## Other

drummers-lowrise

Message boards : Project Staging Area : 2014 PRPNet May Challenge

 Subscribe SortOldest firstNewest firstHighest rated posts first
Author Message
Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 75842 - Posted: 23 Apr 2014 | 22:36:50 UTC

ESP - "North Star Challenge" - 01-16 May

Announced is a sixteen day ESP challenge from 01st until 16th of May. I suggest we start at 18:00 UTC and end at 18:00 UTC.

Suppose that both the Sierpiński problem and the prime Sierpiński problem had finally been solved, showing that 78557 is the smallest Sierpiński number and that 271129 is the smallest prime Sierpiński number. This would not preclude the existence of a composite Sierpiński number k such that 78557 < k < 271129. So we might state an extended Sierpiński problem asking if 271129 is the second Sierpiński number, prime or not.

There remain 12 composite k's for which no primes have been found. They are as follows:
91549, 99739, 131179, 161041, 163187, 193997, 200749, 202705, 209611, 227723, 229673, 238411

More about the extended Sierpinski Problem search can be found here. News and infos about the PRPNet client can be found here.

To take part, you have to activate the following lines in prpclient.ini:

server=ESP:100:1:prpnet.primegrid.com:13005

Stats will be available 15 minutes after the start at the well known place here

All previous PRPNet challenge stats can be found here

Good luck!
____________

rebirther

Joined: 10 Aug 05
Posts: 783
ID: 85
Credit: 175,774,608
RAC: 0

Message 75888 - Posted: 25 Apr 2014 | 4:02:15 UTC

If the port is empty can we add it to BOINC platform?

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 75889 - Posted: 25 Apr 2014 | 5:07:04 UTC - in response to Message 75888.

If the port is empty can we add it to BOINC platform?

I already asked this question. There is talk of moving ESP to BOINC, but no solid plan. Once the ESP port on PRPNet is emptied it will be reloaded.

KEP

Joined: 10 Aug 05
Posts: 299
ID: 110
Credit: 10,360,408
RAC: 3,730

Message 76000 - Posted: 29 Apr 2014 | 9:32:59 UTC

Well, I'm also in. Only with 4 cores. Overall I expect to do 100-120 results over the entire challenge. Of course if we nearly exhaust the testrange entirely, I may only produce as little as 40 results on my I5.

Well let's see, how many primes we find and how many participants that chooses to join the challenge. Currently ~190400 results remain untested. It would be really cool if this conjecture could reach n=10M before the end of the challenge, however that requires some of the Dutch Power Cows (maybe all of them) and Aggie The Pew joins the challenge and starts a little friendly race to the sweet #1 spot :)

Happy crunching everyone. Take care!

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13804
ID: 53948
Credit: 345,369,032
RAC: 1,967

Message 76003 - Posted: 29 Apr 2014 | 11:13:42 UTC

I'm not sure where the idea that using up all the candidates on a PRPNet port means we're moving something to BOINC. Just like on BOINC, when we run low on work, we put more work in. The amount of work in a queue is absolutely meaningless to everyone except the person whose responsibility it is to keep the queue filled with work.
____________
My lucky number is 75898524288+1

JimB
Honorary cruncher

Joined: 4 Aug 11
Posts: 918
ID: 107307
Credit: 977,945,376
RAC: 19

Message 76022 - Posted: 29 Apr 2014 | 23:58:08 UTC

We have sieving on ESP up to n=10M. The PRPNet port was previously loaded to n=5M, now it's got up to n=5.5M. We're not going to run out of work any time soon. I'll keep an eye on it during the challenge and load more new work if needed.

Gary Craig
Volunteer tester

Joined: 30 Dec 09
Posts: 3213
ID: 52890
Credit: 1,005,618,748
RAC: 0

Message 76029 - Posted: 30 Apr 2014 | 7:47:05 UTC - in response to Message 76000.

...
It would be really cool if this conjecture could reach n=10M before the end of the challenge, however that requires some of the Dutch Power Cows (maybe all of them) and Aggie The Pew joins the challenge and starts a little friendly race to the sweet #1 spot :)
...

