Blip (Sep 16 2009)


log in

Advanced search

Message boards : Technical News : Blip (Sep 16 2009)

Previous · 1 · 2 · 3 · Next
Author Message
Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4048
Credit: 32,693,420
RAC: 508
United Kingdom
Message 941052 - Posted: 18 Oct 2009, 9:05:31 UTC - in response to Message 941048.
Last modified: 18 Oct 2009, 9:06:29 UTC

My preferences are
SETI@home Enhanced: yes
Astropulse: yes
Astropulse v5: yes

So, what software version will process Astropulse units I receive ?

If you have set "If no work for selected applications is available, accept work from other applications?" to "no," you won't get any.

Gruß,
Gundolf
[edit]If you wanna know why, ask in the "Questions and answers" forum or in "Number crunching."[/edit]


My apologies for not posting all the options. They are

Run only the selected applications
SETI@home Enhanced: yes
Astropulse: yes
Astropulse v5: yes
If no work for selected applications is available, accept work from other applications? yes

You still won't get any Astroulse_v505 work, because you'll almost always get Seti@home enhanced work when you ask for it, so the last option doesn't come into play,
to get it you need to set SETI@home Enhanced to No, In this case you'll ask for Astropulse and Astropulse_v5 work, but you won't get and of them, (None available)
so you'll now get sent Seti@home enhanced and Astropulse_v505 instead.

Claggy
____________


Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 941084 - Posted: 18 Oct 2009, 12:24:14 UTC - in response to Message 941052.


You still won't get any Astroulse_v505 work, because you'll almost always get Seti@home enhanced work when you ask for it, so the last option doesn't come into play,
to get it you need to set SETI@home Enhanced to No, In this case you'll ask for Astropulse and Astropulse_v5 work, but you won't get and of them, (None available)
so you'll now get sent Seti@home enhanced and Astropulse_v505 instead.

Claggy


Thanks, that seems to confirm my earlier point that there is no point in having more than one Astropulse switch, as the version of the software to process a given work unit is not under the control of the user.
____________

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4048
Credit: 32,693,420
RAC: 508
United Kingdom
Message 941090 - Posted: 18 Oct 2009, 13:26:02 UTC - in response to Message 941084.


You still won't get any Astroulse_v505 work, because you'll almost always get Seti@home enhanced work when you ask for it, so the last option doesn't come into play,
to get it you need to set SETI@home Enhanced to No, In this case you'll ask for Astropulse and Astropulse_v5 work, but you won't get and of them, (None available)
so you'll now get sent Seti@home enhanced and Astropulse_v505 instead.

Claggy


Thanks, that seems to confirm my earlier point that there is no point in having more than one Astropulse switch, as the version of the software to process a given work unit is not under the control of the user.

The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy

Profile Duke of Buckingham SETI.USA
Volunteer tester
Send message
Joined: 19 Jun 09
Posts: 4
Credit: 148,795
RAC: 0
Portugal
Message 941102 - Posted: 18 Oct 2009, 13:52:24 UTC
Last modified: 18 Oct 2009, 14:29:22 UTC

I have one task pending since 30 july, task 1320233368. Is it normal?

Ricardo

Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 941110 - Posted: 18 Oct 2009, 14:48:44 UTC - in response to Message 941090.


The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy


Which brings me back to my original point: why not have just ONE Astropulse user switch ?

____________

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4048
Credit: 32,693,420
RAC: 508
United Kingdom
Message 941118 - Posted: 18 Oct 2009, 15:33:51 UTC - in response to Message 941110.
Last modified: 18 Oct 2009, 15:45:10 UTC

The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy

Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because there's been different Astropulse Applications?

Claggy

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4048
Credit: 32,693,420
RAC: 508
United Kingdom
Message 941120 - Posted: 18 Oct 2009, 15:53:53 UTC - in response to Message 941110.


The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy


Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because the project hasn't had the time to take out the old Astropulse switches and put in the new Astropulse_v505 switch?

Claggy

Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 941143 - Posted: 18 Oct 2009, 17:37:05 UTC - in response to Message 941120.


The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy


Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because the project hasn't had the time to take out the old Astropulse switches and put in the new Astropulse_v505 switch?

Claggy


Since the work units define the software version to be used, only two switches are needed:

SETI@home Enhanced: yes / no
Astropulse: yes / no

The others are superfluous.

____________

Profile [seti.international] Dirk Sadowski
Volunteer tester
Avatar
Send message
Joined: 6 Apr 07
Posts: 7028
Credit: 59,317,484
RAC: 21,172
Germany
Message 941463 - Posted: 19 Oct 2009, 23:52:25 UTC - in response to Message 941102.
Last modified: 20 Oct 2009, 0:24:01 UTC

I have one task pending since 30 july, task 1320233368. Is it normal?

