Deadlines

MikeSW17
MikeSW17
Joined: 18 Jan 05
Posts: 2
Credit: 13115
RAC: 0
Topic 188557

Why are Einstein Report Deadlines so tight?

I have three systems running BOINC all 75% SETI, 12.5% Einstein and 12.5% LHC.

The 2 desktop (running 24/7) systems just manage to run the allocated Einstein work within the deadlines.

I also have a laptop that doesn't run BOINC when on batteries - so its perhaps 12-14 hours per day. It NEVER quite has enough time to run an Enstein WU.

By the time a WU comes to start, it has perhaps 3 days left of the deadline, but will need 4 days.

The deadlines are just too tight.

I don't want to drop Einstein - SETI is my first priority, LHC and Einstein are my backup projects. and LHC hasn't had any work for a while.

RandyC
RandyC
Joined: 18 Jan 05
Posts: 6574
Credit: 111139797
RAC: 0

Deadlines

> Why are Einstein Report Deadlines so tight?
>
> I have three systems running BOINC all 75% SETI, 12.5% Einstein and 12.5%
> LHC.
>
> The 2 desktop (running 24/7) systems just manage to run the allocated Einstein
> work within the deadlines.
>
> I also have a laptop that doesn't run BOINC when on batteries - so its perhaps
> 12-14 hours per day. It NEVER quite has enough time to run an Enstein WU.
>
> By the time a WU comes to start, it has perhaps 3 days left of the deadline,
> but will need 4 days.
>
> The deadlines are just too tight.
>
> I don't want to drop Einstein - SETI is my first priority, LHC and Einstein
> are my backup projects. and LHC hasn't had any work for a while.
>
You need to reduce your cache size by reducing the 'Connect to Network' number of days. I believe the default setting is two days, which (for Einstein) is too long for slow machines, or ones not running 24/7.

Go to your account page and edit General Preferences...Edit Network usage and set Connect to Network to 0.5 or 1.0 (even less if you have SLOOWWW machines).

If you have WUs that have run past deadline, use the BOINC manager to abort them. (This assumes you're using 4.2x). I also recommend you change your resource allocation to 50% Seti, 25% Einstein, 25% LHC...at least for the system that runs past deadline (use your School/Work/Home preferences to differentiate your slow system from the others).

Seti Classic Final Total: 11446 WU.

STE\/E
STE\/E
Joined: 18 Jan 05
Posts: 135
Credit: 144047410
RAC: 15000

You need to reduce your cache

You need to reduce your cache size by reducing the 'Connect to Network' number of days. I believe the default setting is two days, which (for Einstein) is too long for slow machines, or ones not running 24/7.

Go to your account page and edit General Preferences...Edit Network usage and set Connect to Network to 0.5 or 1.0 (even less if you have SLOOWWW machines).

If you have WUs that have run past deadline, use the BOINC manager to abort them. (This assumes you're using 4.2x). I also recommend you change your resource allocation to 50% Seti, 25% Einstein, 25% LHC...at least for the system that runs past deadline (use your School/Work/Home preferences to differentiate your slow system from the others).
==========

A lot of people don't want to do that Randy, all they want to do is run Seti 90-95% of the time and use the side Projects just for a backup. Thats the main reason why they want the deadline increased, so they can take 2 weeks to finish 1 WU and keep everybody that has the same WU waiting for their credit ...

John McLeod VII
John McLeod VII
Moderator
Joined: 10 Nov 04
Posts: 547
Credit: 632255
RAC: 0

What is really needed is a

What is really needed is a change in the client side scheduler so that it doesnt download so much more work than the cache size requested and it pays attention to deadlines. This would be offset by having the low resource share projects have work on the computer less frequently.

genes
genes
Joined: 10 Nov 04
Posts: 41
Credit: 2705410
RAC: 8436

CPDN is a good backup project

CPDN is a good backup project -- you'll get one WU and a year to finish it!

OTOH, I now use 0.5 days for the cache on everything from 933MHz P3 to 3.2GHz P4(HT). I never get too much work on any system. (all 24/7)


John McLeod VII
John McLeod VII
Moderator
Joined: 10 Nov 04
Posts: 547
Credit: 632255
RAC: 0

> CPDN is a good backup

Message 8591 in response to message 8590

> CPDN is a good backup project -- you'll get one WU and a year to finish it!
>
> OTOH, I now use 0.5 days for the cache on everything from 933MHz P3 to 3.2GHz
> P4(HT). I never get too much work on any system. (all 24/7)
>
CPDN really isn't that good a backup project. Given one of my machines, you may have a year to complete the work, but you have more than a month of processing to get it done. That is a factor of 12 for deadline to crunch. S@H on the other hand has a deadline of 2 weeks, but you can finish 5/day, this is a factor of 70 for deadline to crunch. Einstein has a 7 day deadline (I believe) and the same CPU can get about 1.5 done per day for a factor of about 12.

Comment viewing options

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