I'm quite sure that the Vermin Horde will show up. Hope to see some Cows too, as well as others. Hopefully we can knock off some K's, but it will be a tall order to fill.
--Gary

____________
"I am he as you are he as you are me and we are all together"

87*2^3496188+1 is prime! (1052460 digits)
4 is not prime! (1 digit)

KEP

Joined: 10 Aug 05
Posts: 299
ID: 110
Credit: 10,360,408
RAC: 3,730

Message 76031 - Posted: 30 Apr 2014 | 11:24:22 UTC - in response to Message 76029.
Last modified: 30 Apr 2014 | 11:26:05 UTC

Hopefully we can knock off some K's, but it will be a tall order to fill.
--Gary

Well here is something to think about :)

at n=1.25M 20k's remained unprimed
at n=2.50M 17k's remained unprimed (15 % removed)
at n=5.00M 12k's remained unprimed (at least 29.4 % removed)

If we go to n=5.0M without further primes, then the average removal rate for n=1.25M to n=5.0M is 22.2% of the k's remaining, removed per doubleing of n. This gives these statistical chances for the amount of primes found for n=5M to n=10M:

15.0% removal rate for n>5M to n<=10M = 1.800 primes
22.2% removal rate for n>5M to n<=10M = 2.664 primes
29.4% removal rate for n>5M to n<=10M = 3.528 primes

So as can be seen, statistically we can expect anything from 1.8 to 3.528 primes from n>5M up to n<=10M.

The odds is as follows (with 190,348 untested candidates remaining):

1.800 primes gives a 1 in 105,748 chance of a prime for any given test completed
2.664 primes gives a 1 in 71,451 chance of a prime for any given test completed
3.528 primes gives a 1 in 53,953 chance of a prime for any given test completed

As you can see, yes it is a tall order to fill, however the challenge is long and if participation will be great, it is not unrealistical that we will see a prime or 2... maybe even 3 if we are lucky and productive enough ;)

Take care

KEP

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13804
ID: 53948
Credit: 345,369,032
RAC: 1,967

Message 76068 - Posted: 1 May 2014 | 16:27:54 UTC - in response to Message 75889.

If the port is empty can we add it to BOINC platform?

I already asked this question. There is talk of moving ESP to BOINC, but no solid plan. Once the ESP port on PRPNet is emptied it will be reloaded.

Only if we screw it up.

The goal is always to load more work BEFORE it runs out!
____________
My lucky number is 75898524288+1

Brook Harste

Joined: 30 Nov 08
Posts: 1348
ID: 32401
Credit: 805,483,198
RAC: 0

Message 76088 - Posted: 2 May 2014 | 3:44:47 UTC

I don't think the challenge points page is updating...or I'm to impatient.
____________

rebirther

Joined: 10 Aug 05
Posts: 783
ID: 85
Credit: 175,774,608
RAC: 0

Message 76089 - Posted: 2 May 2014 | 3:56:50 UTC - in response to Message 76088.

I don't think the challenge points page is updating...or I'm to impatient.

yes, its stucking.

Volunteer moderator
Volunteer tester
Project scientist

Joined: 5 Feb 08
Posts: 1212
ID: 18646
Credit: 816,613,648
RAC: 185,781

Message 76090 - Posted: 2 May 2014 | 4:56:25 UTC
Last modified: 2 May 2014 | 5:07:27 UTC

there was a bug/typo in my crontab; next update should be at 07:00 MESZ (05:00 UTC); I will keep an eye on it ...

