Why the short due date with long workunits?

DonCorleone
DonCorleone
Joined: 25 Dec 06
Posts: 2
Credit: 60,277
RAC: 0

These WUs can't be completed

These WUs can't be completed when you run BOINC some hours every day, I'm running an average PC with a 2GHz CPU, and well, I just wasted about 50h CPU time because the deadline ended yesterday, and the WU will be completed in about an hour. I think it's a big mistake to force BOINC unsers which process Einstein@HOME WUs to run their PCs all the time, because they won't do it, at least I won't do.

Until this period is over, I'll let BOINC process some WUs from other projects.

Odysseus
Odysseus
Joined: 17 Dec 05
Posts: 372
Credit: 16,766,208
RAC: 6,920

RE: These WUs can't be

Message 66967 in response to message 66966

Quote:
These WUs can't be completed when you run BOINC some hours every day, I'm running an average PC with a 2GHz CPU, and well, I just wasted about 50h CPU time because the deadline ended yesterday, and the WU will be completed in about an hour. I think it's a big mistake to force BOINC unsers which process Einstein@HOME WUs to run their PCs all the time, because they won't do it, at least I won't do.


You mean this one? As it happens your original partner also missed the deadline, so as long as you can return your result before at least one of the two hosts that were issued new copies yesterday, your time won’t have been wasted.

I agree that the deadlines should have been increased, even if not as much as the crunching times did. My systems do run all the time, but the slower ones, running several projects in addition to E@h, are feeling some deadline pressure.

DonCorleone
DonCorleone
Joined: 25 Dec 06
Posts: 2
Credit: 60,277
RAC: 0

RE: RE: These WUs can't

Message 66968 in response to message 66967

Quote:
Quote:
These WUs can't be completed when you run BOINC some hours every day, I'm running an average PC with a 2GHz CPU, and well, I just wasted about 50h CPU time because the deadline ended yesterday, and the WU will be completed in about an hour. I think it's a big mistake to force BOINC unsers which process Einstein@HOME WUs to run their PCs all the time, because they won't do it, at least I won't do.

You mean this one? As it happens your original partner also missed the deadline, so as long as you can return your result before at least one of the two hosts that were issued new copies yesterday, your time won’t have been wasted.

I agree that the deadlines should have been increased, even if not as much as the crunching times did. My systems do run all the time, but the slower ones, running several projects in addition to E@h, are feeling some deadline pressure.

Yes, it was this one. Do you know if there a plans to increase the deadlines for S5R2 WUs? It's great that I was lucky this time, but my luck is nothing I will bet on for the next WUs.

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3,516
Credit: 518,661,214
RAC: 721,657

RE: Three hours? Yeah,

Message 66969 in response to message 66965

Quote:
Three hours? Yeah, sounds realistic. I'm surprised myself that it's so short... My Core Duo (okay not the fastest out there but certainly not a "slow" host either) does a little under 15 credits/hour (running Linux, but we had that in another thread ;-) ) which would mean about 7 hours for a result... okay, but it has two cores, which statistically means it has to get a new WU every 3.5 hours... Take these numbers, think of the Xeon Quads out there, and I think you can understand why the new WUs are so huge... I can.

Yup!

Take this little Xeon for example: http://einsteinathome.org/host/916397

Over 30 credits per CPU hour per core (and it has 8 cores !!!!) > 240 credits per CPU hour . What a monster!

As to the deadline issue: I haven't heard of any plans to extend the deadline, but optimiztion of the app will eventually happen and I don't think the deadlines will be lowered when this happens.

Has anybody done a cross-projects comaparison of (credits)/(deadline length)?

CU

BRM

Winterknight
Winterknight
Joined: 4 Jun 05
Posts: 663
Credit: 221,890,006
RAC: 39,075

I also think the deadlines

I also think the deadlines are too short, even though it does not affect my families computers. But looking at my sons Pent 4, 3.00GHz, it takes approx 250,000 sec (~70hrs) to complete a 425cr unit. That is almost a quarter of the time until the deadline. Therefore the computer needs to be on and Einstein processing un-interrupted for 6 hrs/day. Or if it is an office computer, (Mon - Fri only) 7 hrs/day minimum. This assumes the work cache is minimum.

