mersenneforum.org Prime95 version 25.5
 Register FAQ Search Today's Posts Mark Forums Read

 2007-10-02, 11:34 #67 Andi47     Oct 2004 Austria 46628 Posts What am I doing wrong? I was running a torture test using 25.5 After the torture test I added the "secret line" to prime.txt and restarted prime95, but it started stresstesting again. So I set Stresstester=0 and UsePrimenet=1 in prime95.txt, and then I got this: Code: [Tue Oct 02 13:26:34 2007 - ver 25.5] Updating computer information on the server URL: http://v5.mersenne.org/v5server/....... [rest of the link obscured for I don't know HOW MUCH info about the PC, the universe and the rest of all is included] RESPONSE: pnErrorResult=9 pnErrorDetail=Untrusted program versions currently excluded by PrimeNet ==END== PrimeNet error 9: Access denied Untrusted program versions currently excluded by PrimeNet
 2007-10-02, 13:03 #68 Bundu     Jul 2004 Mid Calder, Scotland 5·37 Posts I found the best method was to join gimps without the secret line (so you connect to the old server and prime then gives you all the correct settings which don't add in when stress testing). Then quit which unreserved all the exponents but left all the settings correct. Then add the secret line; open prime; add your user details (or just leave them anonymous if you prefer) and force a manual communication if the comms thread hasn't started already? I think I got an error message at first but it clears pretty soon Last fiddled with by Bundu on 2007-10-02 at 13:06
 2007-10-02, 13:37 #69 Bundu     Jul 2004 Mid Calder, Scotland 5·37 Posts one of my instances froze today without any error messages of any kind. It froze at 00:41. There was no windows error message and the other instances are running perfectly. Tried forcing a manual communcation which didn't clear it. So I exited and restarted and it's running fine now
2007-10-02, 14:49   #70
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

22·1,873 Posts

Quote:
 Originally Posted by James Heinrich I ran ECM (2 threads) for not much more than an hour and got 5 SUMOUT errors and then finally it crashed outright.
There was a bug in the multi-threaded FFTs. I've uploaded a fix for Windows 32-bit and 64-bit. The linux and Mac OS X versions have not been updated.

2007-10-02, 23:24   #71
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

22×1,873 Posts

Quote:
 Originally Posted by Bundu one of my instances froze today without any error messages of any kind. It froze at 00:41. There was no windows error message and the other instances are running perfectly.
Try changing to "debug=2" in the primenet section of prime.txt. We might learn a little more.

Note, there were similar problems in the sockets implementation of 24.14. I'd call the networking layer and it would never return. I hoped the cURL library would not have the same problem. But there may well be a bug deep within Windows that I cannot workaround.

At least with the comm stuff in its own thread, an LL test will continue even though comm with the server is hung. I could code a timer to wake up when the comm thread has not returned after 15 minutes - but I do not know of a way to safely terminate the comm thread.

2007-10-03, 08:20   #72
Andi47

Oct 2004
Austria

2·17·73 Posts

Quote:
 Originally Posted by Bundu ... Then quit which unreserved all the exponents but left all the settings correct.
When I unreserve an exponent (even with testing stopped), it immediately grabs a new exponent from the server.

Do the settings stay correct (for now adding the secret line) when I click "advanced --> quit GIMPS?"

P.S.: I downloaded the file again because George said that there was a bugfix, and then created a prime95.txt which contained the secret line, before starting Prime95. When I started Prime95 it gave me the dialogue window to join GIMPS and then contacted the V4 server to get exponents for P-1'ing (and LLing afterwards).

2007-10-03, 10:05   #73
James Heinrich

"James Heinrich"
May 2004
ex-Northern Ontario

2×412 Posts

Quote:
 Originally Posted by Prime95 There was a bug in the multi-threaded FFTs. I've uploaded a fix for Windows 32-bit and 64-bit. The linux and Mac OS X versions have not been updated.
Seems to be happy -- it's been running for 9 hours so far without problems.