EDIT:
... seems to me it is working again
____________
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76143 - Posted: 3 May 2014 | 23:31:46 UTC

I put together a list of Conjectures being worked on for Base 2/4:

Riesel Conjectures (Source)
Base 2 - 52x Odd k's being worked on by PrimeGrid's The Riesel Problem (TRP)
Base 2 - 2x Even k's being worked on by CRUS
Base 2 even-n - 3x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 2 odd-n - 5x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 4 - 1x k not covered by Base 2 Conjectures being worked on by CRUS

Sierpinski Conjectures (Source)
Base 2 - 6x k's being worked on by PrimeGrid's Seventeen or Bust Project (SoB)
Base 2 2nd Conjecture - k<78557 as above
- Prime k's>78557 being worked on by PrimeGrid's Prime Sierpinski Problem (PSP)
- Composite k's>78557 being worked on by PrimeGrid's Extended Sierpinski Problem (ESP)
- 2x even k's being worked on by CRUS
Base 2 even-n - 2x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 2 odd-n - 4x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 4 - All k's covered by Base 2 Conjectures

Let me know if I've missed any.

composite
Volunteer tester

Joined: 16 Feb 10
Posts: 1022
ID: 55391
Credit: 888,938,802
RAC: 133,359

Message 76167 - Posted: 4 May 2014 | 18:32:27 UTC - in response to Message 76143.

ALL YOUR BASE ARE BELONG TO US.

MarkHFX

Joined: 12 Feb 12
Posts: 109
ID: 130425
Credit: 6,869,640
RAC: 0

Message 76173 - Posted: 4 May 2014 | 23:00:10 UTC

Set up and 6 cores dedicated

____________
PrimeSearch Team

[DPC]Division_Brabant~TFH|Fony

Joined: 2 Jul 11
Posts: 45
ID: 104145
Credit: 64,217,029
RAC: 0

Message 76211 - Posted: 6 May 2014 | 9:57:54 UTC
Last modified: 6 May 2014 | 9:59:11 UTC

LOL.. Haven't done any PRPnet stuff for ages...
But yesterday afternoon I've set it up again to crunch a few days for this challenge. See if my i5-3450 @ 3,8 Ghz can do some usefull work here. Only 20 - 24 tests a day though, but better then nothing :)
This way I can also finally get rid of that last silver badge ;)
____________

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76274 - Posted: 8 May 2014 | 16:39:29 UTC

I set up to use 4 cores on my Windows 7 desktop for this challenge. No problems seen so far, although I've found nothing telling me how much work was completed. Today, I also set up 2 cores on my Windows Vista desktop. The first two workunits crashed within seconds. Is there a place where I can look at why the crashes occurred, so that I might fix it before any more workunits?

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76283 - Posted: 9 May 2014 | 6:33:43 UTC - in response to Message 76274.

I set up to use 4 cores on my Windows 7 desktop for this challenge. No problems seen so far, although I've found nothing telling me how much work was completed. Today, I also set up 2 cores on my Windows Vista desktop. The first two workunits crashed within seconds. Is there a place where I can look at why the crashes occurred, so that I might fix it before any more workunits?

Is the Vista PC CPU/Memory overclocked? Hard drive full?
Maybe try one core on its own first and then both together.
Have you run BOINC on this PC or PRPNet previously?
There is no FAQ page that I know of. Plenty of advice if you start reading other postings on the Message Boards.

My PCs been crashing lately. I think all CPU Cores together is too much for the level of cooling I have. Gone back to 5 cores + GPU (84% load) rather than 6 no GPU (100% load).

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76287 - Posted: 9 May 2014 | 13:33:41 UTC - in response to Message 76283.

I set up to use 4 cores on my Windows 7 desktop for this challenge. No problems seen so far, although I've found nothing telling me how much work was completed. Today, I also set up 2 cores on my Windows Vista desktop. The first two workunits crashed within seconds. Is there a place where I can look at why the crashes occurred, so that I might fix it before any more workunits?

Is the Vista PC CPU/Memory overclocked? Hard drive full?
Maybe try one core on its own first and then both together.
Have you run BOINC on this PC or PRPNet previously?
There is no FAQ page that I know of. Plenty of advice if you start reading other postings on the Message Boards.

My PCs been crashing lately. I think all CPU Cores together is too much for the level of cooling I have. Gone back to 5 cores + GPU (84% load) rather than 6 no GPU (100% load).

Not overclocked. Hard drive not full. BOINC run previously, but not PRPNet.

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76316 - Posted: 11 May 2014 | 6:43:27 UTC - in response to Message 76287.

