and third hanged. But I am sorry I can not give more details. Right now, doing PRP-CF in same 6 workers, and also doing front P-1 in a single worker (18 cores), not in the same time, of course, haha, everything works fine, as supposed to work. No hanging.

ATH 2021-05-01 12:05

[QUOTE=James Heinrich;576771]Or you could just look at the live report:
[url]https://www.mersenne.org/report_pplus1/?f=1[/url][/QUOTE]

[url]https://www.mersenne.org/report_pplus1/?f=1&all=1[/url]

Only 14 of the 24 current listed factors are real P+1 factors. The rest are old factors found again or combination of old factors found as composite or the same P+1 factors from days ago found for the 2nd time, because people are not using the known factors in the worktodo lines.

Any way to improve the report?

These 10 are not real:

[CODE]
92 821 40000000 5240000000 6 5 6042060144547476324897523777583 2021-04-30 05:09:12.250
296 663 10000000 500000000 6 5 33472736879943425145991507583 2021-04-30 13:15:54.873
19 000 043 1000000 1000000 6 5 3062497001568580991 2021-04-29 06:41:20.290
19 000 087 1000000 55000000 6 5 composite:165612834709033744097242153925591 2021-04-29 21:33:07.540
19 000 099 1000000 1000000 6 5 composite: 6556735832659867930257216861791 2021-04-30 00:17:38.310
19 000 187 1000000 1000000 6 5 composite: 2148714300288186353 2021-04-30 11:50:22.557
19 000 259 1000000 1000000 6 5 composite: 1449071846137904258057 2021-04-29 20:02:07.513
19 000 411 1000000 57000000 6 5 composite: 3759440282766424334351228034173859299863351 2021-04-29 01:12:55.823
19 000 697 1000000 54000000 6 5 457554313955057777 2021-04-30 07:22:56.087
19 000 853 1000000 57000000 6 5 26850678692775826607 2021-04-29 21:36:39.270
[/CODE]


We are at 14 factors from 5038 curves: 9 P+1 smooth and 5 P-1 smooth.

[url]https://mersenneforum.org/showpost.php?p=576652&postcount=199[/url]

nordi 2021-05-01 15:46

One number is actually missing from the [URL="https://www.mersenne.org/report_pplus1/"]P+1 report[/URL]: a factor was found for [URL="https://www.mersenne.org/report_exponent/?exp_lo=92459&full=1"]M92459[/URL] . It was found in [B]stage 1 of P+1[/B], so it probably slipped through the mechanism that creates that report.

Prime95 2021-05-01 17:50

[QUOTE=nordi;577373]One number is actually missing from the [URL="https://www.mersenne.org/report_pplus1/"]P+1 report[/URL]: a factor was found for [URL="https://www.mersenne.org/report_exponent/?exp_lo=92459&full=1"]M92459[/URL] . It was found in [B]stage 1 of P+1[/B], so it probably slipped through the mechanism that creates that report.[/QUOTE]

A bug in the SQL VIEW that tosses rows where a subsequent run used larger bounds

ATH 2021-05-01 18:22

[QUOTE=nordi;577373]One number is actually missing from the [URL="https://www.mersenne.org/report_pplus1/"]P+1 report[/URL]: a factor was found for [URL="https://www.mersenne.org/report_exponent/?exp_lo=92459&full=1"]M92459[/URL] . It was found in [B]stage 1 of P+1[/B], so it probably slipped through the mechanism that creates that report.[/QUOTE]

It has been showing in the factor only report since you found it:

[url]https://www.mersenne.org/report_pplus1/?f=1&all=1[/url]

crasher86 2021-05-01 18:58

I'm kind off clueless.

While pime 30.3.1.0 (beta6) passes all large FFTs

30.5.0.0 and 30.6.1.0(beta6) gets Fatal error on every worker in the first test.

30.6.1.0(beta6):
Test1, 44000 Lucas-Lehmer iterations of M7471106 using FMA3 FFT lenght 384K, Pass1=384, Pass2=1K, clm=1.
Fatal ERROR:Final result was 00000000, expected:3D4F6CBB.
Hardware failure detected running 384K FFT size.


This behaviour is even under factory defaults.
Is there a Problem with ryzen 5800x (2CCD versions, with 1 deacivated CCD)

Prime95 2021-05-01 20:52

[QUOTE=crasher86;577385]I'm kind off clueless.

