PrimeGrid Sophie Germain Challenge April 1-3

[Ken g6]

Welcome back to the PrimeGrid Marie-Sophie Germain Challenge. PrimeGrid is running a short, two-day challenge to find primes, and maybe find a record Sophie Germain prime. No foolin'!*

What you need:

• One or more fast x86 processors. (Even slow ones will do!)
• Windows, Linux, or MacOS 10.4+.
• BOINC, attached to PrimeGrid (http://www.primegrid.com/).
• Your PrimeGrid Preferences with only "Sophie Germain Prime Search (LLR)" selected in the Projects section.

What may help:

• A processor with AVX. Even Bulldozer processors should work with AVX now!
• A 64-bit OS. The benefit of such an OS is unclear at this time, and as of this writing only the Windows 64-bit client was deployed.

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

• A large amount of RAM.
• Any video cards.

This challenge starts April 1 at 18:00 UTC, and runs for only 48 hours. 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, just set their queue level very low (0.01 days). This also makes it more likely that you will be a prime finder rather than a double-checker.

Welcome and good luck to all!

* Well, there sort of is an April Fool's joke this year. If you find a prime, it counts at PrimeGrid, but it won't be big enough to count at The Prime Pages - unless it's a twin or Sophie-Germain prime. Oh, well, you still get BOINC credit.

 

[zzuupp]

There's a new client for the LLR projects. It's definitely worthwhile to d/l some work early. The client is relatively ginormous.

 

[blckgrffn]

Is avx support native now? That would be a huge plus

Actually, it will put me at more of a disadvantage, but I suppose that is how progress works out. Interested to see how avx helps on BD vs SB. Last challenge it was good for about a forty percent performance increase.

 

[petrusbroder]

I am in with all my cores - will take a break from WCG!

 

[zzuupp]

Is avx support native now? That would be a huge plus

Actually, it will put me at more of a disadvantage, but I suppose that is how progress works out. Interested to see how avx helps on BD vs SB. Last challenge it was good for about a forty percent performance increase.

From PG

"This version support AVX (for Intel Sandy Bridge processors)."

 

[Uppsala9496]

Since it is on a Sunday I just need to remind myself to throw my machines at the project. Hopefully I am all in.

 

[blckgrffn]

I am starting to put hosts on this now, hopefully that will help prevent my normal "lateness"

With shallow queues, of course!

 

[biodoc]

These WUs have pretty short crunch times. Hopefully their server can keep up with the demand.

 

[Ken g6]

The race starts in two hours.

Dont forget to run at least one wu on all your host before the challenge start.

You need to get the new app's so we elliminate all app uploads at start.

 

[blckgrffn]

jeeze, thought it just started... man, have to abort all those WU's...

It's on like Donkey Kong!

I have 18 Thuban Cores, the 2500k and i3-2100 in right now, will be adding two more AMD Quads and the Q6700 when I get home today

The AVX enabled 2500k is murdering these things.... I need to beef up my Intel capacity,evidently...

 

[Uppsala9496]

Didn't have the opportunity to upload/run any before the race. Just got all machines up and running now.
Quick question - what is the average run time on these? My 960T is showing under 10min a WU. Is that normal? Actually, all of my machines are with the exception of my 8120. That is showing o9ver an hour run time per WU. What gives?

 

[Uppsala9496]

Nevermind. The FX machine must have been playing an April Fools joke on me. It's running at about 15 minutes a WU. I did scale back the OC a little so the heat doesn't get too high.

 

[biodoc]

I'm all in!! :biggrin:

 

[blckgrffn]

I'm all in!! :biggrin:

That 3930 is going rock the house. My 2500k is doing ~5 Minute WU's!

 

[biodoc]

That 3930 is going rock the house. My 2500k is doing ~5 Minute WU's!

So far I'm not super impressed with it. My 2600K (4.2 GHz) is doing WUs at just under 8 minutes and the 3930K (4.0 GHz) is doing WUs at just over 8 minutes. Of course, there's the 12 threads vs 8 threads factor. The SB-E throws a ton of heat too. Had to jack up the fans on the Corsair H-100 to medium for the race. Significant amount of noise now. :biggrin:

 

[filibusterman]

up in running with most rigs!!

