The SETI.Germany vs. TeAm AnandTech race in Correlizer. The goal has been reached!!!

[filibusterman]

I don't know if this has been discussed in the thread but I am not able to get wu on some rigs. Is there a problem with having two different projects run at the same time such that correlizer doesn't send wu to clients that have a queue?

 

[filibusterman]

what is the proportional resource allocation for Correlizer compared to the other BOINC CPU projects on this particular host? if its relatively small in comparison to the other projects, it won't download new work as often as the other projects do. also, what is your cache set to (this is what Uppsala9496 is talking about)?


Uppsala9496: go to tools -> computing preferences -> network usage tab -> additional work buffer (set value manually)...of course this is assuming you're running BOINC v6.12.xx, but even if you're running a slightly older version, it'll be found in a similar place.

ahh here it is!!

 

[Uppsala9496]

I find that Phenom II CPUs don't have accurate temperature readings with CoreTemp or any temperature monitoring software. From what I've read and personal experience, I would add about 10C to what your reading in CoreTemp.

Unfortunately, the stock heatsinks for the most part are junk if you're serious about overclocking. I'm not 100% sure but I think the X6 HSF are bit beefier than the X2 or X4 HSF. If so, cooling a 6 core CPU with a HSF designed for lesser CPUs might be an issue. I would keep an eye out for a sale for a Cooler Master Hyper 212+ or Corsair A50 for $20. What kind of voltages are you applying? May need a slight increase to the CPU and chipset. As a reference, at 3.5 GHz, my X6 needs 1.425V for the Vcore and 1.25 for the chipset.

@ geecee
The X6 stock heatsink is a lot better than the the standard for the X2 or X4.
I was using the stock for the X6 on a X2 BE with 4 cores unlocked. At a simple overclock (3.2 to 3.4) the temps from X-Sensors were coming in at around 50*C with the fan at full speed. A Hyper 212+ dropped the temps to the mid 40's however the fan is only at about 1,200 rpm's.
Amazon has the Hyper 212+ for around the price as Newegg (~$25). If you have Amazon Prime, might be worth getting it there for the 2 day shipping.

 

[Uppsala9496]

geecee - newegg has the following for sale:
http://www.newegg.com/Product/Produc...-35233029-L07C
$8.96 after $20 mail in rebate.
Coupon code: EMCJJKB39

 

[geecee]

geecee - newegg has the following for sale:
http://www.newegg.com/Product/Produc...-35233029-L07C
$8.96 after $20 mail in rebate.
Coupon code: EMCJJKB39

Nice, thanks.

 

[geecee]

I find that Phenom II CPUs don't have accurate temperature readings with CoreTemp or any temperature monitoring software. From what I've read and personal experience, I would add about 10C to what your reading in CoreTemp.

Unfortunately, the stock heatsinks for the most part are junk if you're serious about overclocking. I'm not 100% sure but I think the X6 HSF are bit beefier than the X2 or X4 HSF. If so, cooling a 6 core CPU with a HSF designed for lesser CPUs might be an issue. I would keep an eye out for a sale for a Cooler Master Hyper 212+ or Corsair A50 for $20. What kind of voltages are you applying? May need a slight increase to the CPU and chipset. As a reference, at 3.5 GHz, my X6 needs 1.425V for the Vcore and 1.25 for the chipset.

Thanks for the advice. Good to know about the core temps. I'm not really an overclocker, but I thought it might be nice to get a little bit more out of this box with the various holiday DC races coming up. If I can pick up an aftermarket cooler cheap, I will do so (Uppsala has already pointed one out). I haven't messed with the voltage at all yet, so other than the change to FSB, everything is pretty much default. Maybe I'll tweak more after I get a cooler.

 

[wayliff]

I've joined the SETI.USA forum to see if I can recruit someone from there ... at least for the race. The worst they can say is no.

I am on hold for some time until I can post...hopefully soon.

Edit: I sent a PM to "Cruncher Peter" over @ SETI.USA. We'll see what happens.
Edit2: My pm generated a thread in SETI.USA. (http://www.setiusa.us/showthread.php?2416-Correlizer-Challange)

I just noticed we got our first guest from SETI.USA. Please welcome our new race guest Shralper.

