PrimeGrid Calendula Challenge September 24 - October 7

[Ken g6]

Welcome to the most grueling PrimeGrid challenge of the year, the Calendula challenge! Your mission, should you choose to accept it, is to complete as many Prime Sierpinski Problem LLR tests as you can, as quickly and correctly as you can. Because these work units take longer than any other challenge WUs this year, you will have a full thirteen days, from September 24 to October 7, to work on them.

What you need:

• One or more fast x86 processors. (Even slow ones might do some work!)
• new! Having hyperthreading off on i3/5/7s is said to help.
• 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 only "Prime Sierpinski Problem (LLR)" selected in the Projects section.
• Patience! These are long WUs. (Not as long as Seventeen or Bust, but each will probably take at least two days each.)

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.
• Any video cards.

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 the challenge strategy thread for ways to start and finish on time.

Good luck! This message will self-destruct in a few hours! But I'm sure it will come back when the forum downtime is over.

 

[Ken g6]

Less than 48 hours to go until the longest race of the year. Has everyone who's participating tested their machines with some Proth Prime Search LLR WUs? If you don't test, don't come complaining to me later how your four-day WU failed 3/4 of the way through!

On the other hand, I do understand why many may not want to try for this challenge. I'm not sure my slowest computer could even complete one WU in the 13-day time limit!

 

[petrusbroder]

I'll be in - at least with some computers ,,,

 

[Ken g6]

And we're off!

Those who aren't monitoring 10 or 20 computers at a time may be interested in SteveRC's WU juggling suggestion. In theory, with proper planning, the last WU you download could be completed as fast as 1/N'th the time of the others, where N is the number of CPU cores you have.

 

[petrusbroder]

Yupp, we are off. 14 comps started within 15 minutes of each others ... lets see how this goes ...
Please join!

 

[djdave404]

I'm in.

Here for the team!

Go Team AnandTech!

 

[RobertE]

Totally forgot to enable my start/stop tasks this morning. So, got about an 8 hour late start.

 

[Ken g6]

Good to see you both all in! (Edit: Not forgetting you, Peter.) This is an awfully long, tough race, and I wasn't sure how many people would join.

With this race, being several hours off isn't too bad. I'm expecting to have 11 surplus hours at the end of the race, even with WU juggling.

 

[petrusbroder]

I can't do any juggling - takes too much time .... for so many comps. But the race is fun anyhow just for fun. Those who take part will get a lot of credits ...

 

[djdave404]

Wow, a guy from team seti has posted the first score.

Must have used the video card, no cores just to get the first posted score is all i can think of.

My machine is 40% at the moment.

so it's not raw CPU power.

 

[Ken g6]

It can't be the GPU; there's no program for the GPU to run LLR tasks. I suspect he's running an i7 at 4GHz or more. I also wonder if he's running with the new LLR client in an app_info.xml file. It's very fast, but not always stable.

I have one machine that insists on running its WUs at high priority and won't download more to allow me to juggle. If it doesn't do so in time, I may try that new LLR.

 

[biodoc]

Wow, a guy from team seti has posted the first score.

Must have used the video card, no cores just to get the first posted score is all i can think of.

My machine is 40% at the moment.

so it's not raw CPU power.

If it's a highly OC'd i7, they may have turned off hyperthreading in the bios, which may make sense in this race (?).

I won't be joining this race. I'm shuffling some equipment around and have just built a new rig (thuban 1090T) so I'll stick with simap for testing. Plus I'm getting close to the 1 M milestone on simap.

 

[djdave404]

i was compairing it to my i7 980x @ 4.0

hyperthreading is on tho.

 

[Ken g6]

WTF? Wrong thread.

Edit: I meant I posted in the wrong thread. Sorry.

 

[petrusbroder]

We have the first TeAm-stats!

Rank _ Name _______________ Score
1 ____ Sicituradastra _____ 533 909.78
2 ____ BOINCstats __________ 279 816.49
3 ____ SETI.Germany ________ 207 655.90
4 ____ Czech National Team _ 146 287.19
5 ____ TeAm AnandTech _______ 70 776.50
6 ____ Ars Technica _________ 36 704.45

Thanks Kimmyl for the first WUs in!

 

[djdave404]

WTF? Wrong thread.