 

[somethingsketchy]

I'm currently getting ~11 minute WUs on an i7-860 Nehalem CPU. Even my Sandy Bridge work laptop is doing ~13 minute WUs. Not great, but not too bad.

 

[petrusbroder]

I am in too - with all my 84 cores ... which may not be so much because many of them are Phenom II, C2Q and i7 of the older kind; none of them are SB .
The FX8150 is not doing too good - 1050 seconds/WU (= 17.5 minutes)

I have to upgrade my crunchers ...

 

[petrusbroder]

The TeAm is still #7. Very nice ... lets see how long that may last.
Crunch on!
Join!
We need all we can get!

 

[biodoc]

Yes, we're still #7 and Peter is catching me quickly!

 

[Uppsala9496]

I am in too - with all my 84 cores ... which may not be so much because many of them are Phenom II, C2Q and i7 of the older kind; none of them are SB .
The FX8150 is not doing too good - 1050 seconds/WU (= 17.5 minutes)

I have to upgrade my crunchers ...

I noticed my FX8120 is my slowest per WU cruncher. My X6's, 960T and 555BE all out perform it on a per WU time basis. Rather disappointing since it is my fastest machine (some 490mhz faster than my X6's).

 

[blckgrffn]

The q6700 couldn't take it with all four cores (slowly overeheated...) so I lost a core

I have to take one of my Thubans down for probably six hours tomorrow night for class

Still, even though 35 cores sounds like an odd number, it is a good number

I am never going to catch the Mad Swede though

I noticed my FX8120 is my slowest per WU cruncher. My X6's, 960T and 555BE all out perform it on a per WU time basis. Rather disappointing since it is my fastest machine (some 490mhz faster than my X6's).

Take your BD FX clock speed, multiply it by 75%, and that is the equivalent Thuban clock speed, so far as I can tell. That's why it needs to be eight cores, otherwise it would never be able to pull its weight...

Heck, I figure the 2500k is doing the work of eight ~3+ Ghz thuban cores !!! With probably 2/3rds the power consumption of six... They either need to straighten our FX AVX on PrimeGrid or I am going to have to build out Intel crunchers, simple as that.

We are currently splitting the difference well in 7th...

Rank Name Score
1 Sicituradastra. 281364.35
2 SETI.Germany 241198.72
3 Czech National Team 136074.61
4 Ukraine 133888.26
5 PrimeSearchTeam 118660.68
6 Team 2ch 111384.24
7 TeAm AnandTech 71637.10
8 Crunching@EVGA 59714.66
9 Aggie The Pew 58040.74
10 BOINC@Poland 52993.35

Too bad we aren't closer to 2ch

This project simply favors Intel. My Q6700 @ 3.0 is completing WUs in ~580 seconds, right in the same range as my non-AVX i3-2100. Compare that to ~700 Seconds on equivalently clocked Deneb...

 

[biodoc]

Jeez, I'm having trouble keeping this beast fed!

There's a limit in tasks in progress so any trouble connecting to the server and I'm toast on new work units.

Nice crunching blckgrffn! You're catching me!!

 

[somethingsketchy]

Yeah, I noticed this morning when I woke up (~5am) that my main desktop had ran out of work. Perhaps bumping it up to 0.05 may work.

 

[blckgrffn]

Rank Name Score
1 Sicituradastra. 635757.91
2 SETI.Germany 515713.67
3 Ukraine 349175.35
4 Czech National Team 302570.24
5 Team 2ch 264906.96
6 PrimeSearchTeam 260440.32
7 TeAm AnandTech 158578.65
8 Crunching@EVGA 138534.26
9 BOINC@Poland 126731.39
10 Aggie The Pew 120368.78

Looks like we are growing the gap below us!

Unfortunately, it looks like #6 is going to be untouchable... 7th would be pretty awesome though

Hoping that the three of us can finish in the top 50, too Four in the top 100? That would be pretty sweet.

BioDoc, how many more cores to do I have on this than you? If we looked at it from a power consumption standpoint, I am certain you are much greener...