Thanks Shralper for your help!

 

[Sunny129]

Welcome Shralper! thank you for helping out our team with the Correlizer race!

i'll also extend a warm welcome to any and all SETI.USA members who join the cause

 

[ZipSpeed]

Welcome aboard Shralper!

 

[petrusbroder]

SG has noticed our recruitment drive to SETI.USA.
Some oppose it, others dont care.
An argument is that the race would not be team vs team race and that the results do not show the power of each team ...
Just to keep you informed. We don't want to look bad ... although I don't know if we look bad.
Let us see how this turns out.

 

[wayliff]

SG has noticed our recruitment drive to SETI.USA.
Some oppose it, others dont care.
An argument is that the race would not be team vs team race and that the results do not show the power of each team ...
Just to keep you informed. We don't want to look bad ... although I don't know if we look bad.
Let us see how this turns out.

I understand the sentiments SG team members may have about TA asking for help at SETI.USA.

I personally do not see this type of recrutining any different than recruiting somewhere else. Recruiting is recruiting, the only difference here being I went to a place where we know people crunch a lot.

But if this would forfeit the race, I guess I can go to SETI.USA and ask for the reversal of my request. I'd feel no shame in doing so.

Maybe next we can make no cross-team recruiting rules...if that is not seen as a valid way of recruiting.

 

[wayliff]

A heads up for all Correlizer racers!

With mod permission, I started a recruitment thread at the CPU forum.
http://forums.anandtech.com/showthread.php?t=2204371

Please refrain from starting recruitment threads in other forums.

I have already made arrangements so we post where we are allowed to post while staying in compliance with forum rules.

You will see one more post at the Video Cards forum made by one of our team members at some point today.

 

[ZipSpeed]

Awesome work wayliff! Hopefully your work won't be in vain.

 

[wayliff]

Welcome ViRGE to the Correlizer race!

Thanks ViRGE! all help is appreciated.

 

[[2W]SEARCHER]

:biggrin: Good Evening TeAm AnandTech,
:biggrin: Hello Petrusbroder,

I don`t care, your Team look for Help by Team SETI.USA. I have Contact with HIM and I must only Phoned HIM and the 2nd Wave become Help from only 1 Specialuser, with a Swiss Knife. WE WIN GUYS.

http://www.youtube.com/watch?v=3hfzlQKKqVU

:whiste::whiste::whiste:

:biggrin: Greatings [2W]SEARCHER

 

[wayliff]

:biggrin: Good Evening TeAm AnandTech,
:biggrin: Hello Petrusbroder,

I don`t care, your Team look for Help by Team SETI.USA. I have Contact with HIM and I must only Phoned HIM and the 2nd Wave become Help from only 1 Specialuser, with a Swiss Knife. WE WIN GUYS.

http://www.youtube.com/watch?v=3hfzlQKKqVU

:whiste::whiste::whiste:

:biggrin: Greatings [2W]SEARCHER

Hey Searcher!

<Google Translated>
Das ist nett! Bringen Sie es!
Wir haben noch mehr Spaß Drücken beider unsere Grenzen!
</Google Translated>

:biggrin:
See ya...

 

[wayliff]

removed...

 

[VirtualLarry]

I don't know if this has been discussed in the thread but I am not able to get wu on some rigs. Is there a problem with having two different projects run at the same time such that correlizer doesn't send wu to clients that have a queue?

I noticed that on my laptop, that I was switching over from PrimeGrid to Correlizer. I clicked "No new work" on PrimeGrid, attached Correlizer, and it's not pulling many if any WUs. There were three PrimeGrid WUs when I last checked, and a single Correlizer.

When I checked Properties for the Project, it showed that WU fetch priority was something like -600. (Negative 600). Could this have something to do with it?

Edit: Nevermind, once it finished up the PrimeGrid units, it downloaded some more Correlizer units.

 

[wayliff]

Well it started good and perhaps because of me ... it went south.

Here are some of the replies at SETI.Germany after they found out we, I should "I and not WE", were recruiting at SETI.USA. Not everyone took it well.