Ricardo


You mean this WU: [http://setiathome.berkeley.edu/workunit.php?wuid=485578903]
Like you see, your first wingman didn't sent a result. He missed the deadline.
So a third WU was sent to your second wingman.

Uhh.. but he have: Average turnaround time 37.56 days
So maybe in a few days - if you have good luck - he will send a result and then if the result match with yours, then you will get your credits.

BTW.
For this kind of questions - the next time the NC subforum. ;-)

BTW.
Welcome in the SETI@home forum! :-)

____________
BR



>Das Deutsche Cafe. The German Cafe.<

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 941522 - Posted: 20 Oct 2009, 5:43:50 UTC - in response to Message 941110.


The version of the software to process a given work isn't supposed to be under the control of the user,
If you decided to run An Astropulse_v505 Wu with one of the Obsolete Astropulse or Astropulse_v5 apps, you'll find it just won't validate.
The switches are there to enable/disable what work is sent by the schedular to the Boinc client on your computer,
and it is the schedular that Brands the WU with what app it requires, so that Boinc can then run the correct app.

Claggy


Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because there isn't just one Astropulse application. There are three different ones.
____________

Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 941526 - Posted: 20 Oct 2009, 7:17:14 UTC - in response to Message 941522.



Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because there isn't just one Astropulse application. There are three different ones.


So what ? The purpose of the switches is to allow volunteers to select either Enhanced, Astropulse, or both because the two applications have different run times and other characteristics. Is there any reason for volunteers to be able to distinguish between the different versions ("applications" if you prefer) of Astropulse ?

The argument given in an earlier message that the output formats from the various Astropulse versions are different is irrelevant as BOINC will always pick the correct version to process a work-unit.
____________

Profile Geek@Play
Volunteer tester
Avatar
Send message
Joined: 31 Jul 01
Posts: 2463
Credit: 85,121,964
RAC: 12,158
United States
Message 941528 - Posted: 20 Oct 2009, 7:25:51 UTC
Last modified: 20 Oct 2009, 7:27:17 UTC

Newly created Astropulse work would be version 505.

At this time the only reason an older Astropulse version would be issued is if someone needs a wingman to process the work to get verification and credit. And of course that someone would need the proper version app to properly crunch it.
____________
Boinc....Boinc....Boinc....Boinc....

Richard Haselgrove
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8377
Credit: 46,805,962
RAC: 23,998
United Kingdom
Message 941534 - Posted: 20 Oct 2009, 8:31:50 UTC

Some people might be willing to help out with a few resends of older work, to help clear the almost-complete run out of the database, but prefer to devote the major part of their CPU donation to Multibeam or another project entirely.

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 941606 - Posted: 20 Oct 2009, 22:43:33 UTC - in response to Message 941526.



Which brings me back to my original point: why not have just ONE Astropulse user switch ?

Because there isn't just one Astropulse application. There are three different ones.


So what ?

BOINC doesn't know that Astropulse, Astropulse V5 and Astropulse V505 are in any way related.

There is no way to express the concept.

So, if the demand is that the switches be "ganged" you either introduce some odd bugs (Astropulse "on" doesn't include new Astropulse versions, but turn it off and back on and it does), or it requires a program to go through and "flip" all of the switches as applications change (expensive in server time).

... or it requires a major restructuring of BOINC.

... or it requires customizing BOINC, which makes updates a lot harder in the future.

"So What?" in this case is a big chunk of unavailable developer time, so no matter how right you are that this is how it SHOULD work, the odds of it being changed in the next few weeks are vanishingly small.

.... no matter how much you want to argue that it SHOULD work that way.

In my opinion, the best solution is to just take all the switches out and not worry about it.

If someone really needs that functionality, they can handle it through APP_INFO.XML, client side.

____________

Profile 52 Aces
Avatar
Send message
Joined: 7 Jan 02
Posts: 497
Credit: 12,918,406
RAC: 6,038
United States
Message 941621 - Posted: 21 Oct 2009, 0:28:06 UTC - in response to Message 941606.
Last modified: 21 Oct 2009, 0:29:12 UTC

Would be nice if they had a process back then to aggressively End-Of-Life any legacy components (in this case, Astropulse & Astropulse v5) so this confussion is not subjected upon the mass of generalists.

How to do that? Have a dedicated mid-tear crunch box at Berkeley running the older version and fork all inventory & all future re-issues to that box. Anyone already processing an AP will finish up fine, anyone who has yet to begin but upgrades somehow will have that AP instantly fail and tossed back into the re-issue pile. Meanwhile the dedicated box at Berkeley will finish up existing and wingman reporting and all re-issues for the next 30 to 90 days. If any of it's own AP WU's time out, they get re-issued back to itself. The box need not be at Berkeley either, it could be a willing volunteer tester, etc.

