I am running SETI on a number of Linux boxes. I had ver. 3.18 installed on some and recently had upgraded them to 3.20. I have noticed that some of them that have been able to "get through" have downloaded ver 3.08. That was "SETI Classic", right?

Sample lines:

2004-08-10 01:39:47 [SETI@home] Requesting 85286 seconds of work
2004-08-10 01:39:47 [SETI@home] Sending request to scheduler: http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi
2004-08-10 01:40:22 [SETI@home] Scheduler RPC to http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi succeeded
2004-08-10 01:40:22 [SETI@home] Started download of setiathome_3.08_i686-pc-linux-gnu
2004-08-10 01:40:23 [SETI@home] Started download of 13ja04aa.29764.19152.497156.205
2004-08-10 01:40:31 [SETI@home] Finished download of 13ja04aa.29764.19152.497156.205

Any thoiughts?

> I am running SETI on a number of Linux boxes. I had ver. 3.18 installed on
> some and recently had upgraded them to 3.20. I have noticed that some of them
> that have been able to "get through" have downloaded ver 3.08. That was "SETI
> Classic", right?
>
> Sample lines:
>
> 2004-08-10 01:39:47 [SETI@home] Requesting 85286 seconds of work
> 2004-08-10 01:39:47 [SETI@home] Sending request to scheduler:
> http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi
> 2004-08-10 01:40:22 [SETI@home] Scheduler RPC to
> http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi succeeded
> 2004-08-10 01:40:22 [SETI@home] Started download of
> setiathome_3.08_i686-pc-linux-gnu
> 2004-08-10 01:40:23 [SETI@home] Started download of
> 13ja04aa.29764.19152.497156.205
> 2004-08-10 01:40:31 [SETI@home] Finished download of
> 13ja04aa.29764.19152.497156.205
>
> Any thoiughts?
>
>

3.08 is the client version of the seti@home project.
Naturally it's same like in seti classic, because the
algorithm for scanning the data is still the same.

Versions 3.19 and 3.20 are the version of boinc.
So the client project Seti@home 3.08 is running in the
'mother' application boinc 3.19 oir 3.20.

Boinc 3.18 you shouldn't use (last beta version ?). It's out of time.
3.20 and 3.19 are nearly the same with few changes.
In the near future version 4.03 will go public (right now in climateprediction in use and) for seti@home.

Then boinc 4.03 will host the seti@home 3.08 client.

Hope I could point the difference.

Actually, it's just a coinsidence the current seti-client is v3.08. ;)

And even CPDN is currently using v4.03, it's not certain seti will use this, but maybe a later version. Also, applications like seti must have the same major-version-number, so then the upgrade to v4 happens, it will be seti-v4.02 or later. ;)

> Also, applications like seti must have the same
> major-version-number, so then the upgrade to v4 happens, it will be seti-v4.02
> or later. ;)

I don't think the major version numbers have to be the same.

If you self-compile, the current version of seti is 4.0 and it runs in boinc 3.2. I've been running this combination for several weeks now.

> 3.08 is the client version of the seti@home project.
> Naturally it's same like in seti classic, because the
> algorithm for scanning the data is still the same.
>
> Versions 3.19 and 3.20 are the version of boinc.
> So the client project Seti@home 3.08 is running in the
> 'mother' application boinc 3.19 oir 3.20.
>
> Boinc 3.18 you shouldn't use (last beta version ?). It's out of time.
> 3.20 and 3.19 are nearly the same with few changes.
> In the near future version 4.03 will go public (right now in climateprediction
> in use and) for seti@home.
>
> Then boinc 4.03 will host the seti@home 3.08 client.

And people wondered why I said that calling it a client was a bad thing to do ...

Even though the term client is deprecieated it is still haunting us ...

I still think my alternatives were better ...

> > 3.08 is the client version of the seti@home project.
> > Naturally it's same like in seti classic, because the
> > algorithm for scanning the data is still the same.
> >
> > Versions 3.19 and 3.20 are the version of boinc.
> > So the client project Seti@home 3.08 is running in the
> > 'mother' application boinc 3.19 oir 3.20.
> >
> > Boinc 3.18 you shouldn't use (last beta version ?). It's out of time.
> > 3.20 and 3.19 are nearly the same with few changes.
> > In the near future version 4.03 will go public (right now in
> climateprediction
> > in use and) for seti@home.
> >
> > Then boinc 4.03 will host the seti@home 3.08 client.
>
> And people wondered why I said that calling it a client was a bad thing to do
> ...
>
> Even though the term client is deprecieated it is still haunting us ...
>
> I still think my alternatives were better ...
>
> [url=http://homepage.mac.com/pauldbuck/index.html]
>
>
>


Beat me to it Paul lol

WE have at least 2 version numbers for each project...

SETI@H uses BOINC ver 3.20.
Cruncher ver 3.08.


