Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!infonode!ingr!b11!dclark From: dclark@b11.ingr.com (Dave Clark) Newsgroups: comp.arch Subject: Re: Instruction Scheduling Keywords: SPEC compilers CDC Mips Message-ID: <1991Apr10.131341.26357@b11.ingr.com> Date: 10 Apr 91 13:13:41 GMT References: <32097@shamash.cdc.com> <1991Apr8.224717.14402@aero.org> Organization: Intergraph Corp. Huntsville, AL Lines: 38 jordan@aero.org (Larry M. Jordan) writes: >Where and when does instruction scheduling/reorganizing occur? Is >this function performed by the HOL compilers? (If so, the symbolic >assembly I'm reading obscures this fact. If not, then during >assembly?) Yes and no. Most HOL compilers do some sort of optimizing on the back-end. Although I haven't (yet) seen an assembler that does optimizations, I have seen utilities that can optimize assembly code (part of an HOL compiler back-end). One of the compilers I'm using actually does instruction reordering within the linker! >In a word or two, could someone summarize the pros/cons for either >approach (HOL or independent lang. independent pass)? HOL or high-level intermediate representation: Control structures are more easily recognizable. Restructuring is straightforward. Machine- specific optimizations are often omitted from this phase so that the bulk of the compiler can be easily ported. Assembly level: This is a good place to put machine-specific optimi- zations. Instruction timing and reordering is a simple search prob- lem, especially on the advanced RISC chips. >--Larry ---------------------------------------------------------------------------- Dave Clark | System Development | I among men bear the same wounded hand, Intergraph Corp., CR1102 | suffer the same reddened cup Huntsville, AL 35894-0001 | and live an identical rage. UUCP: uunet!ingr!b11!dclark | -- Pablo Neruda Internet: dclark@b11.ingr.com | ----------------------------------------------------------------------------