I am doing Greg's test on an SGI for Jeremy's network, and do not see any
stabilization in the memory. I even put the Option module in with the
cache attribute set to 0 as David suggested. As Jeremy noted previously,
the network is rather large with a couple of macros, including a loop with
Get/SetLocal. That is why I not tried to narrow the problem down to one
module.
Well I finally got the "Out of Memory" error running in script mode.
Can we now say that there is a memory leak?
I orginally tested on Windows 98, but after Suhaib's warning that memory
leaks are a known problem there, I ran the network on Windows NT and so
did Jeremy. And the memory leak occured there, too.
Jeff
On Fri, 5 Nov 1999 [EMAIL PROTECTED] wrote:
> The standard way to test for memory leaks in DX is this.
>
> 1. Prepare a network that you believe demonstrates a memory leak - the
> smaller the better.
>
> 2. Run DX with the flags "-script -readahead off -cache off" That'll
> bring up dx at a prompt.
>
> 3 type:
>
> include "foo.net"
> Executive("flush cache");
> Usage("memory", 0);
>
> where foo.net is the name of your test script. It'll run the network,
> flush out everything thats in the cache and dictionary, and tell you how
> much memory remains allocated.
>
> 4. go to 3 - eg. do it a few times. Memory usage *should* stabilize. If
> it rises after the second or third time, there's a leak.
>
> Greg
>
> (Embedded
> image moved to "Suhaib Siddiqi" <[EMAIL PROTECTED]>
> file: 11/05/99 11:24 AM
> pic09223.pcx)
>
>
>
>
> Please respond to [email protected]
>
>
> To: [email protected]
> cc:
> Subject: RE: [opendx-dev] Memory Leak clarifications
>
>
>
> > In response to Suhaib's post:
> > > The OpenDx/Cygwin version Windows 98? That is a known
> > problem. You
> > > have right to yell at M$ for selling the gloriously
> > shitty product.
> > > Win98 itself has memory leak, which gets horrible when an
> > > X-application (X-server) is running because Win98 tries
> > to be selfish
> > > and hat to share
> > > resources with any other server... With X-server trying to share
> > > Display causes Win98/95 to start eating CPU resources
> > and may once
> > > in a while will force you to reboot the PC..... because it gets
> > > horribly slow.
> > Actually, I was able to create the problem under NT as
> > well. The problem is
> > not with CPU resources, but memory usage. It seems like
> > OpenDx isn't releasing
> > memory somewhere, so eventually it fills up its memory
> > limit (-memory 200) and
> > won't release any of it. Nothing effects the system,
> > except the dx network
> > won't run until you disconnect/reconnect the dxexec to
> > force the release of
> > memory.
> >
> > Thanks for the responses! Let's keep at it so we can
> > figure out what's going
> > on here...
>
>
>
> Get a CPU resources monitor and try to watch resources consuption
> and Windows 98/95 not releasing resources to X applications.
> You will notice this problem with any X app, not just OpenDX.
> I have tested this over and over again during developemnt of several
> X clients. it is Win98/95. The M$ engnieers do not know
> how write an OS which could be stable and share resources properly.
>
> >
> > Jeremy Zoss
> > Southwest Research Institute
> > (210)522-3089
> > [EMAIL PROTECTED]
> >
>
>
>
>
----------------------------------------------------------------------
Jeff Braun Geophysics Dept.
mailto: [EMAIL PROTECTED] Montana Tech
(406) 496-4206 1300 W. Park St.
Butte, MT 59701
