WU's should be recycled starting today

Message boards : Number crunching : WU's should be recycled starting today
Message board moderation

To post messages, you must log in.

AuthorMessage
EclipseHA

Send message
Joined: 28 Jul 99
Posts: 1018
Credit: 530,719
RAC: 0
United States
Message 5145 - Posted: 8 Jul 2004, 4:45:02 UTC

Based on the 6/22 (or 22/6) startdate, many WU's expired today (7/7) (all from 6/22 without 3 results). We all know they are out there (per UCB), but has anybody seen them?
(finish a WU, get credit right away in some cases)
ID: 5145 · Report as offensive
Heffed
Volunteer tester

Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 5158 - Posted: 8 Jul 2004, 5:20:31 UTC

Until the validator gets caught up, there will be no way to tell.

<a> [/url]
ID: 5158 · Report as offensive
EclipseHA

Send message
Joined: 28 Jul 99
Posts: 1018
Credit: 530,719
RAC: 0
United States
Message 5172 - Posted: 8 Jul 2004, 6:35:50 UTC - in response to Message 5158.  

> Until the validator gets caught up, there will be no way to tell.


There's "no way to tell" about many aspects of Seti/Boinc!

So, why do you think the validator is behind? (implies that it's slow and running!)

I think it's just down right broken! (like most of the Seti/Boinc launch, IMHO)

If the validator was just "slow" why don't I have the credit for the WU's that had 3 valid results a week ago! (last time they let us peons see our own work)!

Heck the "live launch" has ~30k users, ~50k CPUS, while the beta had ~20k users.. Not a big difference (I got CPU 52xxx just last night, and a bunch of folks have more than 1 cruncher) The load here is not much above what could be done by the beta testers!

There's no excuse for the mess that UCB has created here! This is NOT an inital release as Set 1 was! This crap has been in beta for over a year, and had a resource of +100k active users and 5 years of history to learn from! Seti 1 didn't have that!
ID: 5172 · Report as offensive
Profile Rom Walton (BOINC)
Volunteer tester
Avatar

Send message
Joined: 28 Apr 00
Posts: 579
Credit: 130,733
RAC: 0
United States
Message 5179 - Posted: 8 Jul 2004, 7:07:14 UTC - in response to Message 5172.  

> Heck the "live launch" has ~30k users, ~50k CPUS, while the beta had ~20k
> users.. Not a big difference (I got CPU 52xxx just last night, and a bunch of
> folks have more than 1 cruncher) The load here is not much above what could
> be done by the beta testers!

Hmmmmmmmm, On beta's best day we received 23,000 results in a day. Before things went south for public we were receiving 80,000 results a day. I haven't been able to run any of the management queries in two weeks because it would take to much DB Server crunch time away from the scheduler/transitioner/feeder/validator/splitter.

After making a few changes to various components on the system, like a new feeder today, we have handed out 100,000 results in just the last 3 hours.

If I had to guesstimate about the number of results we will see a day after this feeder change, I would say we have probably broken the 150,000-200,000 results a day marker.

I would say, the load is quite different than what we experienced with the Beta project.

----- Rom
ID: 5179 · Report as offensive
Angstrom

Send message
Joined: 20 Sep 99
Posts: 205
Credit: 10,131
RAC: 0
United Kingdom
Message 5180 - Posted: 8 Jul 2004, 7:08:56 UTC

I'm with Woody on this.

Looking at some of the posts it would seem that the launch of Boinc is OK because it hasnt gone quite as badly as the original Seti classic launch.

The original launch was carried out as a pioneering attempt to connect computing power around the world. That was ground breaking stuff. This is just a new piece of software. Yes bugs are expected even though its beta tested however there is no excuse for the WU bottleneck.

The people running this project should be the best placed in the world to know what was goingf to happen when they went live and make suitable preparations to handle it.

Neil

PS Seti team please update the news with the current situation etc....
ID: 5180 · Report as offensive
Heffed
Volunteer tester

Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 5397 - Posted: 8 Jul 2004, 22:00:29 UTC - in response to Message 5180.  