I set up to use 4 cores on my Windows 7 desktop for this challenge. No problems seen so far, although I've found nothing telling me how much work was completed. Today, I also set up 2 cores on my Windows Vista desktop. The first two workunits crashed within seconds. Is there a place where I can look at why the crashes occurred, so that I might fix it before any more workunits?

Is the Vista PC CPU/Memory overclocked? Hard drive full?
Maybe try one core on its own first and then both together.
Have you run BOINC on this PC or PRPNet previously?
There is no FAQ page that I know of. Plenty of advice if you start reading other postings on the Message Boards.

Not overclocked. Hard drive not full. BOINC run previously, but not PRPNet.

Check what's written at the end of the *.log files in the prpclient-* directories?

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76317 - Posted: 11 May 2014 | 7:58:57 UTC

With 5 days to go ardo and PrimeSearchTeam are solidly in the lead.

Trailing Edge 4,813,480

At the start of the challenge:
Trailing Edge 4,649,251

10.97% chance we would have on average found a prime during the challenge so far (Trailing start to Trailing edge now).

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76349 - Posted: 12 May 2014 | 1:58:20 UTC - in response to Message 76316.

I set up to use 4 cores on my Windows 7 desktop for this challenge. No problems seen so far, although I've found nothing telling me how much work was completed. Today, I also set up 2 cores on my Windows Vista desktop. The first two workunits crashed within seconds. Is there a place where I can look at why the crashes occurred, so that I might fix it before any more workunits?

Is the Vista PC CPU/Memory overclocked? Hard drive full?
Maybe try one core on its own first and then both together.
Have you run BOINC on this PC or PRPNet previously?
There is no FAQ page that I know of. Plenty of advice if you start reading other postings on the Message Boards.

Not overclocked. Hard drive not full. BOINC run previously, but not PRPNet.

Yes.

Check what's written at the end of the *.log files in the prpclient-* directories?

No log files present. Only the executables and the DLL files. No input or output files found.

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76355 - Posted: 12 May 2014 | 4:50:56 UTC - in response to Message 76349.

Yes.

Check what's written at the end of the *.log files in the prpclient-* directories?

No log files present. Only the executables and the DLL files. No input or output files found.

For 2 core setup you will need to first fill in the master_prpclient.ini with your username, email address etc. Then for 2 core setup double click the 2-dual-install-prpclient.bat, then double click the 2-dual-update-prpclient-ini.bat (even for the first time), then finally double click the 2-dual-start-prpclient.bat.

Doc No

Joined: 2 Jan 09
Posts: 135
ID: 33689
Credit: 735,525,610
RAC: 5,642

Message 76359 - Posted: 12 May 2014 | 6:29:01 UTC - in response to Message 76274.

I've found nothing telling me how much work was completed.

May not be exactly what you are looking for but I stumbled across this in another thread - you can search for your name at the bottom of the page: http://u-g-f.de/PRPNet/

Volunteer moderator
Volunteer tester
Project scientist

Joined: 5 Feb 08
Posts: 1212
ID: 18646
Credit: 816,613,648
RAC: 185,781

Message 76374 - Posted: 12 May 2014 | 19:10:41 UTC - in response to Message 76359.
Last modified: 12 May 2014 | 19:11:52 UTC

I've found nothing telling me how much work was completed.

May not be exactly what you are looking for but I stumbled across this in another thread - you can search for your name at the bottom of the page: http://u-g-f.de/PRPNet/

direkt link of user robertmiles
challenge stats rank #22 as of now
____________
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76376 - Posted: 12 May 2014 | 20:20:50 UTC - in response to Message 76355.

Yes.

Check what's written at the end of the *.log files in the prpclient-* directories?

No log files present. Only the executables and the DLL files. No input or output files found.

For 2 core setup you will need to first fill in the master_prpclient.ini with your username, email address etc. Then for 2 core setup double click the 2-dual-install-prpclient.bat, then double click the 2-dual-update-prpclient-ini.bat (even for the first time), then finally double click the 2-dual-start-prpclient.bat.

Already done. When I started 2-dual-start-prpclient.bat, two windows showed up on the bottom toolbar, them whatever was running in those windows crashed within a few seconds.