While pime 30.3.1.0 (beta6) passes all large FFTs

30.5.0.0 and 30.6.1.0(beta6) gets Fatal error on every worker in the first test.[/QUOTE]

A bug was introduced somewhere in the 30.4/30.5 changes. Your problem is fixed in 30.6 build 4.

James Heinrich 2021-05-02 00:00

[QUOTE=ATH;577363]Only 14 of the 24 current listed factors are real P+1 factors.
Any way to improve the report?[/QUOTE]I have updated the report to show prime component factors when a composite was reported by P+1.
And the discovery date of each prime factor is compared to the P+1 run date, and already-known factors are now shown smaller (and with a tooltip indicating the discovery date). If you have UI suggestions for how to present this better it can be modified.

kriesel 2021-05-02 13:02

[QUOTE=James Heinrich;577396]I have updated the report to show prime component factors when a composite was reported by P+1.
And the discovery date of each prime factor is compared to the P+1 run date, and already-known factors are now shown smaller (and with a tooltip indicating the discovery date). If you have UI suggestions for how to present this better it can be modified.[/QUOTE]Hi, thanks for all you do. But please, not the tiny font from the bottom line on the vision test chart. How about italic font or a different font color or both?
Also, perhaps a column for bitsize of factors? Getting a rough comparison by length of the string is obscured by using different font sizes.

ATH 2021-05-02 16:20

[QUOTE=kriesel;577431]Hi, thanks for all you do. But please, not the tiny font from the bottom line on the vision test chart. How about italic font or a different font color or both?
Also, perhaps a column for bitsize of factors? Getting a rough comparison by length of the string is obscured by using different font sizes.[/QUOTE]

Why do you need to be able to read them or see their size? They are old factors, that just happened to be "rediscovered" because people are not using the correct syntax in worktodo.


I understand the report is based on the report of the found factor to the server. I guess it is not possible to search for the "F-PP1" on exponents? Those lines are not added to the old exponents that are "rediscovered".

James Heinrich 2021-05-02 16:42

[QUOTE=kriesel;577431]please, not the tiny font from the bottom line on the vision test chart. How about italic font or a different font color or both?
Also, perhaps a column for bitsize of factors? Getting a rough comparison by length of the string is obscured by using different font sizes.[/QUOTE]I have increased the font size a bit, and set the text color to gray + italic. I have also added a Bits column.

edit: the composite splitting and previously-discovered factor display functionality has been copied to the P-1 report as well now.

James Heinrich 2021-05-02 16:47

[QUOTE=ATH;577454]I understand the report is based on the report of the found factor to the server. I guess it is not possible to search for the "F-PP1" on exponents? Those lines are not added to the old exponents that are "rediscovered".[/QUOTE]I'm not quite sure what you're asking here? The report is generated from the list of all P+1 results sent to the server. The F-PP1 results are always reported as composite factors if applicable, Prime95 makes no effort to split composite factors if it finds one, that's all done server-side.

ATH 2021-05-02 18:53

[QUOTE=James Heinrich;577459]I'm not quite sure what you're asking here? The report is generated from the list of all P+1 results sent to the server. The F-PP1 results are always reported as composite factors if applicable, Prime95 makes no effort to split composite factors if it finds one, that's all done server-side.[/QUOTE]

Yeah, I know you are capturing the results coming in, I was wondering if it could be based on data in the database instead.

On the exponents with "real" P+1 factors: [M]92419[/M]
2021-04-29 nordi F-PP1 Start=6/5, B1=40000000, B2=5240000000, Factor: 26566754812550500792150275766183

That "F-PP1" line is not on the exponents with composite/rediscovered factors.

drkirkby 2021-05-06 12:42

On dual-processor machines, would it be wise to limit the RAM for P-1 to a bit under half the system memory? The reason I ask is that any particular DIMM will be close to one CPU where access is fast, but for a second CPU to access that DIMM, it will be slower.

There might be some advantage in allowing a RAM limit to be set for each worker, just like there are for CPU cores. I have 384 GB RAM, and when the machine is not being used for real work, I could let mprime have 350 GB. But it might not be a good idea for any worker to use more than half that (175 GB).

