climateprediction.net home page
i586-pc-linux-gnu

i586-pc-linux-gnu

Questions and Answers : Unix/Linux : i586-pc-linux-gnu
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user19294

Send message
Joined: 19 Sep 04
Posts: 4
Credit: 0
RAC: 0
Message 4428 - Posted: 21 Sep 2004, 13:39:39 UTC

I can\'t get this client, there is only i686-pc-linux-gnu. Why?
ID: 4428 · Report as offensive     Reply Quote
old_user1
Avatar

Send message
Joined: 5 Aug 04
Posts: 907
Credit: 299,864
RAC: 0
Message 4429 - Posted: 21 Sep 2004, 13:48:12 UTC

I dunno, that's what BOINC has always "called" their Linux x86 version, I guess it's the "autoconf name." Can you not run the i686 version? I can't imagine a 586 would run CPDN so well? Even on a Pentium III/500MHz you're looking at probably 6-9 months to finish a model.

ID: 4429 · Report as offensive     Reply Quote
old_user19294

Send message
Joined: 19 Sep 04
Posts: 4
Credit: 0
RAC: 0
Message 4434 - Posted: 21 Sep 2004, 15:38:20 UTC - in response to Message 4429.  

> I dunno, that's what BOINC has always "called" their Linux x86 version, I

BOINC works on a i586, I compile it from daily cvs.
It's the CP client that doesn't support i586, this is the message:

[climateprediction.net] Message from server: platform 'i586-pc-linux-gnu' not found

It's the same on my Alphastation:

[climateprediction.net] Message from server: platform 'alpha-unknown-linux-gnu' not found

> guess it's the "autoconf name." Can you not run the i686 version? I can't

On these two computer no :(

> imagine a 586 would run CPDN so well? Even on a Pentium III/500MHz you're
> looking at probably 6-9 months to finish a model.

This is a good point.

Thanks
Matteo
ID: 4434 · Report as offensive     Reply Quote
old_user1
Avatar

Send message
Joined: 5 Aug 04
Posts: 907
Credit: 299,864
RAC: 0
Message 4439 - Posted: 21 Sep 2004, 16:54:44 UTC - in response to Message 4434.  

the "official" BOINC (well SETI/BOINC) releases are Win32/Mac/Linux x686/Solaris so that's what we were shooting for (although we failed on Win98/ME and have no access to Solaris). It's probably best to have x586's do SETI/LHC/Predictor as it is probably risky and too long to finish a run for CPDN (probably 1-1.5 years?)


ID: 4439 · Report as offensive     Reply Quote
Profile tullio

Send message
Joined: 6 Aug 04
Posts: 264
Credit: 965,476
RAC: 0
Message 4484 - Posted: 22 Sep 2004, 13:54:27 UTC - in response to Message 4434.  

> > I dunno, that's what BOINC has always "called" their Linux x86 version,
> I
>
> BOINC works on a i586, I compile it from daily cvs.
> It's the CP client that doesn't support i586, this is the message:
>
> [climateprediction.net] Message from server: platform 'i586-pc-linux-gnu' not
> found
>
> It's the same on my Alphastation:
>
> [climateprediction.net] Message from server: platform
> 'alpha-unknown-linux-gnu' not found
>
> > guess it's the "autoconf name." Can you not run the i686 version? I
> can't
>
> On these two computer no :(
>
> > imagine a 586 would run CPDN so well? Even on a Pentium III/500MHz
> you're
> > looking at probably 6-9 months to finish a model.
>
> This is a good point.
>
> Thanks
> Matteo
>
>
My Pentium II at 400 MHz ahould finish a model in 155 days (AVG 17 s/step). What gets me is that running in parallel with BOINC seti@home AVG does not increase.


ID: 4484 · Report as offensive     Reply Quote
Profile tullio

Send message
Joined: 6 Aug 04
Posts: 264
Credit: 965,476
RAC: 0
Message 4485 - Posted: 22 Sep 2004, 13:56:03 UTC - in response to Message 4434.  