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76378 - Posted: 12 May 2014 | 20:37:47 UTC - in response to Message 76376.

Yes.

Check what's written at the end of the *.log files in the prpclient-* directories?

No log files present. Only the executables and the DLL files. No input or output files found.

For 2 core setup you will need to first fill in the master_prpclient.ini with your username, email address etc. Then for 2 core setup double click the 2-dual-install-prpclient.bat, then double click the 2-dual-update-prpclient-ini.bat (even for the first time), then finally double click the 2-dual-start-prpclient.bat.

Already done. When I started 2-dual-start-prpclient.bat, two windows showed up on the bottom toolbar, them whatever was running in those windows crashed within a few seconds.

Tried again. This time, I saw an error message in each of those windows, but
only displayed for less than a minute:

Configuration file 'prpclient.ini' could not be opened for reading.

Currently, none of my PRPNet directories contain a file named 'prpclient.ini'.
Is 2-dual-start-prpclient.bat supposed to create this file, or do I have to create it manually? If manually, where and what should go in it?

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76379 - Posted: 12 May 2014 | 20:41:43 UTC - in response to Message 76374.

I've found nothing telling me how much work was completed.

May not be exactly what you are looking for but I stumbled across this in another thread - you can search for your name at the bottom of the page: http://u-g-f.de/PRPNet/

direkt link of user robertmiles
challenge stats rank #22 as of now

Thanks, that's it.

Toshio Yamaguchi

Joined: 19 May 11
Posts: 135
ID: 99209
Credit: 693,512
RAC: 0

Message 76380 - Posted: 12 May 2014 | 21:32:23 UTC - in response to Message 76378.

Currently, none of my PRPNet directories contain a file named 'prpclient.ini'.
Is 2-dual-start-prpclient.bat supposed to create this file, or do I have to create it manually?

This should be created automatically either when running the install-prpclient script or the update-prpclient script, I am not sure which of them does it. start-prpclient only initializes the PRPNet client.

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76382 - Posted: 12 May 2014 | 23:39:28 UTC - in response to Message 76380.

Currently, none of my PRPNet directories contain a file named 'prpclient.ini'.
Is 2-dual-start-prpclient.bat supposed to create this file, or do I have to create it manually?

This should be created automatically either when running the install-prpclient script or the update-prpclient script, I am not sure which of them does it. start-prpclient only initializes the PRPNet client.

Finally started. Appears to have been some problem if the batch files are started by some means other than double-clicking.

Thank you.

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76426 - Posted: 16 May 2014 | 5:12:48 UTC

With 14 hours to go:

Trailing Edge 4,938,158

At the start of the challenge:
Trailing Edge 4,649,251

Would be good to get the Trailing edge over 5M before it heads to BOINC.

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 76442 - Posted: 17 May 2014 | 3:29:36 UTC

Stats are final. Congratulation to ardo and the PrimeSearchTeam!
Top of the challenge rankings is as follows:

rank userid tests score 1 ardo 2,564 53,319,498 2 sm5ymt 1,920 40,659,795 3 288larsson 1,756 37,506,625 rank teamid tests score 1 PrimeSearchTeam 6,287 132,946,106 2 Sicituradastra. 1,761 37,614,666 3 Aggie_The_Pew 1,048 22,499,538

We had 61 users and 32 teams compete with 24 first time users on the ESP port join during the challenge!

Currently:
Trailing Edge 4,970,306

At the start of the challenge:
Trailing Edge 4,649,251

So trailing edge has increased 6.6% during the challenge, leading edge 8.1%.

Clicking on the primelist I can see ESP's been going on PRPNet since ~n=100,000 so its come along way.

See you at the next challenge.

Gary Craig
Volunteer tester

Joined: 30 Dec 09
Posts: 3213
ID: 52890
Credit: 1,005,618,748
RAC: 0

Message 76447 - Posted: 17 May 2014 | 9:00:10 UTC

For future challenge standings pages, how about including a "Team" column even in the individual standings table? Sometimes I'm curious to see who's crunching for what team when looking at that page.

