Michael J. Smith (smith++at++vsl.ist.ucf.edu)
Wed, 13 Sep 1995 16:14:28 -0400 (EDT)
> On Tue, 12 Sep 1995, Anita Kishore wrote:
>
> > Is it possible to have any type of conflict between the shared arena of
> > Performer and other shared arenas obtained by "shmat" and "shmget" inside
> > a Performer application? There seems to be a time lag between when the data
> > is written in (non-performer) shared arena and reading from it. The reading
> > is done inside Performer.
>
> There certainly is! For starters "shmat" and "shmget" are not IRIX arena
> IPC calls, but System V IPC calls... The two IPC paradigm's don't co-exist
> that happily. In addition the System V IPC calls are significantly slower
> and don't seem to be MP safe. Unless you really need to use Sys V IPC, I'd
> advise using IRIS arena memory with Performer applications, it's a much
> happier solution. You can use arena's either directly through IRIX (see the
> manpages for usinit et al) or through Performer - pfDataPools etc...
>
Well, I'm not sure why you say that IPC calls don't co-exist happily with
IRIX REACT semaphores and shared memory. This just doesn't follow what
system I have running. For my underlying drivers that I have ported to
about 4 different UNIX platforms, I use System V IPC semaphores and
shared memory segments and have had absolutely no problems using it under
a performer 1.2 application. The speed hit may be true, but since the
SYS V semaphores are portable its just a matter of what you are after on
your product and what your programmers are familiar with. I have to,
however, agree that the IRIX REACT seamphores and shared memory segments
are faster, and a preferred choice on SGI platforms -- its just I hate to
write code that is completely unusable on other platforms unless I have a
resource bottleneck by that choice.
-----------------------------------------------------------------------------
| Michael J. Smith University Of Central Florida |
| Visual Systems Laboratory Institute for Simulation & Training |
| Graduate Research Assistant 3280 Progress Drive |
| smith++at++vsl.ist.ucf.edu Orlando, FL 32826-0544 |
| ++at++cs.ucf.edu |
| Codesmith for hire |
-----------------------------------------------------------------------------
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:51:52 PDT