Too fast

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157718
RAC: 0

RE: Michael, I think you

Message 24381 in response to message 24380

Quote:
Michael, I think you should have more faith in the project Admins on their MDQ change. I'm sure they have done the equivelent of a 'cost/benefit' analysis on this and they realise that more and more dual core/faster boxs are coming on line and they will have to adjust the MDQ sooner or later. Yes, you may have to wait longer for a 'problem' WU but they are still 'money in the bank'...Cheers, Rog.

Rog,

I'd say the change is more like a knee-jerk reaction, should have at least waited until either benchmark/WU-length or MDQ/WU-length matching and abort-queued-work-at-uninstall could be implemented, and add that the Admins are notoriously absent on the helldesk.

I too have been plaqued with the "fast processor/short WU blues", so I am not entirely unsympathic to you guys. I don't know how many more of them remain on the datafile I've been assigned, but the in-between # on them started at ~540 and is only down to 309 after 9 days, implying that they'll be around for another 2 weeks, at least. Worse than that, I'm usually paired with 2 PowerMacs, claiming only 8 credits/WU, so the "money in the bank" is pennies on the dollar. I love E@H, but I've put way too much time and effort into making this Athlon XP rig perform like an Athlon 64 4000 to "sell" it's time for only 8 credits/hour. I'm as frustrated as the rest of you speed demons here, but a large share of my rig is now working at earning ~21 credits/hour at Rosetta and ~15 c/h at Seti, at least until the 1-hour shorties have finished.

Respects to all,

Michael

edited for explanation

microcraft
"The arc of history is long, but it bends toward justice" - MLK

AnRM
AnRM
Joined: 9 Feb 05
Posts: 213
Credit: 4346941
RAC: 0

RE: I'd say the change is

Message 24382 in response to message 24381


Quote:
I'd say the change is more like a knee-jerk reaction, should have at least waited until either benchmark/WU-length or MDQ/WU-length matching and abort-queued-work-at-uninstall could be implemented, and add that the Admins are notoriously absent on the helldesk.

>Yes, Michael, you've made good points here but I wonder how easy those changes would be or how long they would take. Also your point about attaching to another project is well taken (as I have had to do) although I have problems with just giving them such a tiny slice of time out of respect. I give most back up projects 10% minimum.....Cheers, Rog.

Wurgl (speak^Wcrunching for Special: Off-Topic)
Wurgl (speak^Wc...
Joined: 11 Feb 05
Posts: 321
Credit: 140550008
RAC: 0

RE: I'm sorry, but I don't

Message 24383 in response to message 24379

Quote:
I'm sorry, but I don't have time to be patient. I'm on a rather short "deadline" (bad pun) myself.

Cool down man, the next heart attack is just waiting :-)

To be serious. There are people/machines who support only one project, for whatever reason. And there are people supporting multiple projects, for whatever other reason. One group likes this change, the other doesn't.

In my team there is a group of people arguing similar as you do. "Oh my god, so many pendings, this brain dead scheduler" and so on. I do not care. Pendings are as good as granted. Pendings are the future.

And if you want to reach a better position, well if you have fewer pendings, the other will have fewer too, so it does not raise your position.

A last point: There is a big cluster at the Potsdam Einstein institute. Some of machines of this cluster got hit too. Do you really think that a cluster which is almost dedicated to this project should support some other BOINC project just for a few guys here complaining about pendings?

They changed it and this is fine.

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157718
RAC: 0

RE: A last point: There is

Message 24384 in response to message 24383

Quote:

A last point: There is a big cluster at the Potsdam Einstein institute. Some of machines of this cluster got hit too. Do you really think that a cluster which is almost dedicated to this project should support some other BOINC project just for a few guys here complaining about pendings?

They changed it and this is fine.

That cluster has possibly ONE host out of 191 total, the Opteron, possibly capable of processing more than 16 WU/day/CPU, so your point is vacuous. My single-core Athlon XP is capable of nearly as much throughput as any of their Athlon XP multicores.

microcraft
"The arc of history is long, but it bends toward justice" - MLK

Wurgl (speak^Wcrunching for Special: Off-Topic)
Wurgl (speak^Wc...
Joined: 11 Feb 05
Posts: 321
Credit: 140550008
RAC: 0

RE: That cluster has

Message 24385 in response to message 24384

Quote:
That cluster has possibly ONE host out of 191 total, the Opteron, possibly capable of processing more than 16 WU/day/CPU, so your point is vacuous. My single-core Athlon XP is capable of nearly as much throughput as any of their Athlon XP multicores.