That column would be blank for crunchers with no team, of course. Not sure what to do if someone changes team in the middle of a challenge; probably just show the current one, or something like "(several)". I wouldn't advocate multiple rows for a single UID since the individual standings table should show a single overall total per UID.

I've no idea if this would be easy or hard, so it's just a thought. Cheers and thanks for another successful challenge, though apparently no new primes were found :-(

--Gary

Volunteer moderator
Volunteer tester
Project scientist

Joined: 5 Feb 08
Posts: 1212
ID: 18646
Credit: 816,613,648
RAC: 185,781

Message 76456 - Posted: 17 May 2014 | 12:46:53 UTC - in response to Message 76447.

I've no idea if this would be easy or hard, so it's just a thought.

it is not hard, as of now it is impossible because stats are builded up by loading individual or team stats from the server and there is no relation between user and team.
____________
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76464 - Posted: 18 May 2014 | 3:42:14 UTC

Could you modify the master_prpclient.ini file to make it say what to do in order to finish and return the workunits already downloaded, but don't download any more? My first two tries at doing that both appear to be wrong.

JeppeSN

Joined: 5 Apr 14
Posts: 1727
ID: 306875
Credit: 41,456,369
RAC: 13,221

Message 76465 - Posted: 18 May 2014 | 6:07:49 UTC - in response to Message 76464.

Could you modify the master_prpclient.ini file to make it say what to do in order to finish and return the workunits already downloaded, but don't download any more? My first two tries at doing that both appear to be wrong.

I would also like to know. I think that "setting" could be useful before upgrading to a new version of the PRPNet client, and before abandoning PRPNet on a particular computer. /JeppeSN

Volunteer moderator
Volunteer tester
Project scientist

Joined: 5 Feb 08
Posts: 1212
ID: 18646
Credit: 816,613,648
RAC: 185,781

Message 76466 - Posted: 18 May 2014 | 6:15:51 UTC - in response to Message 76465.

Could you modify the master_prpclient.ini file to make it say what to do in order to finish and return the workunits already downloaded, but don't download any more? My first two tries at doing that both appear to be wrong.

I would also like to know. I think that "setting" could be useful before upgrading to a new version of the PRPNet client, and before abandoning PRPNet on a particular computer. /JeppeSN

There is a stop as soon as possible option in the prpclient.ini.
// stopasapoption= tells the client that it needs to be shutdown automatically, i.e. without // a CTRL-C. It is evaluated after each test is completed. It should be 0 upon startup. // The accepted values are: // 0 - Continue processing work units // 2 - Return completed work units and abandon the rest // 3 - Return completed work units (keep the rest) // 6 - Complete all work units and return them // 7 - Complete all work units then shut down (do not report completed work) stopasapoption=0

You can modify this while the client is running. So your feature request is implemented, if I got you right.
____________
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76471 - Posted: 18 May 2014 | 12:36:45 UTC

Doesn't work for stopaspoption under Windows.

I'm now trying it for startoption and stopoption as well.

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13804
ID: 53948
Credit: 345,369,032
RAC: 1,967

Message 76472 - Posted: 18 May 2014 | 12:52:41 UTC - in response to Message 76471.

Doesn't work for stopaspoption under Windows.

I'm now trying it for startoption and stopoption as well.

It's stopasapoption. You're missing the second "a" if that's the way it appears in your prpclient.ini.
____________
My lucky number is 75898524288+1

robertmiles

Joined: 26 Dec 09
Posts: 121
ID: 52682
Credit: 101,555,708
RAC: 2,311

Message 76473 - Posted: 18 May 2014 | 14:49:53 UTC - in response to Message 76472.
Last modified: 18 May 2014 | 14:50:59 UTC

Change stopasapoption in which file? Change it in master_prpclient.ini, the prpclient.ini file in the same directory, or the prpclient.ini file in each of the instance directories?

Changing it only in master_prpclient.ini doesn't work.

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13804
ID: 53948
Credit: 345,369,032
RAC: 1,967

Message 76475 - Posted: 18 May 2014 | 15:55:23 UTC - in response to Message 76473.