When Seti first started, Berkeley only assumed a computer would be processing 10% of the time. And their minimum spec machine, I believe was 100MHz, this has since been raised.

I assume that the average cruncher, i.e. not a 24/7 BOINCaholic like us, only has their computer on either during office hrs, or evenings and w/ends for a home computer. Therefore any computer less than a 3 GHz, or eqiuvalent, is probably going to miss the deadline.

Andy

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

RE: I also think the

Message 66971 in response to message 66970

Quote:

I also think the deadlines are too short, even though it does not affect my families computers. But looking at my sons Pent 4, 3.00GHz, it takes approx 250,000 sec (~70hrs) to complete a 425cr unit. That is almost a quarter of the time until the deadline. Therefore the computer needs to be on and Einstein processing un-interrupted for 6 hrs/day. Or if it is an office computer, (Mon - Fri only) 7 hrs/day minimum. This assumes the work cache is minimum.

When Seti first started, Berkeley only assumed a computer would be processing 10% of the time. And their minimum spec machine, I believe was 100MHz, this has since been raised.

I assume that the average cruncher, i.e. not a 24/7 BOINCaholic like us, only has their computer on either during office hrs, or evenings and w/ends for a home computer. Therefore any computer less than a 3 GHz, or eqiuvalent, is probably going to miss the deadline.

Andy

Alternatively, if we could only have MMX at this point, not even SSE, the deadline issues would likely be moot for all but vintage Pentium class and older... I can't believe that someone would be using a 486, but just a caveat just in case there is...

Donald A. Tevault
Donald A. Tevault
Joined: 17 Feb 06
Posts: 439
Credit: 73,516,529
RAC: 0

RE: Alternatively, if we

Message 66972 in response to message 66971

Quote:
Alternatively, if we could only have MMX at this point, not even SSE, the deadline issues would likely be moot for all but vintage Pentium class and older... I can't believe that someone would be using a 486, but just a caveat just in case there is...

Hey, I still want to use my 200 Mhz Cyrix 6x86.

w1hue
w1hue
Joined: 28 Aug 05
Posts: 18
Credit: 208,388,223
RAC: 233,447

"the largest potential WUs

"the largest potential WUs (acording to a graph posted by Bernd) are about 630 credits, noone's seen any above 530. Unless your machine spends alot of cpu time doing other stuff than crunch, you should be able to get a 630credit WU completed within the deadline at 24/7."

1) My machine is doing other work, and 2) I ain't gonna run my machine 24/7 just so the project can get one more lousy data point.

Dr. Atanu Maulik
Dr. Atanu Maulik
Joined: 5 Jan 06
Posts: 16
Credit: 5,308,855
RAC: 0

One must realise that not all

One must realise that not all computers are running 24/7. For those which are running 3-4 hrs a day it is impossible to finish the job on time even if they are running only one project. So the workunits must be shortened. Untill that is done I see no option but to move on to some other projects for my home computer. I am sure many more must be feeling the same way. Are the people concerened taking note ?

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494,410
RAC: 0

The problem is that CPU

The problem is that CPU performance varies more and more (meaning the difference between the slowest and the fastest hosts is growing simply gigantic). If the WUs are long, slow hosts can't cope. If the WUs are short, the fast dual, quad or "whatever" cores will kill off the server. I think what we need is what we had during the S5R1 science run: Decide whether a host counts as "slow" or "fast" and send only the shorter kind of WU to slow ones. It's the best solution I can come up with atm and I do hope the project staff is going to implement it asap (yes, I know they're up to their necks in work already but I have no idea how the crunchers should get along without that feature... honestly... not with 200 MHz Pentiums as well as hyperthreaded Quad Xeons out there)

Comment viewing options

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