Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!cs.utexas.edu!uunet!mcsun!ukc!stl!stc!root44!jgh From: jgh@root.co.uk (Jeremy G Harris) Newsgroups: comp.lang.c Subject: Re: A solution to the multiple inclusion problem Summary: Integrated preprocessor not always faster Message-ID: <1037@root44.co.uk> Date: 27 Oct 89 10:25:28 GMT References: <14240@well.UUCP> <2185@dataio.Data-IO.COM> Reply-To: jgh@root44.UUCP (Jeremy G Harris) Distribution: comp Organization: UniSoft Ltd, London, England Lines: 16 In article <2185@dataio.Data-IO.COM> bright@dataio.Data-IO.COM (Walter Bright) writes: >Integrating the preprocessor into the parser is a big performance win >because you save a file write and read of the source. Except on a multiprocessor system. Assuming decent comms bandwidth between processors, it should make sense to split the compiler up fairly finely and pipeline the work. I know that more modern 'make' variants help somewhat, but I'd like to give those sixteen processors in our Sequent real work to do even when I'm only compiling one or two files. So, has anyone done this, and measured the results? We want information! -- Jeremy Harris jgh@root.co.uk