Prime95 version 25.5

[Prime95]

Originally Posted by Mini-Geek

I just got some "Server sent bad exponent" errors when trying to request DCs on the v5 server.

That's why you're not supposed to check that work preference!

Client rejects all DCs below 10,000,000 as a sanity check. Server assigned a tiny exponent because I'm testing LL and DC on small exponents (for obvious reasons).

[Mini-Geek]

Originally Posted by Prime95

That's why you're not supposed to check that work preference!

Client rejects all DCs below 10,000,000 as a sanity check. Server assigned a tiny exponent because I'm testing LL and DC on small exponents (for obvious reasons).

Oh, ok. I thought it might be something like that, but I wasn't sure if it was just a real problem that you hadn't noticed because you were specifically testing with other work.

[Bundu]

Sent an email to Scott today about the server pages, thought I would drop a copy here to check if any of you folks are getting the same problem:

Hi Scott,

I don’t know if this is something that is still a work in progress?  When I 
log into my account page on the v5 server I am asked If want 
to download the following files?

piegraph.php
bargraph.php

I take it this is an error?

My Account is Team_Bundu

Cheers

[Mini-Geek]

I noticed the comm thread gives a much more detailed report of what's going on. When this (eventually) comes out, I for one, won't have communication often, but when it does I don't want it to take up the whole screen, so I'd like to request an option to only show the limited communications (like it did in v24), so that I can combine it with the worker and main threads without covering up all their info when communication happens.

Also, on the new PrimeNet, all of the pages take a very long time to load.
e.g. here's some of the time log for a Results Detail page with 62 results (which, by the way, I think there should be an option to limit that, at least per page...multiple pages, I mean):

odbc_query_rows_array() 17035.73 ms 69.05%
include() 7138.78 ms 28.93%
...
TOTAL = 24672.41 ms 100%

EDIT: And one more thing, on the new PrimeNet, when you enter your member login, you have to press tab twice to get to the password field. I'm not sure if there's a way to define that the password field is the next to tab to in HTML, but you could just move the [ forgot? ] link to below the GO button or somewhere.

[Prime95]

Originally Posted by Mini-Geek

I noticed the comm thread gives a much more detailed report of what's going on. When this (eventually) comes out, I for one, won't have communication often, but when it does I don't want it to take up the whole screen, so I'd like to request an option to only show the limited communications (like it did in v24).

Debug=0 in the Primenet section of prime.txt will do what you want. That will be the default value in a production release of v25.

[Mini-Geek]

Originally Posted by Prime95

Debug=0 in the Primenet section of prime.txt will do what you want. That will be the default value in a production release of v25.

Thanks. Is this the right place to report problems like what I posted with PrimeNet, or should I e-mail it somewhere, or are we not focusing on PrimeNet v5 being good at this point, or what?

[Prime95]

Originally Posted by Mini-Geek

Is this the right place to report problems like what I posted with PrimeNet, or should I e-mail it somewhere?

Web page problems / suggestions should be emailed to Scott at primenet at mersenne dot org. I've already forwarded your suggestions.

[Andi47]

Feature request: It would be nice if I could set max. memory useage seperately for each thread:

I am currently running ecm on small mersenne numbers in thread 1 and ecm on small fermat numbers in thread 2. I had set daytime available memory to 128 MB (nighttime 512 MB) and thread 2 gets idle every time when thread 2 starts stage 2 of an ECM curve. I now increased daytime available memory to 192 MB, and now thread 1 grabbed 181 MB for stage 2, and thread 2 still has to become idle when thread 1 runs ECM stage 2.

[Mini-Geek]

Originally Posted by Prime95

Web page problems / suggestions should be emailed to Scott at primenet at mersenne dot org. I've already forwarded your suggestions.

Thanks, I'll send them to the right place from now on.
Another problem with Prime95, but that I'm not quite sure is fixable: Many times, especially after many new lines of output have been...um...outputted, when I click Open Window, it freezes for a bit (a few seconds) as it loads the window.
Another: I'm not sure if it's because it's 25.5, because it has so many new numbers coming in and out of it (as I'm on TF to low limits), or because I'm now running two copies (25.5 and 24.14 to catch 25.5's missed CPU time), but it seems to impact my computer's performance a noticeable bit (in that it seems to be like a program using much memory...programs take longer to start, etc.). (It's using 2-3 MB of memory, so it's not that it's actually using a lot of memory, just impacting the computer like it is.)
Speaking of running both to catch the other's missed CPU time, I've noticed that as it is getting a steady stream of numbers, the computer is finishing them faster than the computer is requesting them, so there's about a second of CPU time that 24.14 is getting, but would otherwise be wasted, per exponent completed.

[Prime95]

Originally Posted by Andi47

I am currently running ecm on small mersenne numbers in thread 1 and ecm on small fermat numbers in thread 2. I had set daytime available memory to 128 MB (nighttime 512 MB) and thread 2 gets idle every time when thread 2 starts stage 2 of an ECM curve. I now increased daytime available memory to 192 MB, and now thread 1 grabbed 181 MB for stage 2, and thread 2 still has to become idle when thread 1 runs ECM stage 2.

When thread 1 goes into high memory usage, thread 2 should temporarily move on to the next line in worktodo.txt. Question: are there multiple lines of work in thread 2? If so, can you email me the prime.txt, local.txt, worktodo.txt files next time you see this behavior? Thanks.

[Andi47]

Originally Posted by Prime95

When thread 1 goes into high memory usage, thread 2 should temporarily move on to the next line in worktodo.txt. Question: are there multiple lines of work in thread 2? If so, can you email me the prime.txt, local.txt, worktodo.txt files next time you see this behavior? Thanks.

I am currently not sitting on this computer (I will be back there on wednesday), so I can't email the files today.

I have switched thread 2 to ECM on fermat numbers, and it has got a couple of curves on F21 in the worktodo list. Prime95 tried to move to other lines in worktodo.txt, but it did not find a line with less memory useage, so it went idle.

I have not checked, but I guess that it possibly has done stage 1 for all of these lines, so it possible has only stage 2 left for all curves.