> The original launch was carried out as a pioneering attempt to connect
> computing power around the world. That was ground breaking stuff. This is just
> a new piece of software. Yes bugs are expected even though its beta tested
> however there is no excuse for the WU bottleneck.

Comparing BOINC to S@H classic is comparing apples to oranges.

<a> [/url]
ID: 5397 · Report as offensive
JAF
Avatar

Send message
Joined: 9 Aug 00
Posts: 289
Credit: 168,721
RAC: 0
United States
Message 5413 - Posted: 8 Jul 2004, 22:40:44 UTC - in response to Message 5397.  

> > The original launch was carried out as a pioneering attempt to connect
> > computing power around the world. That was ground breaking stuff. This is
> just
> > a new piece of software. Yes bugs are expected even though its beta
> tested
> > however there is no excuse for the WU bottleneck.
>
> Comparing BOINC to S@H classic is comparing apples to oranges.
>
> <a>
ID: 5413 · Report as offensive
Heffed
Volunteer tester

Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 5443 - Posted: 8 Jul 2004, 23:31:59 UTC - in response to Message 5413.  

> You mean because Classic started at zero users and Boinc Seti started with
> quite a few? Seems like that should have been taken into account, if that's
> the case.
>
> I'm not sure why they couldn't have a good estimate of the number of initial
> participants when Boinc Seti went "live". They had all the Seti Classic data
> to work from.
>
> I'm not saying you are wrong - I just would like to understand more about the
> project.

No, what I'm saying is the support infrastructure for BOINC is totally different from what was needed for classic. I think it's trying to get all the systems to play well together is one of the big issues.

<a> [/url]
ID: 5443 · Report as offensive
Profile nightowl
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 70
Credit: 948
RAC: 0
Canada
Message 5657 - Posted: 9 Jul 2004, 15:38:49 UTC - in response to Message 5397.  


> Comparing BOINC to S@H classic is comparing apples to oranges.

no -- oranges to pears
ttyl
Jeff
ID: 5657 · Report as offensive
Profile nightowl
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 70
Credit: 948
RAC: 0
Canada
Message 5658 - Posted: 9 Jul 2004, 15:40:04 UTC - in response to Message 5443.  

> No, what I'm saying is the support infrastructure for BOINC is totally
> different from what was needed for classic. I think it's trying to get all the
> systems to play well together is one of the big issues.
>

not to mention some of the needed computers are still running classic seti servers!! People have to calm down and give new seti a break! until the crossover is complete, there will be problems.
ttyl
Jeff
ID: 5658 · Report as offensive
Profile Stephen Macy
Avatar

Send message
Joined: 8 May 99
Posts: 167
Credit: 1,774,063
RAC: 0
United States
Message 6054 - Posted: 10 Jul 2004, 11:37:01 UTC - in response to Message 5172.  

> > Until the validator gets caught up, there will be no way to tell.
>
>
> There's "no way to tell" about many aspects of Seti/Boinc!
>
> So, why do you think the validator is behind? (implies that it's slow and
> running!)
>
> I think it's just down right broken! (like most of the Seti/Boinc launch,
> IMHO)
>
> If the validator was just "slow" why don't I have the credit for the WU's that
> had 3 valid results a week ago! (last time they let us peons see our own
> work)!
>
> Heck the "live launch" has ~30k users, ~50k CPUS, while the beta had ~20k
> users.. Not a big difference (I got CPU 52xxx just last night, and a bunch of
> folks have more than 1 cruncher) The load here is not much above what could
> be done by the beta testers!
>
> There's no excuse for the mess that UCB has created here! This is NOT an
> inital release as Set 1 was! This crap has been in beta for over a year, and
> had a resource of +100k active users and 5 years of history to learn from!
> Seti 1 didn't have that!
>
>If you think it is so bad, why are you still here?
ID: 6054 · Report as offensive

Message boards : Number crunching : WU's should be recycled starting today


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.