Newsgroups: comp.parallel.pvm
From: John Biddiscombe <j.biddiscombe.yum.SPAM.yum.@rl.ac.uk>
Reply-To: j.biddiscombe.yum.SPAM.yum.@rl.ac.uk
Subject: Re: Odd Behaviour of PVM on Pentium II MMX machine
Organization: Rutherford Appleton Laboratory
Date: Fri, 21 Nov 1997 10:10:37 +0000
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <34755E1D.2714@rl.ac.uk>

> No I haven't. I can't see why it would make a difference, since I'd be
> using exactly the same compiler on the other machine, so the generated
> code would be identical. In the absence of any other ideas, I'll give it
> a try though.

Well I recompiled it, and guess what ? - I found that if you use byte
alignment of record structures. It works OK, but Word, DWord it doesn't.
Don't know why it only failed on one machine, but problem is fixed now.

Thanks for making me rethink my strategy.

PS : Anyone know if byte alignment is correct, couldn't find anything in
the makefiles to indicate which was expected.

ta taa

John B

--  
SPAM tastes yummy, but I've had too much lately.
Please note doctored reply-to address.