One reply:
"hmm, bin noch am &#252;berlegen aber vermutlich ist f&#252;r mich das Race dann hier zu Ende. Bin erstmal auf NNW.
SG hat gewonnen und TeAm AnandTech hat sich disqualifiziert."

Another reply:
"Also f&#252;r mich hat sich TAAT auch selbst aus dem Rennen genommen...

Das hat dann mit Race nix mehr zutun. Aber egal, freut mich das wir die 1 so ohne Probleme halten k&#246;nnen und der Gegner schon zu Tiefschl&#228;gen ausholen muss um uns &#252;berhaupt die Bretter zeigen zu k&#246;nnen

"

The first reply basically declares SG victorious and TA self-disqualified. The race ends end for this cruncher at this point.

The second reply also considers the race finished and also mentions how TA (the adversary) uses a low blow by recruiting across boards in order to be able to show off. (Please do not take this literally as I may not be understanding it well, perhaps Petrus can clarify if I misrepresented the person's reply).

Before anything I just want to say: I did not mean to cause the impression of cheating nor for this race to go south. I took it upon myself to do this recruitment at SETI.USA and also at the CPU and the Video Card forums.
I was not aware we should not be recruiting to increase output and particularly from another board but I suppose at that point it is no longer TA ground...mea culpa.

My apologies to TA and SG.

I wanted to help and I guess it backfired. I take full blame...
Next time I will ask first...
:$

 

[sangyup81]

What I'm wondering is whether the SETI.USA member was participating in correlizer beforehand or not. If not, then we're adding to correlizer by getting more work done for them and frankly deserve the points

 

[ZipSpeed]

I don't think we can paint anything victorious or disqualified seeing as this race is unofficial friendly competition. Doesn't affect Correlizer one bit as the only thing victorious in the end is the progression of the science.

Nothing against SG but some people take distributed computing way too seriously. If they want to declare victory, go ahead. Our emotions shouldn't get in the way of something that in the end, isn't that big of a deal.

If SG thinks the race is over, perhaps they can stop crunching. But I have a feeling if we take the lead, they would think twice about their "victory".

 

[petrusbroder]

OK. Let us get this straight:
Look at this post in this thread:
It says among other things:

3. Have fun.
4. Recruit your best.
5. Have fun.
So: Asslaminate, borgh, borrow, get permissions for borghing and asslaminating, persuade your friends, make peace with your adverseries (to get access to their computers), install new software (and get paid in credits), etc, etc,

This has been seen by wayliff as a encouragement to recruit. Nowhere in this whole thread has anybody forbidden the recruitment of crunchers, as long as they signed up for TeAm AnandTech.
Those from SETI.USA who joined the TeAm have crunched for the TeAm.
We have not stated anywhere, quite the opposite, that recruitng was forbidden.
Nowhere in the SG-thread has been posted that recruiting is forbidden.
In Team AnandTech recruiting is encouraged because it helps the scientific project tpo get their WUs crunched, and that is important for us in TeAm AnandTech.
Please check post #6: here I wrote

Remember the ways:
to borgh, to asslaminate, to borrow, to persuade, to buy, to incorporate, to add ...
and then there are some more shady ways, but those are for more mature minds ...

Nobody in SG has had any objections to that post. The TeAm has been open with their culture, nowhere have we hidden what we are doing. Now, at this time point SG complains about recruiting members from other teams.
IMHO they should have read the whole thread, especially the beginning, and stated their own preferences or discussed their culture. We would have accepted thjat.

wayliff, please do not feel bad abou this. This is a culter clash - between two teams with different ways of doing DC.
I am sorry that this has put a a shadow on the race. Since SG has pulled out, nobody wins, the project looses. That is sad.

 

[ZipSpeed]

Those from SETI.USA who joined the TeAm have crunched for the TeAm.

This quote really drives it home. The credit goes toward TA and to have earned that credit, that individual would have had to join our team.

 

[wayliff]

well race or no race...what are we waiting for to try for #1?! :biggrin:

Several users at the CPU forum appear to be ready to jump in!

 

[petrusbroder]

I have posted this in the SG-forum.
I have also recieved a PM from one SG member who sendss his greetings to wayliff and who does not think bad about wayliff.
It seems to me that the race will go on, so don't let up.
Let us race.