Of course there is no doubt more complexity to it than outlined, but the primary take away is never subject avg users to a hierarchy of obtuse switches, and I understand it's because that's what was permissable with the Boinc framework to begin with.

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 941640 - Posted: 21 Oct 2009, 2:40:31 UTC - in response to Message 941621.
Last modified: 21 Oct 2009, 2:42:15 UTC

Of course there is no doubt more complexity to it than outlined, but the primary take away is never subject avg users to a hierarchy of obtuse switches, and I understand it's because that's what was permissable with the Boinc framework to begin with.

Part of the nature of software development (I'm just shy of 40 years experience, which is a scary thought) is that the developers do their best to dream up something that works, and works well, and then they turn it over to users, and hope it does well.

In this particular case, it's BOINC, and the users are projects like SETI@Home, Einstein, LHC.

As I understand it, the work going out to Astropulse is the same for all three versions. The difference is in the result files, and that means different validators.

They could have written a super-validator that recognized the difference, but they'd also have to make sure that work originally crunched by Astropulse V5 is always crunched by V5, or someone gets cheated. That's a lot harder. The easy way is to make each a completely separate application.

So, now we have four different applications "in play" at the moment, and BOINC has to keep them all separate. No problem. It does that.

Even that is only an issue internally if the project doesn't add the ability to select and deselect the kind of work that they want to do.

It would have been much easier (in hindsight) to just not put the switches at all.

The fix is to remove the visible Astropulse and Astropulse V5 switches and add an Astropulse v505 switch, and write a program that looks at the two existing switches and sets the new 505 switch to match, then add the 505 choice.

Those crunchers who used to do older AP will still be available, along with everyone who said they don't care.

To have a dedicated "mid-tier" cruncher would mean going through the database (again) and turning off AP and APv5 for every machine except the dedicated cruncher, and that's the same kind of PITA that they'd run into if they have to implement Mr. Gadfly's imperative.

Ultimately, the question becomes: is it more important to fix that switch or is it better to get the radar blanking working and tested?

I know how I'd fix it: in the veterinary sense.
____________

Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 941707 - Posted: 21 Oct 2009, 8:48:06 UTC - in response to Message 941640.

Well my options are to accept any type of unit S@H sends. As long as BOINC is capable of selecting the correct software to crunch a unit, and doesn't use APv505 to crunch an APv5 or an APvprevious, for example, my computer's efforts will not have been in vain.

____________

1mp0£173
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 941829 - Posted: 21 Oct 2009, 22:27:54 UTC - in response to Message 941707.

Well my options are to accept any type of unit S@H sends. As long as BOINC is capable of selecting the correct software to crunch a unit, and doesn't use APv505 to crunch an APv5 or an APvprevious, for example, my computer's efforts will not have been in vain.

BOINC takes care of keeping the applications straight, so there is no danger of losing credit because BOINC used the wrong application.
____________

Profile KWSN THE Holy Hand Grenade!
Volunteer tester
Avatar
Send message
Joined: 20 Dec 05
Posts: 1902
Credit: 9,227,534
RAC: 13,551
United States
Message 943687 - Posted: 29 Oct 2009, 21:23:50 UTC - in response to Message 941829.

Well my options are to accept any type of unit S@H sends. As long as BOINC is capable of selecting the correct software to crunch a unit, and doesn't use APv505 to crunch an APv5 or an APvprevious, for example, my computer's efforts will not have been in vain.

BOINC takes care of keeping the applications straight, so there is no danger of losing credit because BOINC used the wrong application.



As I understand it, the three OUTGOING file formats are different, that is why there were two switches to start with!
____________
.

Norwich Gadfly
Avatar
Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 943737 - Posted: 30 Oct 2009, 0:17:21 UTC - in response to Message 943687.

Well my options are to accept any type of unit S@H sends. As long as BOINC is capable of selecting the correct software to crunch a unit, and doesn't use APv505 to crunch an APv5 or an APvprevious, for example, my computer's efforts will not have been in vain.

BOINC takes care of keeping the applications straight, so there is no danger of losing credit because BOINC used the wrong application.



As I understand it, the three OUTGOING file formats are different, that is why there were two switches to start with!


Sorry I cannot follow that logic. Whatever type of work unit the server sends to be crunched, it has to be able to accept the file format produced by the crunching software. This would be true if volunteers did not have switches to control what types of work-unit to crunch. So it would be possible to have one switch to control several work unit types.

____________

Previous · 1 · 2 · 3 · Next

Message boards : Technical News : Blip (Sep 16 2009)

Copyright © 2014 University of California