PrimeGrid Leonids Challenge Nov 18-23

[Ken g6]

Welcome to the last chance to find primes in a race this year. It's the Leonids Challenge! Your mission, is to complete as many 321 Prime Search LLR tests as you can, as quickly and correctly as you can. You will have five days, from November 18-23, to work on them.

What you need:

• One or more fast x86 processors. (Even slow ones could probably do some work!)
• All of those processors running stably. One good way to check that is to run PPS or SGS LLR for a few days. If those make your processor too hot, try TRP LLR. If you get any errors, your processor probably isn't stable at your current clock/voltage.
• Windows, Linux, or MacOS 10.4+.
• BOINC, attached to PrimeGrid (http://www.primegrid.com/).
• Your PrimeGrid Preferences with "321 Prime Search (LLR)" the only CPU project selected in the Projects section.
• Patience! These are fairly long WUs. (Not as long as Seventeen or Bust, but each could take a day.)

What won't help (but won't hurt either):

• A 64-bit OS. (But if it's Linux, type "sudo apt-get install ia32-libs".)
• A large amount of RAM.
• A GPU. Sorry, still not working. I hit a server misconfiguration, and another tester has crashes.

What would help, but isn't absolutely necessary:

• Reading [post=30943822]Ken's WU Juggling Guide[/post] and applying it once you figure out how long each WU will take. Note that the automatic time estimate isn't very good; estimate total time manually from progress and elapsed time.
• Not running a GPU project. They don't all take much CPU time, but they do all take some.

Work not downloaded and uploaded within the challenge will not be counted. Should you not be able to be in front of one or more computers at that time, see [post=29529001]the challenge strategy thread[/post] for ways to start and finish on time.

Good luck to all!

P.S. Sorry I'm getting later and later with these. :$

 

[petrusbroder]

As usually, I'll be priming ... sorry crunching.

 

[somethingsketchy]

I'll be there, as always.

 

[GLeeM]

I think I've done all the others this year, I'll do this one too

 

[biodoc]

I'll be in.

 

[somethingsketchy]

I'm not sure if my i7-860 is older or what, but after running a few 321 LLR 6.10 WUs, it'll take approximately 20 hours per WU. Considering this post (see below), mentioned 8-12hrs/WU, the extra hours has me concerned.

http://www.primegrid.com/forum_thread.php?id=3744#43303

CPU speed is at stock (2.8GHz). I'm wondering if overclocking to 3.0 would make a difference?

 

[Uppsala9496]

I'm in for the challenge. Been getting some errors in F@H on one of my machines (which I was 100% positive was stable). So this will be a good test for it.

 

[Ken g6]

I'm in for the challenge. Been getting some errors in F@H on one of my machines (which I was 100% positive was stable). So this will be a good test for it.

I'd rather you did some Prime95 to test stability, or ran some short PPS LLR WUs first. If you run an unstable computer in the race, it may appear to give us credits, but those will be removed later.

Edit:

I'm not sure if my i7-860 is older or what, but after running a few 321 LLR 6.10 WUs, it'll take approximately 20 hours per WU.

First, John may have been quoting last year's times; test times increase over months and years. Second, do you have hyperthreading on? Turning it off shouldn't slow things down too much, and might allow you to get an extra set of work in.

 

[somethingsketchy]

I'll give it a try tonight and see if turning off Hyperthreading will work. I have not turned it off before, so I will be curious to see how it will run, or if WUs are a little faster than with HT on.

 

[Uppsala9496]

Running Prime95 now to see how things go. I'll run it until it fails or for 18 hours, whichever comes first.

 

[VirtualLarry]

Got most of my computers switched over to 321 LLR. Yeah, I know it's early, but I won't be up in the morning to switch them over manually, so might as well get them running now.

 

[geecee]

I timed the start of the 321 LLRs incorrectly. I thought I had enough of the shorter PPS LLRs to last until the start, but they apparently went faster than I thought, and I had already switched my selections over to the 321s in my primegrid settings. So the machine I am doing primegrid on is already doing 321s as of sometime earlier this morning. Should I abort these and just wait another hour and a half to pull new work eligible for the race? Or do I just wait until 20 hrs or so in after these have finished? I feel guilty killing the 2 units I just started crunching.

 

[petrusbroder]

I would abort them - they will come back very soon to you or to someone else.

BTW: I am allset for the race ...

 

[brownstone]

Running Prime95 now to see how things go. I'll run it until it fails or for 18 hours, whichever comes first.

Just for the record, I've had OC's that passed 24+ hours of prime95 that failed in F@H. I now use prime95 to test initial stability (30 minutes) and F@H to verify that it is 100% stable. If you are getting errors in F@H you may want to increase voltage or decrease the OC. Just my personal experience.

 

[Uppsala9496]

I figured out that if I shut down via ctrl-c on the machine, I get an error when starting back up. Very bizarre. I went ahead and uninstalled F@H and reinstalled. Took a while to track down the cause.

But back on topic - All 3 of my machines are now crunching PrimeGrid.

 

[petrusbroder]

15 computers ( 90 cores) crunching PrimeGrid 321LLR WUs.

 

[Ken g6]

I got started 1/2 hour late - forgot about daylight savings time.

 

[waffleironhead]

6 cores crunching away. Hopefully I have my cooling under control this race. Losing two units last race was a bit disappointing for me.

 

[biodoc]

18 cores going. Switched off hyperthreading on my 2600K so that should be interesting. Go TeAm!!!

 

[somethingsketchy]

I finally got new tasks nearly an hour late

 

[Uppsala9496]

14 cores. Have F@H gpu clients running on 2 machines (10 cores) where they will pause when done.

 

[ZipSpeed]

12 cores in the race.

 

[sangyup81]

9 cores here with more possibly on the way!

 

[blckgrffn]

When the malaria work clears out, 36 cores on the project.

 

[biodoc]

We're in sixth place and sandwiched between the Cows and Rats again!