Path: utzoo!mnetor!uunet!husc6!cca!mirror!ima!johnl From: johnl@ima.UUCP Newsgroups: comp.compilers Subject: Re: Assemblers can be fast Message-ID: <822@ima.ISC.COM> Date: 22 Dec 87 20:46:35 GMT References: <813@ima.ISC.COM> Sender: johnl@ima.ISC.COM Reply-To: ucbvax.Berkeley.EDU!sun!wrs!dg (David Goodenough) Organization: Wind River Systems, Emeryville, CA Lines: 32 Approved: compilers@ima.UUCP In article <813@ima.ISC.COM> Green Eric Lee writes: >Still, on the great assembly-language vs. direct-to-object debate, I >still must conclude that emitting assembly language slows things down >a lot. This is especially true on microcomputers, with their slow disk >drives, which is why most microcomputer compilers directly emit object >files. One solution that I have worked on is to emit a hybrid assembler / object file. For instance if the assember source was to be mov d7,-(a7) then the object for this can be generated by the compiler as a 16 bit word, whereas for something like bcc L42 the compiler emits the assembler as shown. Now you have the advantage of assembler / HLL mixing, and IF you want to mess with the assembler, a separate translator will convert from the hybrid language to full assembly in a twinkle; however you still keep the advantage of much less disk I/O: the 'mov' above is 14 bytes, against the 2 bytes for the assembled output - this probably means about a 60% saving in file being processed. -- dg@wrs.UUCP - David Goodenough -- Send compilers articles to ima!compilers or, in a pinch, to Levine@YALE.EDU Plausible paths are { ihnp4 | decvax | cbosgd | harvard | yale | bbn}!ima Please send responses to the originator of the message -- I cannot forward mail accidentally sent back to compilers. Meta-mail to ima!compilers-request