Prime95 version 25.5

[ATH]

Got this bug doing in the windows version 25.5 doing trial factoring to low limits:

[Wed Sep 26 13:02:02 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 82

When I restarted prime95 it worked again.

[chappjc]

Originally Posted by Prime95

Does anyone have a good reason why I should keep the -B command line argument for Linux?

My rc.d start up scripts use the -B syntax instead of &? But seriously, I guess there is no reason to keep it.

Is there any chance for a FreeBSD 6 compile? Both 32-bit and 64-bit would be very cool to test on our cluster.

[S00113]

Originally Posted by Prime95

That was a nasty bug to find. Please follow the new line to a 64-bit Linux executable.

The new one works!

I can't set the username. It keeps changing back to ANONYMOUS:

Updating computer information on the server
URL: http://v5.mersenne.org/v5server/?v=0.95&px=GIMPS&t=uc&g=be320c6b3f1f6118b05
822dee86976d5&hg=9fe8219314a8555fdf692edf41b6a637&wg=&a=Linux64,Prime95,v25.5,bu
ild+1&c=AMD+Opteron(tm)+Processor+250&f=RDTSC,CMOV,Prefetch,3DNow!,MMX,SSE,SSE2&
L1=64&L2=1024&np=2&hp=1&m=1881&s=2406&h=24&r=1000&u=S00113&cn=minister&ss=41568&
sh=8A32517223DB092459F262CDAAF046A0
RESPONSE:
pnErrorResult=0
pnErrorDetail=SUCCESS
g=be320c6b3f1f6118b05822dee86976d5
u=admin_user_anon
un=ANONYMOUS
cn=minister
od=4

[Joël Harismendy]

Hello,

I am french with version p95v2414 and testing actually exponent 39 864 023

I have pc with xp sp2 home and pentium4 3,2 ghz processor

How can change and use new version p95v255 ?

Thx

[Prime95]

Originally Posted by ATH

Got this bug doing in the windows version 25.5 doing trial factoring to low limits:

PHP Fatal error: Maximum execution time of 60 seconds exceeded in C:\v5\server\v5server\0.95_database.inc.php on line 82

This is only a bug if prime95 does not recover gracefully. One of the reasons for load testing is to make sure the server doesn't generate these types of server-busy errors too often and to see if prime95/mprime recovers properly by resending the result/request at a later time.

[Prime95]

Originally Posted by Joël Harismendy

I am french with version p95v2414 and testing actually exponent 39 864 023

How can change and use new version p95v255 ?

I've edited the first post to show that this is pre-beta software - not for everyday use. Please continue to use version 24.14.

[Prime95]

Originally Posted by S00113

I can't set the username. It keeps changing back to ANONYMOUS:

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.

[jbristow]

Prime95 stopped communicating with the server for a few hours today, no error messages or anything, just no communication. Stopping and continuing didn't fix it, but closing and restarting the program did. When I restarted it, it immediately sent results to the server and got new assignments. I don't see anything interesting in the log file. Thanks.

[James Heinrich]

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.

Posts #18 (& #25 I guess) above shows the log output you get when you try and fill in your username in the PrimeNet dialog (which shows up either blank, or "admin_user_anon"). If I close Prime95 (x64 on Vista64) and edit prime.txt to set the correct value of V5UserID=JamesHeinrich it seems to stick, for a little while at least. In v25.4 when I was doing small-TF it would flop back and forth between "JamesHeinrich", "ANONYMOUS" and "admin_user_anon". Even now in v25.5 it was reporting results as "admin_user_anon" when the PrimeNet dialog username was blank, then as "ANONYMOUS" when I tried to fill in my username and it did the server-communication log as in post #18. For now it's logging results as "JamesHeinrich", but from past experience it may yet revert to one of the anonymous users after a little while.

If I bring up the PrimeNet dialog and don't change anything, no server communication is done, and nothing gets broken. If I change a value and click OK, it communicates with the server and resets the username (in both dialog and prime.txt) to "admin_user_anon". I this log extract I set the username to a non-existant user to force communication, then tried to reset it to my actual username. You'll notice the fake name gets a server error, but my name doesn't, yet both result in setting the user to anonymous:

// trying to set username "INVALID_TEST_USERNAM"
[Sep 26 21:00] Updating computer information on the server
[Sep 26 21:00] URL: http://v5.mersenne.org/v5server/?v=0.95&px=GIMPS&t=uc&g=c4c1132ea5fa74fce4c0e3712cb9ed2d&hg=a558bb4549a8a940bfa2390cf9355e97&wg=ca82aec0c07
[Sep 26 21:00] RESPONSE:
[Sep 26 21:00] pnErrorResult=21
[Sep 26 21:00] pnErrorDetail=Find t_users failed: INVALID_TEST_USERNAM
[Sep 26 21:00] ==END==
[Sep 26 21:00] 
[Sep 26 21:00] PrimeNet error 21: Invalid user
[Sep 26 21:00] Find t_users failed: INVALID_TEST_USERNAM
[Sep 26 21:00] Updating computer information on the server
[Sep 26 21:00] URL: http://v5.mersenne.org/v5server/?v=0.95&px=GIMPS&t=uc&g=c4c1132ea5fa74fce4c0e3712cb9ed2d&hg=a558bb4549a8a940bfa2390cf9355e97&wg=ca82aec0c07
[Sep 26 21:00] RESPONSE:
[Sep 26 21:00] pnErrorResult=0
[Sep 26 21:00] pnErrorDetail=SUCCESS
[Sep 26 21:00] g=c4c1132ea5fa74fce4c0e3712cb9ed2d
[Sep 26 21:00] u=ANONYMOUS
[Sep 26 21:00] un=ANONYMOUS
[Sep 26 21:00] cn=3600x2v255
[Sep 26 21:00] od=12
[Sep 26 21:00] ==END==
[Sep 26 21:00] 
[Sep 26 21:00] Done communicating with server.
// trying to set username "JamesHeinrich"
[Sep 26 21:00] Updating computer information on the server
[Sep 26 21:00] URL: http://v5.mersenne.org/v5server/?v=0.95&px=GIMPS&t=uc&g=c4c1132ea5fa74fce4c0e3712cb9ed2d&hg=a558bb4549a8a940bfa2390cf9355e97&wg=ca82aec0c07
[Sep 26 21:00] RESPONSE:
[Sep 26 21:00] pnErrorResult=0
[Sep 26 21:00] pnErrorDetail=SUCCESS
[Sep 26 21:00] g=c4c1132ea5fa74fce4c0e3712cb9ed2d
[Sep 26 21:00] u=admin_user_anon
[Sep 26 21:00] un=ANONYMOUS
[Sep 26 21:00] cn=3600x2v255
[Sep 26 21:00] od=12
[Sep 26 21:00] ==END==
[Sep 26 21:00] 
[Sep 26 21:00] Done communicating with server.

[James Heinrich]

If communication & worker windows are merged, when the communication thread goes inactive the child window icon goes red and it looks like the worker thread is inactive (but it's still working).

[StarQwest]

Originally Posted by Xyzzy

You are using a much larger FFT size.

How much L2 cache do you have?

L1 Cache Size: 16 KB
L2 Cache Size: 1024 KB

What would be a recommended size to get M332,192,831 down to 0.480 seconds per iteration? If it ran at 0.960 seconds per iteration when operating on one core then it should run twice as fast when running on two cores.