• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Stanford, what's next....

M

marty9876

[H]ard|DCer of the Month - February 2006
Joined
Jun 11, 2003
Messages
4,906
I dunno guys/gals/moose..... This is why people quit folding, and get fired.

With this latest round form Stanford, QMD's going to just any old machine regardless of configuration (flags) I'm left with a very uneasy feeling.

How the hell can anyone run this software in a corporate environment with any degree of certainty tomorrow things are not going to blow up? Yea, this might be over reacting a touch but I happen to like my job (well the checks don't bounce anyways). I've mentioned flirting with the idea of borging many, many more machines but this is basically the reason I have not. Well....good thing for my mortgage company I did not.

Permission letter to boss: Dear boss I want to install a program which someday might tear you open a new one- sound good?

Well so much for that plan.

People get pissed when Stanford pulls all QMD's from their assignment servers, well now I know why. Stanford can't control where they go, and I have little faith (Read put my job on it) what Stanford will have better luck next time.

For the record I love QMD's, where I can run them. Stanford let the genie out of the bottle with these, I just hope the benfit to all mankind with the QMD program out wieghs the damage Stanford is doing to themselves.

Sorry for the rant, I just hate seeing all that power go to waste.

Offical thread at Stanford (has been posted here)

OC-AMD turning things way down
 
damn, that sucks... guess now wouldn't be the time to approach my dad with the "try it and see" idea for 75-90ghz of dells. especially cause they only have 256mb mem each and with this... very bad.

sorry to see oc-amd fading, too. least he's willing to stick around and see if they can fix it.

and yeah, they really shot themselves in the foot with this. a lot of corporate folders/school folders (including flecom, i believe) are gonna get hurt by this... hopefully people don't lose any jobs over this. (outside of stanford...)

the black knight always triumphs!
 
good grief; seems like stanford needs to get their act together -- they just lost (for good, this time?) one of their best producers ever...

good job, stanford :rolleyes:

 
Spectre said:
I say we send Stanford my leftover vegemite!
Click to expand...

And convince them to do what with it....?
Like have to eat it when they screw up like this!

 
Mayhem33 said:
And convince them to do what with it....?
Like have to eat it when they screw up like this!

Click to expand...
i say we force not only them, but their children, wives/husbands, and small pets to eat it as well!...





...and then make them use it as thermal compound :eek:

 
Ever since early/mid summer when they QMD memory issues came up, there's been something seemingly off with those units. Alot of the problem the way I see it, is they've let it get out of hand only to throw a bandaid on the issue.

Some of the talk has centered around well if more people ran the beta... well it's labeled as beta for a reason.. not everyone is going to run it. I guess at this point, I'd almost rather see them offer a handful of clients that hit specific cores only, and let the end user choose. The estimatining of peoples capability and such just isn't working.

 
p[H]ant0m said:
Ever since early/mid summer when they QMD memory issues came up, there's been something seemingly off with those units. Alot of the problem the way I see it, is they've let it get out of hand only to throw a bandaid on the issue.

Some of the talk has centered around well if more people ran the beta... well it's labeled as beta for a reason.. not everyone is going to run it. I guess at this point, I'd almost rather see them offer a handful of clients that hit specific cores only, and let the end user choose. The estimatining of peoples capability and such just isn't working.

Click to expand...

I don't know if the beta client would have stopped this or not. This time it's not about misconfigured clients or what not, Stanford bypassed the clients configuration and gave the clients QMD's. Not on purpose, Stanford ran out of WU, and gave the clients what they had left- QMD's.

The existing safety measures were bypassed by no fault of the end user.
 
Does this mean we're all jumping to UD?? :confused:

What has been done? Have they pulled the plug on the QMD's? Are my borgs safe?

<worried>

 
Aratech said:
Does this mean we're all jumping to UD?? :confused:

What has been done? Have they pulled the plug on the QMD's? Are my borgs safe?

<worried>

Click to expand...


Stanford is redoing things on their end. The damage, while very limited from what I can tell has already been done- I'd say your borgs are safe.
 
marty9876 said:
I don't know if the beta client would have stopped this or not. This time it's not about misconfigured clients or what not, Stanford bypassed the clients configuration and gave the clients QMD's. Not on purpose, Stanford ran out of WU, and gave the clients what they had left- QMD's.

The existing safety measures were bypassed by no fault of the end user.
Click to expand...

The prevailing mindset on folding-community is that if people were running the beta and specifying low memory levels, this *could* have been avoided. But I agree it's unlikely, again just a bandaid for a larger problem.

Who honestly would have thought they would have run out of WU's though? That in itself is probably more surprising than the fact that they all just started assigning something.

 
I really, really hope nome of my 256mb 1ghz celerons didn't get any of these... I should send an email out to my contact and ask if any major slowdowns have occured. :( It'll be all good since really they just type emails on those computers.

It's nice to see that Stanford reacted so quickly and got this issue patched up. As it was said in one of those threads it's not like Microsoft hasn't messed up, but sadly Microsoft programs are a necessity and F@H isn't - in many eyes this is sad but true.

-MN Scout
 
upriverpaddler said:
I'm confused. Could you sum up the issue in laypersons terms?
Click to expand...

QMD's can cripple systems for anything but folding.
QMD's are restricted to a specific subset of machines after much uproar from the community.
Allocation program ran out of units so started giving all types of systems QMD's.
Uh-oh :eek:
 
If you system doesn't have 1 gig of ram then you shouldn't run QMDs. I've seen them take up to 500mbs of memory, and they really saturate the memory bus, so you definately don't want 2 clients with QMDs running because you're system wont respond unless you have 2 gbs of ram, it'll seriously crawl (I've had the happen on my xeon system, my fault totally).

The problem is that the main servers ran out of normal wus, and the server that assigns everything decided to fall back to the server that distributes QMDs and gave them to all computers no matter what flags and what settings and what memory amount they had. To get them you have to have, big wus enabled and be running the advmethods flag.

The problem is fixed now, so it's all good.

-MN Scout
 
You must log in or register to reply here.
Back
Top