Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site ritcv.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxj!mhuxn!houxm!ihnp4!zehntel!hplabs!hao!seismo!rochester!ritcv!mjl From: mjl@ritcv.UUCP (Mike Lutz) Newsgroups: net.lang.c Subject: Re: C stack frame sizes Message-ID: <1397@ritcv.UUCP> Date: Mon, 3-Dec-84 19:31:01 EST Article-I.D.: ritcv.1397 Posted: Mon Dec 3 19:31:01 1984 Date-Received: Thu, 6-Dec-84 05:45:35 EST References: <2400@pur-ee.UUCP> Organization: Rochester Institute of Technology, Rochester, NY Lines: 17 >From: ecn-ee!malcolm Nov 30 21:27:00 1984 >I commonly put up to a megabyte into a single stack frame...I find >it is much more elegant to put all of my temporary arrays on >the stack...I wonder what this type of programming style would do to a >Berkeley style RISC machine? Probably wouldn't have much effect, as the philosophy of RISC is to hold scalars in the register bank. However, it does point up an interesting problem -- does the Berkeley RISC require a parallel stack, maintained by software protocols, to hold structures, arrays, and other humongous local variables? -- Mike Lutz Rochester Institute of Technology, Rochester NY UUCP: {allegra,seismo}!rochester!ritcv!mjl ARPA: ritcv!mjl@Rochester.ARPA