Change stopasapoption in which file? Change it in master_prpclient.ini, the prpclient.ini file in the same directory, or the prpclient.ini file in each of the instance directories?

Changing it only in master_prpclient.ini doesn't work.

The prpclient.ini file in the prpnet-# directories are what controls each individual instance of prpclient. You therefore have two choices:

1) Change the master_prpclient.ini file, then run the appropriate update script to propagate that to the individual prpclient.ini files in the sub directories. NOTE: If you've manually changed the contents of the individual prpclient.ini files, this will overwrite your prior changes.

2) Change the individual prpclient.ini file in one or more sub directories, as appropriate. NOTE: If at any time in the future you run one of the update scripts, you will overwrite your changes.

It's the individual prpclient.ini files that are important, but master_prpclient.ini, combined with the update scripts, are there to make it easy for you to update multiple instances at once.

What I do depends on whether I want to affect just one PRPNet instance, or all of them. If it's just one, I update the individual prpclient.ini file. If I want to affect all the instances, I update master_prpclient.ini and run the update script.
____________
My lucky number is 75898524288+1

composite
Volunteer tester

Joined: 16 Feb 10
Posts: 1022
ID: 55391
Credit: 888,938,802
RAC: 133,359

Message 76477 - Posted: 19 May 2014 | 1:41:41 UTC - in response to Message 76475.

How do you get prpclient.ini to switch projects without causing workunits to be "abandoned" when resuming the original work after a few days?

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13804
ID: 53948
Credit: 345,369,032
RAC: 1,967

Message 76479 - Posted: 19 May 2014 | 3:06:55 UTC - in response to Message 76477.

How do you get prpclient.ini to switch projects without causing workunits to be "abandoned" when resuming the original work after a few days?

Let the old tasks run down to zero, then switch to another project.
____________
My lucky number is 75898524288+1

Toshio Yamaguchi

Joined: 19 May 11
Posts: 135
ID: 99209
Credit: 693,512
RAC: 0

Message 76486 - Posted: 19 May 2014 | 9:38:07 UTC - in response to Message 76477.

How do you get prpclient.ini to switch projects without causing workunits to be "abandoned" when resuming the original work after a few days?

Do you mean you want to receive work from, lets call it port 1 for simplicity, then pause crunching work from port 1, receive work from port 2, crunch that work to end and then resume the unfinished work from port 1?

If this is correct, then I don't think that is possible with the settings provided by PRPNet. What you could try is to manually move the checkpoint file for port 1 work out of the folder, switch to port 2 in prpclient.ini, crunch that work and return it and then comment out port 2 and uncomment port 1 again. Then move the checkpoint file back to the folder where it was and start prpclient again.

Please note that the above procedure comes without any warranty of correct functionality, I never tried that myself.

composite
Volunteer tester

Joined: 16 Feb 10
Posts: 1022
ID: 55391
Credit: 888,938,802
RAC: 133,359

Message 76491 - Posted: 19 May 2014 | 10:41:02 UTC - in response to Message 76486.
Last modified: 19 May 2014 | 10:42:40 UTC

Do you mean you want to receive work from, lets call it port 1 for simplicity, then pause crunching work from port 1, receive work from port 2, crunch that work to end and then resume the unfinished work from port 1?
yes
What you could try is to manually move the checkpoint file for port 1 work out of the folder, switch to port 2 in prpclient.ini, crunch that work and return it and then comment out port 2 and uncomment port 1 again. Then move the checkpoint file back to the folder where it was and start prpclient again.
I tried that already and it works, but it's inconvenient on a 24-core hotbox (it can be done by writing scripts, which some people might also consider "inconvenient"). What I would like to see is the following procedure:
stop-prpclient.sh
change workload in master_prpclient.ini
update-prpclient-ini.sh to distribute
start-prpclient.sh, which immediately picks up the new workload (a challenge is a suitable use case)
crunch for a few days on the new workload (but don't let the original workunits expire!)
reset master_prpclient.ini back to original workload
update-prpclient-ini.sh to finish the current workunit and resume the original one

This might work or it might not. What I did wrong in the last 2 steps was uncomment both workloads and set stopasapoption=6, which I thought would complete workunits from both ports and then stop, but it abandoned the original work after it finished the current one. I seems like a reasonable expectation for both workunits to complete since the comments for stopasapoption say

// 6 - Complete all work units and return them

Lumiukko
Volunteer tester

Joined: 7 Jul 08
Posts: 165
ID: 25183
Credit: 785,072,250
RAC: 40,488

Message 76515 - Posted: 20 May 2014 | 10:09:58 UTC - in response to Message 76491.

