[QUOTE=cheesehead;208865]Or, to put it another way, to keep the P-1 workers from holding up LL assignments.[/QUOTE]

From my perspective P-1 actually appeared to be pulling ahead of LL.

Mind you I have no numbers to support this observation.

Yeah I asked George about this and it was indeed done to avoid P-1 holding up LL and TF!

[QUOTE=garo;208898]Yeah I asked George about this and it was indeed done to avoid P-1 holding up LL and TF![/QUOTE]

The only small risk is assigning these exponents to PC's with inadequate RAM allocated to do the P-1 first ... mind you I believe I read some where that the server can check for that before assigning them.

[quote=cheesehead;208865]Or, to put it another way, to keep the P-1 workers from holding up LL assignments.[/quote]That was supposed to be

"to keep the [I]shortage of dedicated [/I]P-1 workers [i](who are more likely to allocate sufficient RAM for decent stage 2)[/i] from holding up LL assignments."

[QUOTE=petrw1;208867]From my perspective P-1 actually appeared to be pulling ahead of LL.

Mind you I have no numbers to support this observation.[/QUOTE]

Other than the stats I presented here:

[url]http://www.mersenneforum.org/showpost.php?p=208541&postcount=64[/url]

which suggest about 10K more P-1 than LL in 2010.
Of which only about 4500 appear to be P1-small.

P-1 abend after factor found in Stage 1 anomaly(?)
 

I've had Prime95 ocassionally abend when a factor is found in P-1; sometimes in Stage 1 and sometimes in Stage 2. Others have mentioned the same problems. The DIFFERENCE this time is that in every previous case immediately upon restarting Prime95 it would send the required result; give me the credit and proceed with the next assignment. THIS TIME after sending in the result (and credit) it proceeded to P-1 factor the SAME exponent. A few hours later during the regularly scheduled "Send new completion dates" process Prime95 noted that this exponent was already P-1'd (by me) and abandoned it to continue with the next one.

Has anyone else had this happen?
Any updates?

[QUOTE=petrw1;218774]I've had Prime95 ocassionally abend when a factor is found in P-1; sometimes in Stage 1 and sometimes in Stage 2. Others have mentioned the same problems. The DIFFERENCE this time is that in every previous case immediately upon restarting Prime95 it would send the required result; give me the credit and proceed with the next assignment. THIS TIME after sending in the result (and credit) it proceeded to P-1 factor the SAME exponent. A few hours later during the regularly scheduled "Send new completion dates" process Prime95 noted that this exponent was already P-1'd (by me) and abandoned it to continue with the next one.

Has anyone else had this happen?
Any updates?[/QUOTE]

Just happened again ... P-1 for M52868369.
For what it is worth I have P-1 factored on several PC's over the years but it is on both my E6550 Duos and ONLY those PC's that I have ever gotten an abend after a P-1 factor or this above situation.

So, to help out, I set a machine to do P-1 on one "core", while doing 100M TF's on the other. I gave it the following memory settings:
Day 40MB
Night 300MB
The machine has 1GB and is running a hyperthreaded Pentium D at 3GHZ, P95 v26.2.
PrimeNet won't give it P-1's. It handed out a DC. The account has an overall preference of TF-LMH.

I manually move all but one of the remain TF's to the other core. Those (plus some that were already assigned to the 2nd core) just disappeared (unreservedays=365, maxexponents=400) and new TF's were handed out.

While I know that this machine is not a P-1 monster, I want to help that area. I don't want to do DC's on the machine.

Well?

[QUOTE=Uncwilly;231316]So, to help out, I set a machine to do P-1 on one "core", while doing 100M TF's on the other. I gave it the following memory settings:
Day 40MB
Night 300MB
The machine has 1GB and is running a hyperthreaded Pentium D at 3GHZ, P95 v26.2.
PrimeNet won't give it P-1's. It handed out a DC. The account has an overall preference of TF-LMH.

I manually move all but one of the remain TF's to the other core. Those (plus some that were already assigned to the 2nd core) just disappeared (unreservedays=365, maxexponents=400) and new TF's were handed out.

While I know that this machine is not a P-1 monster, I want to help that area. I don't want to do DC's on the machine.

Well?[/QUOTE]

From the Thresholds page:
[CODE]Thresholds for P-1 factoring assignments
Required memory for P-1 assignments 300 MB [/CODE]

Maybe you need to set the daytime memory higher.

These sections in undoc.txt might be worth a read:
[CODE]Since P-1 stage 2 runs faster with more memory available you can have the
program only run stage 2 at night when more memory is available. In
prime.txt set:
OnlyRunStage2WithMaxMemory=1[/CODE]

[CODE]LowMemWhileRunning is similar to PauseWhileRunning. This option does not
allow workers use a lot of memory. This example, will make sure the program
is using the minimum amount of memory possible while photoshop is running:
LowMemWhileRunning=Photoshop[/CODE]

And finally on 2 occasions I have had Prime95 drop most of my assignments and then immediately get new ones. George and I discussed but I think we left it as a "once in a lifetime -- let me know if it happens again" event.

[QUOTE=petrw1;231326][CODE]Required memory for P-1 assignments 300 MB [/CODE]

Maybe you need to set the daytime memory higher.[/quote]<Lt. Scott>I cannot do it Capt'n.</scott>

[quote]These sections in undoc.txt might be worth a read:[/QUOTE]Yeah, but it is a combo of programs: Excel, Firefox, IE (pig 1), Acrobat (pig 2), [B]and[/B] 2 others.

Would it work if you set the client to the default assignment (GIMPS - what makes sense) and then use the Primenet site to set the appropriate core to the P-1 assignment?