 2007-10-03, 11:04 #74 garo     Aug 2002 Termonfeckin, IE 22·691 Posts There is a bug in the server comm logic somewhere. I have been assigned several TF exponents (3 of which are sitting in the Team_Prime_Ribaccount right now) which never make it to my worktodo.ini Looking at the prime.log it seems that this happens with the very first exponent the server gives me right after the "Maximum Execution time exceeded" error occurs. As an example, I have been assigned 123030191 since yesterday. My computer never saw the exponent and never communicated the expected return date to the server either. Here is what happened: Code: Getting assignment from server PrimeNet success code with additional info: Server assigned trial factoring work Got assignment 9ED6CD3F3C8689C0EF5CD13E62503AF7: Trial factor M123030179 Sending expected completion date for M123030179: Oct 2 2007 Getting assignment from server [Tue Oct 2 22:03:22 2007 - ver 25.5] PnErrorResult value missing. Full response was: PHP Fatal error: Maximum execution time of 60 seconds exceeded in C:\v5\server\v5server\0.95_database.inc.php on line 42 Sending result to server: UID: Team_Prime_Rib/garo_v26, M123017413 no factor to 2^57, Wd2: 9CC732A3, AID: 94A535F1F545C717675267E4E2402711 Getting assignment from server PrimeNet success code with additional info: Server assigned trial factoring work Got assignment 4FC36766E6E6986D2A764EA4C9EBC982: Trial factor M123030893 Sending expected completion date for M123030893: Oct 2 2007 And this is the line in my individual report: garo_v260123030191 TF 0.0% 1 2007-10-02 14:022007-10-02 14:02 Last fiddled with by garo on 2007-10-03 at 11:05
2007-10-03, 17:36   #75
Bundu

Jul 2004
Mid Calder, Scotland

5·37 Posts

Quote:
 Originally Posted by Andi47 When I unreserve an exponent (even with testing stopped), it immediately grabs a new exponent from the server. Do the settings stay correct (for now adding the secret line) when I click "advanced --> quit GIMPS?" P.S.: I downloaded the file again because George said that there was a bugfix, and then created a prime95.txt which contained the secret line, before starting Prime95. When I started Prime95 it gave me the dialogue window to join GIMPS and then contacted the V4 server to get exponents for P-1'ing (and LLing afterwards).
My fault I posted in a hurry:
• Start 25.5 and join gimps as normal and it should connect to the old server and yeah it will grab an exponent
• click "advanced --> quit GIMPS
• It will then ask if you want to finish your exponent... say no
• exit the program
• add the secret line to the prime text
• start it again and set your worker threads to Trial Factoring to low limits or ECM on small mersenne numbers (whatever your preference from what I can tell TF tests the server and ECM has had more bugs so both seem to help)
• have fun and post bugs here :-)!

 2007-10-04, 00:16 #76 Old man PrimeNet     Jan 2003 Altitude>12,500 MSL 101 Posts If there's a 5.0 server bug DO NOT post it here and expect a fix. Email me at primenet@mersenne.org, instead. I probably won't fix anything unless I get an email.
2007-10-04, 11:11   #77
James Heinrich

"James Heinrich"
May 2004
ex-Northern Ontario

D2216 Posts

Quote:
 Originally Posted by Prime95 Can someone send me a prime.log file showing this happening? I'm trying to reproduce on 64-bit WinXP so far without any luck.
My results are still being reported as ANONYMOUS:
Code:
Member Name          Computer Name    Exponent  Type     UTC Time Received    Days GHz-days  Result
-------------------- ---------------- --------- -------- ------------------- ----- --------  -------------------------------------------------
ANONYMOUS            3600x2v255       1086989   NF-ECM   Oct  4 2007  9:57AM   6.0   0.0282  completed 1 ECM curves, B1=50000, B2=5000000
ANONYMOUS            3600x2v255       1086791   NF-ECM   Oct  4 2007  8:56AM   6.0   0.0282  completed 1 ECM curves, B1=50000, B2=5000000
ANONYMOUS            3600x2v255       388519    NF-ECM   Oct  4 2007  8:40AM   6.1   0.0092  completed 1 ECM curves, B1=50000, B2=5000000
ANONYMOUS            3600x2v255       1086763   NF-ECM   Oct  4 2007  8:30AM   6.1   0.0282  completed 1 ECM curves, B1=50000, B2=5000000
ANONYMOUS            3600x2v255       1086461   NF-ECM   Oct  4 2007  8:15AM   6.2   0.0282  completed 1 ECM curves, B1=50000, B2=5000000
The log snippets are still quoted above (e.g. post #18). I've tried both 32- and 64-bit versions of Prime95 on Vista64 and they both give the same broken result. Please let me know if I can provide further logs or testing.

Last fiddled with by James Heinrich on 2007-10-04 at 11:12

 Similar Threads Thread Thread Starter Forum Replies Last Post Prime95 Software 441 2020-02-16 15:18 Prime95 Software 148 2012-03-18 19:24 Prime95 Software 76 2010-12-11 00:11 Prime95 PrimeNet 143 2007-09-24 21:01 pacionet Software 74 2006-12-07 20:30

All times are UTC. The time now is 13:15.

Sun May 16 13:15:48 UTC 2021 up 38 days, 7:56, 0 users, load averages: 1.40, 1.74, 2.04