Nope! The Opteron is lucky and gets long ones. A simple AMD Athlon(TM) MP 2200 is hit. http://einsteinathome.org/host/76687/tasks or scroll down then you see WUs which took 2000 seconds here: http://einsteinathome.org/host/386795/tasks

And for example my Powerbook, sure not a top model machine. Almost every reasonable fast machine can be a victim of this problem. It is just a question of luck.

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157718
RAC: 0

RE: Nope! The Opteron is

Message 24386 in response to message 24385

Quote:
Nope! The Opteron is lucky and gets long ones. A simple AMD Athlon(TM) MP 2200 is hit. http://einsteinathome.org/host/76687/tasks

Bad example! It's a MP, so already gets 32 WUs/day. Possible to be idle no more than 1 hr/day/core.

Quote:
... or scroll down then you see WUs which took 2000 seconds here: http://einsteinathome.org/host/386795/tasks

An even slower host, "crippled" by what? 9 shorties? That host doesn't even approach 16 WU/day/core! Sorry, more homework necessary. :-)

Quote:
And for example my Powerbook, sure not a top model machine. Almost every reasonable fast machine can be a victim of this problem. It is just a question of luck.

I hope that you're not trying to posit that Einstein is likely to lose the Institute as a contributor. On the other hand, do you really suppose that waiting an eternity for credits to start returning is not going to cost E@H a whole slew of crunchers, especially noobs?

edited for BBcode boo-boo

microcraft
"The arc of history is long, but it bends toward justice" - MLK

Wurgl (speak^Wcrunching for Special: Off-Topic)
Wurgl (speak^Wc...
Joined: 11 Feb 05
Posts: 321
Credit: 140550008
RAC: 0

RE: On the other hand, do

Message 24387 in response to message 24386

Quote:
On the other hand, do you really suppose that waiting an eternity for credits to start returning is not going to cost E@H a whole slew of crunchers, especially noobs?

An eternity? This is my oldest one. And I remember one, where I had to wait 1 month, this is not an eternity. Keep cool, live is hard enough :-)

BTW: Why did you abort so many (19) WUs?

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157718
RAC: 0

RE: An eternity? This is my

Message 24388 in response to message 24387

Quote:

An eternity? This is my oldest one. And I remember one, where I had to wait 1 month, this is not an eternity. Keep cool, live is hard enough :-)

BTW: Why did you abort so many (19) WUs?

Eternity? Yes, to a noobie. Three days, and many start to post about "Where are my credits?" I spend a lot of time on the helldesk, and some on this side of the boards, and I believe I have a pretty good understanding of the psyche involved in starting up.

An answer to your question, and it is a fair question, I'll add. :-)
I aborted those WUs so that they can be re-issued and off the DB and into the evaluation process ASAP. As to Why did I refuse to crunch them (what you probably really meant to ask), look more carefully at any of them and tell me how much credit was possible out of any one of them. I made a statement, three or four posts ago, about "selling" my lovingly-built and tweaked rig cheaply, and I'll stand by it.

edited for typos

Addition: The WU you linked to is the oldest one now, before the 32 MDQ goes into effect. Check back a month from now, and tell me how wrongly I guesstimated.

microcraft
"The arc of history is long, but it bends toward justice" - MLK

Wurgl (speak^Wcrunching for Special: Off-Topic)
Wurgl (speak^Wc...
Joined: 11 Feb 05
Posts: 321
Credit: 140550008
RAC: 0

RE: ... As to Why did I

Message 24389 in response to message 24388

Quote:
... As to Why did I refuse to crunch them (what you probably really meant to ask), look more carefully at any of them and tell me how much credit was possible out of any one of them. ...

A quote from your posting a little bit deeper in this thread:

Quote:
... How selfish can a few people be, to cause problems for the majority instead of using the designed-in solution? ...

Have a nice day.

Michael Roycraft
Michael Roycraft
Joined: 10 Mar 05
Posts: 846
Credit: 157718
RAC: 0

RE: RE: ... As to Why did

Message 24390 in response to message 24389

Quote:
Quote:
... As to Why did I refuse to crunch them (what you probably really meant to ask), look more carefully at any of them and tell me how much credit was possible out of any one of them. ...

A quote from your posting a little bit deeper in this thread:

Quote:
... How selfish can a few people be, to cause problems for the majority instead of using the designed-in solution? ...

Have a nice day.

Problem? How? They went back for re-issue almost immediately. You sound as if it is your opinion that the project has some sort of proprietary interest in my rig, that I am obligated to someone. That might...marginally be the case if they properly valuated Tflop-hours, electric costs for said Tflops, and maintanance costs, accounted for them, and issued me a yearly or quarterly statement for tax-exempt donations to a non-profit organization. That isn't likely to happen any time soon, it isn't even under consideration, but the same University involved is required to do that for any other donations, so please explain to me - why not for this one?

microcraft
"The arc of history is long, but it bends toward justice" - MLK

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.