Climate P@H uses BOINC ver 4.03
Cruncher ver 2.2.29


Predictor@H uses BOINC ver 3.20
Cruncher ver 3.13

Pirates@H uses BOINC ver 3.20
Cruncher unknown

As far as I know you will need to have the same BOINC ver number on all projects if you what to run them all, which is a manual install.

The cruncher program (s) will be installed automatically on DL of the WU...

This 'SPACE' for Rent.

M7

>
> Then boinc 4.03 will host the seti@home 3.08 client.
>
> Hope I could point the difference.
>
Actually the fact that the version numbers between the S@H BOINC and S@H classic crunch programs are the same is a pure coincidence. When S@H moves to BOINC 4.x, you can expect that the S@H version will be 4.y. The same is true of the two predictor crunch programs.

Major version changes change the communication between some of the components in such a way that they are not backward compatible. When S@H goes to 4.x, everyone will have to switch to get new work.

> >
> > Any thoiughts?
> >
> >
>
> 3.08 is the client version of the seti@home project.
> Naturally it's same like in seti classic, because the
> algorithm for scanning the data is still the same.
>
> Versions 3.19 and 3.20 are the version of boinc.
> So the client project Seti@home 3.08 is running in the
> 'mother' application boinc 3.19 oir 3.20.
>
> Boinc 3.18 you shouldn't use (last beta version ?). It's out of time.
> 3.20 and 3.19 are nearly the same with few changes.
> In the near future version 4.03 will go public (right now in climateprediction
> in use and) for seti@home.
>
> Then boinc 4.03 will host the seti@home 3.08 client.
>
> Hope I could point the difference.
>
>
>
Ok. Just color my original post "confused due to lack of sleep." I had noticed that originally (difference between BOINC version and SETI version) but after reinstalling one system completely, due to HD errors, I thought something was odd after reinstalling BOINC. Thanks for pointing it out, though.

*move along. there's nothing to see here*

>
> Climate P@H uses BOINC ver 4.03
> Cruncher ver 2.2.29
>
>

If you means Climateprediction.net, they're "officially" using v4.02 for both core & application, but there seems to have been some confusing then building core-client for windows so it report itself as v4.03. The windows-application being downloaded is very much v4.02 atleast.


As for having the same major-version-number, "When a project makes a major-version change in its server software, it may need to create new versions of its applications. It must invalidate (by incrementing min_version) all app versions that are incompatible with the new server software." So it seems it's possible to continue using old applications, but in most cases it's not just some client-changes that's made so must upgrade the application in any case, and upgrading this with a lower major-version-number doesn't make much sence.

Atleast for the v4-upgrade, appart for better multi-project-support, trickle-fixes & other bug-fixes, the main new feature is the applications checking for the core-client is running.
AFAIK for this reason seti-v4 shouldn't be compatible with old v3.20-core, since atleast in my understanding this core doesn't send the "hello I'm still running"-signal so the seti-application should terminate in under a minute if there's not a bug here, or the self-compiler has changed something... Of course, being self-compiled it's no problem changing the version# on the core-client before compiling it, so is actually running v4 masquerading as v3.20 ;)

> AFAIK for this reason seti-v4 shouldn't be compatible with old v3.20-core,
> since atleast in my understanding this core doesn't send the "hello I'm still
> running"-signal so the seti-application should terminate in under a minute if
> there's not a bug here, or the self-compiler has changed something... Of
> course, being self-compiled it's no problem changing the version# on the
> core-client before compiling it, so is actually running v4 masquerading as
> v3.20 ;)

I've been using self-compiled seti v4 since 22 July with self-compiled boinc v3.20, and nothing was changed to make it "masquerade" as anything other than the actual version numbers it compiled with. On execution, it identifies with the correct version numbers - and obviously isn't incompatible.

Here's what my log looks like when I start it:

[darren@platinum boinc]$ ./boinc_client
2004-08-11 02:01:05 [---] Starting BOINC client version 3.20 for i686-pc-linux-gnu
2004-08-11 02:01:05 [SETI@home] Project prefs: using your defaults
2004-08-11 02:01:05 [Pirates@Home] Project prefs: using your defaults
2004-08-11 02:01:05 [SETI@home] Found app_info.xml; using anonymous platform
2004-08-11 02:01:05 [SETI@home] Host ID is 3994
2004-08-11 02:01:05 [Pirates@Home] Host ID is 319
2004-08-11 02:01:05 [---] General prefs: from SETI@home (last modified 2004-07-24 16:19:13)
2004-08-11 02:01:05 [---] General prefs: using your defaults
2004-08-11 02:01:05 [SETI@home] Resuming computation for result 10ja04ab.4477.21168.122158.208_4 using setiathome version 4.00

As you see, at the beginning it identifies the boinc client as version 3.20, and the last line identifies the seti version as 4.00. The same version numbers appear in my client_state.xml file.