I tried that already and it works, but it's inconvenient on a 24-core hotbox (it can be done by writing scripts, which some people might also consider "inconvenient"). What I would like to see is the following procedure:
stop-prpclient.sh
change workload in master_prpclient.ini
update-prpclient-ini.sh to distribute
start-prpclient.sh, which immediately picks up the new workload (a challenge is a suitable use case)
crunch for a few days on the new workload (but don't let the original workunits expire!)
reset master_prpclient.ini back to original workload
update-prpclient-ini.sh to finish the current workunit and resume the original one

This might work or it might not. What I did wrong in the last 2 steps was uncomment both workloads and set stopasapoption=6, which I thought would complete workunits from both ports and then stop, but it abandoned the original work after it finished the current one. I seems like a reasonable expectation for both workunits to complete since the comments for stopasapoption say

// 6 - Complete all work units and return them

I would (if there is enough disk space) just create another copy of prpclient directories. For a 24-core I would then have the original prpclient-1 ... prpclient-24, and a new copy with prpclient-25 ... prpclient-48. Then I just shut the original "1-24" clients, and start other project by running "25-48" clients. After finished working with those I can just restart the "1-24" clients, and they would resume from checkpoint and continue normally. You can even have separate directory sets for each project (prpclient-mega, prpclient-esp, prpclient-prs, ...).
Note: do not comment out ports that have unfinished work, just set that port percentage to 0%. After all work has been reported you can comment the port out.

--
Lumiukko

composite
Volunteer tester

Joined: 16 Feb 10
Posts: 1022
ID: 55391
Credit: 888,938,802
RAC: 133,359

Message 76518 - Posted: 20 May 2014 | 12:56:18 UTC - in response to Message 76515.

create another copy of prpclient directories
A very reasonable suggestion. Thanks.

Note: do not comment out ports that have unfinished work, just set that port percentage to 0%. After all work has been reported you can comment the port out.
This note would be a worthwhile addition to the inline documentation in master_prpclient.ini

Roger
Volunteer developer
Volunteer tester

Joined: 27 Nov 11
Posts: 1138
ID: 120786
Credit: 268,621,444
RAC: 0

Message 77064 - Posted: 9 Jun 2014 | 12:35:43 UTC - in response to Message 76143.

I put together a list of Conjectures being worked on for Base 2/4:

Riesel Conjectures (Source)
Base 2 - 52x Odd k's being worked on by PrimeGrid's The Riesel Problem (TRP)
Base 2 - 2x Even k's being worked on by CRUS
Base 2 even-n - 3x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 2 odd-n - 5x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 4 - 1x k not covered by Base 2 Conjectures being worked on by CRUS

Sierpinski Conjectures (Source)
Base 2 - 6x k's being worked on by PrimeGrid's Seventeen or Bust Project (SoB)
Base 2 2nd Conjecture - k<78557 as above
- Prime k's>78557 being worked on by PrimeGrid's Prime Sierpinski Problem (PSP)
- Composite k's>78557 being worked on by PrimeGrid's Extended Sierpinski Problem (ESP)
- 2x even k's being worked on by CRUS
Base 2 even-n - 2x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 2 odd-n - 4x k's being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
Base 4 - All k's covered by Base 2 Conjectures

Let me know if I've missed any.

60849*2^3067914+1 is prime! Congrats to Jean Penné
http://primes.utm.edu/primes/page.php?id=117992

Base 2 even-n - now 1x k left, being worked on by CRUS, Liskovets-Gallot Conjecture (3 divides k)
If anyone wants to give us a hand sieving...