From owner-freebsd-hackers Wed Oct 15 00:54:47 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id AAA27031 for hackers-outgoing; Wed, 15 Oct 1997 00:54:47 -0700 (PDT) (envelope-from owner-freebsd-hackers) Received: from word.smith.net.au (vh1.gsoft.com.au [203.38.152.122]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id AAA27023 for ; Wed, 15 Oct 1997 00:54:41 -0700 (PDT) (envelope-from mike@word.smith.net.au) Received: from word.smith.net.au (localhost.gsoft.com.au [127.0.0.1]) by word.smith.net.au (8.8.7/8.8.5) with ESMTP id RAA02149 for ; Wed, 15 Oct 1997 17:21:47 +0930 (CST) Message-Id: <199710150751.RAA02149@word.smith.net.au> X-Mailer: exmh version 2.0zeta 7/24/97 To: hackers@freebsd.org Subject: Call for Fortran assistance. Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 15 Oct 1997 17:21:46 +0930 From: Mike Smith Sender: owner-freebsd-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Ok, if the 'F' word upsets you, you can ignore this thread with impunity. If you know something about fortran, a few brief words would be quite helpful. We're currently engaged in a joint venture with a customer to produce an all-sky meteor radar. Not what you'd call an earth-shaking device, but likely to be fairly unique and interesting to the relevant community. Unfortunately, the customer is providing some of the software for the system, and their language of preference is Fortran. We've seen their code before, and it's disgusting. Unfortunately, they have the algorithmic technology that we need, and they're not willing to document it well enough for us to reimplement in another language. (They have also stalled us up against a deadline, so we really don't have the time either.) In order for their software to interoperate with ours, we need to be able to pipe binary data from one of our processes to one of theirs, at a fairly convincing rate (about 1.5M/sec) on a Pentium-class processor. After I/O overheads we have about 1/2 of the processor left. Our initial design assumed that they would be able to read this on standard input; wrong. Our second design threw the data through a FIFO, but Fortran has funny formatting requirements for sequential-access data that have stymied us. We tried converting to ASCII and then parsing it back in, but that's too slow. We've offered to write a Fortran-callable C procedure that can manage reading the FIFO, but they don't use procedures much (!) and want an all-fortran solution. If you have ever done anything like this before, I'd *really* like to know how you did it, before we bludgeon the customer and commit to a total rewrite... Thanks, Mike