> > I dunno, that's what BOINC has always "called" their Linux x86 version,
> I
>
> BOINC works on a i586, I compile it from daily cvs.
> It's the CP client that doesn't support i586, this is the message:
>
> [climateprediction.net] Message from server: platform 'i586-pc-linux-gnu' not
> found
>
> It's the same on my Alphastation:
>
> [climateprediction.net] Message from server: platform
> 'alpha-unknown-linux-gnu' not found
>
> > guess it's the "autoconf name." Can you not run the i686 version? I
> can't
>
> On these two computer no :(
>
> > imagine a 586 would run CPDN so well? Even on a Pentium III/500MHz
> you're
> > looking at probably 6-9 months to finish a model.
>
> This is a good point.
>
> Thanks
> Matteo
>
>
My Pentium II at 400 MHz should finish a model in 155 days (AVG 17 s/step). What gets me is that running in parallel with BOINC seti@home AVG does not increase.


ID: 4485 · Report as offensive     Reply Quote
old_user1
Avatar

Send message
Joined: 5 Aug 04
Posts: 907
Credit: 299,864
RAC: 0
Message 4486 - Posted: 22 Sep 2004, 14:41:13 UTC - in response to Message 4485.  

> My Pentium II at 400 MHz should finish a model in 155 days (AVG 17 s/step).
> What gets me is that running in parallel with BOINC seti@home AVG does not
> increase.

well it does, what BOINC is measuring and reporting is CPU time, so your model will take 155 CPU-days (on your computer), but in "wall-clock-time" that could be 3 years if you have a lot of other apps running in parallel!
ID: 4486 · Report as offensive     Reply Quote
Profile tullio

Send message
Joined: 6 Aug 04
Posts: 264
Credit: 965,476
RAC: 0
Message 4509 - Posted: 23 Sep 2004, 4:45:37 UTC - in response to Message 4486.  

> > My Pentium II at 400 MHz should finish a model in 155 days (AVG 17
> s/step).
> > What gets me is that running in parallel with BOINC seti@home AVG does
> not
> > increase.
>
> well it does, what BOINC is measuring and reporting is CPU time, so your model
> will take 155 CPU-days (on your computer), but in "wall-clock-time" that could
> be 3 years if you have a lot of other apps running in parallel!
>
>
You're right, of course. I have restarted seti@home to have some results in reasonable times, and CP will run to fill the intervals when seti@home work units are not available (very often). Thanks.
ID: 4509 · Report as offensive     Reply Quote
old_user3711

Send message
Joined: 30 Aug 04
Posts: 1
Credit: 320,622
RAC: 0
Message 4588 - Posted: 24 Sep 2004, 8:04:48 UTC - in response to Message 4486.  

> > My Pentium II at 400 MHz should finish a model in 155 days (AVG 17
> s/step).
> > What gets me is that running in parallel with BOINC seti@home AVG does
> not
> > increase.
>
> well it does, what BOINC is measuring and reporting is CPU time, so your model
> will take 155 CPU-days (on your computer), but in "wall-clock-time" that could
> be 3 years if you have a lot of other apps running in parallel!
>
>


g'day carl,

i'm new to CPDN and got BOINC i686-pc-linux-gnu running fine on an Athlon 1800+ without a hitch.

I also have a number of older PCs lying around that i was planning to leave on 24x7x365 _solely_ running CPDN BOINC.

... performance-wise they are mostly in the league of a PII @ 400 MHz ...

... the Pentium II & III are in the i686 family, but AMD chips with similar performance like the K6-2 and K6-3 are actually more like i586s ( or so i have read! ) .. so they wont run i686 binaries reliably ...

... on the _assumption_ that this vintage of machine (running the CPDN client exclusively) could complete a model in less than 6 months, i want to ask if i586 binaries the CPDN clients (hadXXXs) could be made available ...

what do you think ... worth the effort ???

thanks,
terse.
ID: 4588 · Report as offensive     Reply Quote

Questions and Answers : Unix/Linux : i586-pc-linux-gnu

©2024 climateprediction.net