Drastic BigAdv changes

link

So from the 16th of Jan 2012 no one with a machine with less than 16 threads will be give ANY bigadv work ..... LOL!

 

WOW, so you need to have a Xeon, or better still 2, to get the higher ppd? That means no desktop system (other than, of course, a few dual-processor systems) will get bigadv, as I don't recall ever seeing a standard desktop PC with 16 cores.

 

They do this after everyone buys 2600k's

 

So there will be zero desktop rigs that will be able to run BIGADV units - only the SR2 at the moment? Someone's pulling some strings at Stanford...

Well it's back to SMP units for the foreseeable. I might as well have bought 2500K's instead of the 2600K's.

 

hehe, I just sold my 2600K and ordered a 2700K. It's only to keep me happy until SR-3 and SB-EP are released anyway I guess it'll motivate me to get the GTX570/580 for this rig sooner instead of later, it might make up for the bigadv loss

 

They're good overclockers. I've just upgraded my HTPC with a 2700K and with the same board as your sig, hit 5Ghz stable but settled on 4.8 as it was a HTPC I set the fans to whisper quiet.

 

What Vcore did you have to set for 5GHz stable?

Temps won't be a problem for me as I have a chiller, and once the cooling loop is build and I am settled with a decent overclock, the chiller is going in the kitchen, so it will be ultra quiet

 

I'm happy now that I have no reason any more to buy a 2600K, though it will now be impossible for me to build a dedicated folding rig any time soon.

 

1.425v, internal PLL enabled, LLC 75% but for 4.8ghz 1.35v and everything on auto, temps were around 70 degress underload at 5ghz, it's watercooled (here) but I had to have the fans at 1200rpm, it's now runs low to mid 60s at 600rpm.

 

Struth!! I was just about to order an I7 3930K + Asus P9X79 Pro in place of one of my 2600K rigs.

No point now as even this (12 threads) wouldn't be powerful enough to run -bigadv

Looks like its a dual processor server board or nothing then

 

Cool, thanks. That gives me a starting point, although I'm sure mine will vary a little from those settings it's good to know I was on the right track when testing the 2600K in my sig.

It's a shame that case didn't give you room to hide the cables, that looks like a nice build for the room you have.

 

Some of the server chips will have all 8 cores enabled, so will have 16 threads. That will make a cheaper option than going dual-processor (unless, of course, you have the finances and the inclination anyway).

 

Agreed, will just have to see how the market develops over the next few months.

 

There's a good thread here most seem to be getting similar results.

It looks messy I know, I used red sata cables which didn't help, but there's no window so it's not on permanent display

 

If I upgrade it will be to the new sr2, I don't trust PG enough to just scrape through with an octo core only for them to change the requirements again.

 

I read through that thread a couple of days ago. It's what motivated me to replace my 2600K with a 2700K only 3 weeks after buying it

 

It would be interesting to see how many folders they lose each time they make changes like this. Lets face it, we are all doing medical research but the measure of how much we feel like we contribute is based upon the points system, and everytime they make changes like this and we contribute less points, it makes you feel like you contribute less. If SMP units are SO IMPORTANT now because everyone with powerful machines likes to use them on units which 'push' their machines, then why not INCREASE THE POINTS of SMPs . Or why not create large smps with multiple units together or something. So many options? I hate smp units and having to go searching all the time for new units. There just seems to be so much down time inbetween units, which means that there are usually hours a day that the computer is not folding, rather than 45minutes every couple days. (I dont know if there is an smp server issue but my machine seems to go 'fishing' for quite a while after each unit).

The stupid thing is Prof. Vijay states 'Big Advanced (BA) is an experimental type of Folding@home WUs intended for the most powerful machines in FAH.' well then why not create a WU that needs the most powerful machine (bigbeta???) rather than just stopping the majority of 8 threaded users who machines are more than powerful enough from being able to do the work. If so many can do the units, wouldn't stanford be better making new units that fully utilize the current computing.

My thoughts are as a result of this people will fold less, people wont be buildings as many rigs to fold, now a 2600k will be a sub 30k ppd machine, it wont be as 'worth it'. Lets hope it doesnt hurt the science in the long run.

Thats my rant.

 

A remarkable change to be given two months notice - as we all know, in the past, changes like these have happened with only a few days notice, if any. Could it be that all the grumbles and the new Advisory Board are having an effect?

So now I have two months to finally wrestle with the heat sinks and get my dual Xeon machine up-and-running again. Should be possible

 

We build the machines..
We pay the electricity bills..
We fold the units ..

All we have is Kudos, In house rivalry and some points.. Now they are making it harder. Hmmm feels like we are getting shafted.

Looks like i will go from half a million points a week to nothing now.

 

That's my old q6600 on ebay then ....gee I am well behind the times