[prev in list] [next in list] [prev in thread] [next in thread] 

List:       hpux-cxx-dev
Subject:    FW: CXX-DEV: limits on data size application size etc
From:       "Anglister, Shlomo" <shlomo.anglister () intel ! com>
Date:       2000-11-23 5:32:43
[Download RAW message or body]


Michael,

The problem is that we have thousands of HP workstations (!) It's obviously
a big challenge to recompile the kernel and populate it all over Intel.
I thought there might be some compilation/linker flags that can increase
those default sizes.

Thanks,

Shlomo

-----Original Message-----
From: Michael Ehrig [mailto:michael_ehrig@hp.com]
Sent: Wednesday, November 22, 2000 4:44 PM
To: Anglister, Shlomo
Cc: cxx-dev@cxx.cup.hp.com
Subject: Re: CXX-DEV: limits on data size application size etc



Shlomo,

Currently you have to rebuild the kernel if you change these parameters. Sam
is
probably the most convenient way to accomplish this. The parameters in
question
are maxdsiz, maxdsiz_64bit, maxssiz, maxssiz_64bit, and maybe the shmmax and
maxtsiz. An overview of parameter descritions is available under:

http://docs.hp.com/hpux/onlinedocs/os/KCparams.OverviewAll.html

Naturally you must have root permissions to rebuild the kernel and reboot
the
system.

Michael

"Anglister, Shlomo" wrote:
> 
> Hi,
> 
> Where are the limits for max data size and max application size defined?
> (How) can I control them (without compiling the kernel)?
> I already looked at /usr/inlude/sys/params.h
> 
> Thanks,
> 
> Shlomo
> 
> -----Original Message-----
> From: Dennis Handly [mailto:dhandly@cup.hp.com]
> Sent: Wednesday, November 22, 2000 1:13 PM
> To: cxx-dev@cxx.cup.hp.com
> Subject: Re: AW: CXX-DEV: Coroutines versus pthreads with PROCESS_SCOPE
> 
> >during stack unwinding or a catchhandler. The throw/unwind would
therefore
> >be continued after the other coroutine switches back. It would be
continued
> >on the same stack. So its stack is contiguous.
> 
> The stack from the throw, the catch and any destructors must all be
> contiguous and present, until the end of catch.
> 
> >The thrown object is copied into a thread local storage, which is
> overwritten
> >later on
> 
> The object is either copied into the local frame or to the heap.
> 
> >Is there any MxN support planned yet?
> 
> They are working on it.  Not sure when it will be released.
> 
> >If I create more than 256 SYSTEM_SCOPE threads, pthread_create returns 11
> jrg
> 
> You can configure the kernel to make this bigger.
>  _________________________________________________________________
>  To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
>     with the message UNSUBSCRIBE cxx-dev
>  _________________________________________________________________
> 
>  _________________________________________________________________
>  To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
>     with the message UNSUBSCRIBE cxx-dev
>  _________________________________________________________________

 _________________________________________________________________
 To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
    with the message UNSUBSCRIBE cxx-dev
 _________________________________________________________________

[prev in list] [next in list] [prev in thread] [next in thread]