Dave

kruoli 2021-05-06 12:50

There is already something like this. From undoc.txt:
[CODE]The Memory=n setting in local.txt refers to the total amount of memory the
program can use. [B]You can also put this in the [Worker #n] section to place
a maximum amount of memory that one particular worker can use.[/B][/CODE]

moebius 2021-05-07 19:18

I have .proof files for the following exponents. I coppied them to the p95v30.6b4.win64 dir, but the software won't upload them since 24h. The .proof files are generated with p95v30.3b6win64. What should I do, software bug possible?
[M]105012367[/M]
[M]107986393[/M]
[M]108007709[/M]
[M]108890597[/M]

drkirkby 2021-05-07 19:26

[QUOTE=moebius;577900]I have .proof files for the following exponents. I coppied them to the p95v30.6win64 dir, but the software won't upload them since 24h. The .proof files are generated with p95v30.3b6win64. What should I do, software bug possible?
[M]105012367[/M]
[M]107986393[/M]
[M]108007709[/M]
[M]108890597[/M][/QUOTE]
If your results have not been submitted (i.e. you have no CPU credit), then you can manually submit the results at

[url]https://www.mersenne.org/manual_result/[/url]
I found once that was done, proof files that were previously not uploaded, then got uploaded.

moebius 2021-05-07 19:29

[QUOTE=drkirkby;577901]If your results have not been submitted (i.e. you have no CPU credit), then you can manually submit the results at [/QUOTE]
The results are submited..proof files still can't be uploaded manually at mersenne.org

moebius 2021-05-07 19:48

The .proof file for [M]105012367[/M] was uploaded recently with Advanced> Manual communication. Thx

drkirkby 2021-05-07 19:48

[QUOTE=ATH;577363][URL]https://www.mersenne.org/report_pplus1/?f=1&all=1[/URL]
Only 14 of the 24 current listed factors are real P+1 factors. The rest are old factors found again or combination of old factors found as composite or the same P+1 factors from days ago found for the 2nd time, because people are not using the known factors in the worktodo lines.[/QUOTE]


Is there a document that defines precisely the format of the worktodo.txt file used by mprime/Prime95? On the assignments page
[URL]https://www.mersenne.org/workload/[/URL]
there's a copy of the worktodo.txt file contents, in case ones file gets corrupted. But that is often slightly different to the contents in my worktodo.txt file, and I don't know what the differences are. The website shows ,3,1 on the end, but my worktodo.txt file does not have that.

[CODE]PRP=xxxxx,1,2,108399059,-1,76,0 // in worktodo.txt
PRP=xxxxx,1,2,108399059,-1,76,0,3,1 // on website. [/CODE]Is there a way of requesting a double-check of a P-1 factoring? If so, what does one put in the worktodo.txt file to try to get that assigned? Maybe there's no such work type anyway, but I was just interested if there was.

Uncwilly 2021-05-07 20:24

[QUOTE=drkirkby;577904]Is there a document that defines precisely the format of the worktodo.txt file used by mprime/Prime95?[/QUOTE]Ken has compiled all of that here:
[url]https://www.mersenneforum.org/showpost.php?p=522098&postcount=22[/url]

kriesel 2021-05-07 20:58

[QUOTE=Uncwilly;577909]Ken has compiled all of that here:
[URL]https://www.mersenneforum.org/showpost.php?p=522098&postcount=22[/URL][/QUOTE]:snipe:
Also [url]https://www.mersenneforum.org/showpost.php?p=510736&postcount=2[/url]

As usual, use the [URL="https://mersenneforum.org/showthread.php?p=521922#post521922"]reference info[/URL]!

drkirkby 2021-05-07 21:03

[QUOTE=Uncwilly;577909]Ken has compiled all of that here:
[URL]https://www.mersenneforum.org/showpost.php?p=522098&postcount=22[/URL][/QUOTE]
Thank you

drkirkby 2021-05-07 21:33

[QUOTE=kriesel;577913]:snipe:
As usual, use the [URL="https://mersenneforum.org/showthread.php?p=521922#post521922"]reference info[/URL]![/QUOTE]You might want to consider changing the first sentence of that thread

[I]"Welcome to the GIMPS gpu computing reference blog site map / table of contents."[/I]

It implies it is all about [B]GPUs[/B], but actually it has history, CPUs and many other things. But when I first saw a pointer to that, I did I not bother reading past the first sentence, as I assumed it's about GPUs, and only GPUs


All times are UTC. The time now is 15:17.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.