Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!caen!spool.mu.edu!munnari.oz.au!bruce!goanna!minyos.xx.rmit.oz.au!godzilla!mg
From: mg@godzilla.cgl.rmit.oz.au (Mike Gigante)
Newsgroups: comp.sys.sgi
Subject: Re: Does anyone have any experience with 'cord'?
Message-ID: <mg.677721662@godzilla>
Date: 24 Jun 91 00:01:02 GMT
Article-I.D.: godzilla.mg.677721662
References: <1991Jun22.204610.5025@leland.Stanford.EDU>
Sender: usenet@minyos.xx.rmit.oz.au (Njuiz noveles nova newes)
Organization: RMIT Computer Centre, Melbourne Australia.
Lines: 29
Originator: mg@

dhinds@elaine18.Stanford.EDU (David Hinds) writes:

]I've been playing around with the cache optimizing procedure rearranging
]program 'cord', and have had lots of problems.  I have yet to get it to
]terminate successfully.  Firstly, it doesn't seem to work for me when I
]link a program with the multi-threaded C library.  It just core dumps
]with a segmentation violation.  When I re-linked the same program with
]the regular library (the program is set up so it can run single-threaded),
]I get lots of cord warnings about 'procedure(s) not found' naming things
]in the standard library - several hundred, actually.  Then I get a fatal
]error: 'cord: error: cannot write proc bxform'.  This happens at the stage
]when cord is trying to 'relocate section .text'.  This procedure isn't
]anything special, as far as I can tell.  This is all with the Irix 3.3.1
]stuff.

] -David Hinds
]  dhinds@cb-iris.stanford.edu

I had exactly the same experience under 3.3.2, I haven't tried it with
3.3.3 or with the 4.0 alpha

This was a multi-threaded program that used the m_* routines in the
task library.

I abandonded (reluctantly) the attempted use of cord after a frustrating
couple of days.

Mike Gigante
mg@godzilla.cgl.rmit.oz.au