LOL, Nope, right thread.

I may be missing something i'm thinking, Most of my WU are over 50% is all.

 

[SlangNRox]

I starting crunching when the race started. Should be good for 10-12 total wu during the race. It suddenly heated up in here and my electric company instituted a smart day today and tomorrow (and maybe the rest of the week) where my power rates go to over $1/kwh from 2pm-7pm so that will cut into my crunching

 

[Ken g6]

People are starting to return WUs now, so I suppose I should post some four-day stats:

Rank___Credits____Username
17_____612002_____BlackBox
25_____498038_____kimmyl
93_____146758_____Robert Everly
131____106249_____petrusbroder
200____34669______Ken_g6

Rank__Credits____Team
8_____2732742____Team 2ch
9_____2670402____BOINC@Poland
10____1570920____The Knights Who Say Ni!
11____1397717____TeAm AnandTech
12____1117883____US Navy
13____962345_____PrimeSearchTeam
14____886679_____BOINC.Italy

Dasm, Dave, is that just that one machine?!

Should be interesting what happens when people with more slow machines return results.

 

[RobertE]

Another batch of 4 WUs on the i5 should finish up between 4-6 am eastern 9/29.

This is one tough challenge. About 50-55hrs per WU on this box.

 

[petrusbroder]

The Team's stats:
Rank ______ Team ___________________ Points
09 ________ BOINC@Poland ____________ 2 670 402.70
10 ________ The Knights Who Say Ni! _ 1 570 920.19
11 ________ TeAm AnandTech __________ 1 433 438.85
12 ________ US Navy _________________ 1 117 883.97
13 ________ PrimeSearchTeam ___________ 962 345.75

The TeAm member's stats:
Rank _______ Name ________________ Points
017 ________ BlackBox ____________ 612 002.20
025 ________ kimmyl ______________ 498 038.73
096 ________ Robert Everly _______ 146 758.63
110 ________ petrusbroder ________ 141 970.13
202 ________ Ken_g6 _______________ 34 669.16

I'll probably drop som 16 - 20 WUs within the next 12 hours ... and I don't know how many points that will be.

 

[petrusbroder]

There is one interesting observation: two of my machines sent in WUs with errors. Both machines are at stock, not hot (approx 70ºC), and pass the torture tests perfectly for at least 24 hours ... are these WUs more error prone?

 

[Ken g6]

202 ________ Ken_g6 _______________ 34 669.16

I'll probably drop som 16 - 20 WUs within the next 12 hours ... and I don't know how many points that will be.

I'm guessing each will be about 35,000 or so.

Edit: These aren't more error-prone specifically; though there's some debate about that and Windows 7.

 

[petrusbroder]

The Team's stats:
Rank ______ Team ____________________ Points
09 ________ BOINC@Poland ____________ 2 670 402.70
10 ________ TeAm AnandTech __________ 1 682 428.61
11 ________ The Knights Who Say Ni! _ 1 570 920.19
12 ________ US Navy _________________ 1 117 883.97
13 ________ PrimeSearchTeam ___________ 962 345.75

The TeAm member's stats:
Rank _______ Name __________________Points
017 ________ BlackBox ____________ 612 002.20
025 ________ kimmyl ______________ 498 038.73
034 ________ petrusbroder ________ 390 959.89
098 ________ Robert Everly _______ 146 758.63
204 ________ Ken_g6 _______________ 34 669.16

Just dumped 6 more WUs ... let us continue doing that - OK?

 

[petrusbroder]

I'm guessing each will be about 35,000 or so.

Edit: These aren't more error-prone specifically; though there's some debate about that and Windows 7.

Nice.
But I use Win XP - which has been stable for such a long time ...
I'll keep an eye on that computer ...

 

[djdave404]

3 machines Ken.

The above score is for 17 WU

But i lost 7 WU that round.
Wife closed the door to the computer room, To hot, 6 hour BSOD.....bummer.
when i resumed, 7 WU Errors, toast.

Started a 4th, 4 core. today
that makes 4 total computers now.

i7 980x
i7 920
Q9450
just startedhenom II X4 940

that will make 28 cores/threads total.

All i can do and not have the wife go off.

we are in the top 10!

keep pumping out the WU guys.

Go TeAm anandtech!