Path: utzoo!attcan!uunet!mcvax!ukc!stl!stc!idec!prlhp1!paszkows
From: paszkows@prlhp1.prl.philips.co.uk (paszkows)
Newsgroups: comp.sys.transputer
Subject: Increasing Transputer Efficiency
Message-ID: <671@prlhp1.prl.philips.co.uk>
Date: 31 Oct 88 11:52:14 GMT
Reply-To: paszkows@prlhp1.UUCP ()
Organization: Philips Research Laboratories, Redhill, UK
Lines: 13

An article entitled "Parallel Processing with the DisPuter" by 
C. Philip Winder in the "Proceedings of the 8th occam User Group
Technical Meeting"   ISBN 90 5199 002 4 has a small section about increasing
Transputer efficiency.

In particular he mentions  as follows:

"The compiler produces code to check array bounds. This is necessary to conform
to the formal model and can be useful in finding bugs. However, if the checks 
are suppressed, instruction throughput can increase by up to 40 per cent."

My question is how do you do this? ( For information the software is written
using Inmos' TDS system ).