The problem is not with actual memory being used but with the amount of virtual memory being allocated. My current program instance that was doing TF has been hung for the past 10 hours and virtual memory is pegged at 3059M. Nothing abnormal in the prime.log

yeah I've had the same memory problems as well. I had p95 set to use 1024MB of memory out of 2Gb but I've seen it spike up to using 1.8Gb of the physical as well as using 2Gb of my pagefile. I set p95 back to using 500 last night and it's been running fine so far. I'm running 3 ECM and 1 TF

setting the memory allocation back down to 500mb has done the trick. P95 hasn't stopped working in 3 days now. Still using a hefty pagefile of 1-2Gb but it's not causing the PC to thrash at least :-)

I have a question on the server pages. I set up an account by giving my first PC the name Team_Bundu, and I now want to create a team using the same name. Yet when I do I get an error messge, how do i get around this?

I've installed the 25.4 on my new Q6600, and it started with 4 threads. I only wanted to stress-test for starters, but when it consumed too much RAM (all of the 2Gb) I swiotched to primenet. I left the option 'whatever makes the most sense' active, and the program decided for 4 First-time tests.

Now it is my understanding that this isn't optimal for 'fake' quad cores, and that the best distribution is 1 fist time, 1 trial factoring, another first time and another TF.

1. Why doesn't the program allocate the numbers that way?
2. Should I manually unreserve the exponents and ask for TF for the others? How? I only want to unreserve 2 of the 4 assignments...

Hi sonjohan, I think the simplest answer (based on my limited understanding) is that p95 is looking at each core seperately and since your processors are all fairly high end they have been given first time tests.

If you want to alter the settings to the "optimal" setting you outlined then yes, just set the threads and then unreserve the 2 LL's you don't want.

Of course since you're using 25.4 you might want to help out testing the new server? If you do just mail George (or PM Prime95 on this forum) for the addtional line to add to prime.txt. Your stats get zeroed once the test phase is over but you'll get the sweet satisfaction of knowing you helped?

Hope my answers helped

PS - I think George is on holiday at the moment but should be returning soon

Bundu, I asked George the same question and he said that teams have not been enabled yet and all load-testing accounts will most likely be deleted so you can set up teams etc. later.

thanks Garo, it's weird though because there are 2 teams set up on the server, Primenet and Team Prime Rib?

No Team_Prime_Rib is not a team, just an ordinary account like Team_Bundu. The only teams I see on the server are Primenet and GW's Legions which were both set up by George/Scott.

Thank, I think I'll do that, as I want the darn thing to be roughed up and checked out.
Now I have 3 PCs working:
1 at work (dual core: LMH & double check)
Old AMD 2400+: firrst time check
Q6600 :D

[QUOTE=Bundu;113437]Hi sonjohan, I think the simplest answer (based on my limited understanding) is that p95 is looking at each core seperately and since your processors are all fairly high end they have been given first time tests.

If you want to alter the settings to the "optimal" setting you outlined then yes, just set the threads and then unreserve the 2 LL's you don't want.

Of course since you're using 25.4 you might want to help out testing the new server? If you do just mail George (or PM Prime95 on this forum) for the addtional line to add to prime.txt. Your stats get zeroed once the test phase is over but you'll get the sweet satisfaction of knowing you helped?

Hope my answers helped

PS - I think George is on holiday at the moment but should be returning soon[/QUOTE]

Small bug: Prime95 was stopped, then I edited a setting in the PrimeNet dialog, and on clicking OK it communicated with the server (presumably to update the setting) but then started the worker thread (I don't think it should have done so).
Running 64-bit version on Vista64, 1 worker thread doing TF to low limits.

On a different note, I wonder if the communication could be optimized somewhat, perhaps by getting batches of work (eg 10 or 100 exponents at a time) -- I noticed it couldn't communicate fast enough to keep one core of an AM2 X2 @ 2.6GHz busy doing TF-to-low-limits, let alone two cores. Now granted, I should probably be doing something harder than TF to 56-bit with a somewhat-powerful CPU, but still...