Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watnot!watmath!clyde!cbatt!ihnp4!homxb!gemini From: gemini@homxb.UUCP Newsgroups: comp.lang.c Subject: Re: C compiler for Intel 8051 Message-ID: <199@homxb.UUCP> Date: Sun, 5-Apr-87 04:31:41 EST Article-I.D.: homxb.199 Posted: Sun Apr 5 04:31:41 1987 Date-Received: Sun, 5-Apr-87 22:38:49 EST References: <569@moogvax.UUCP> <1955@hoptoad.uucp> Distribution: na Organization: PC Research, Inc. Lines: 32 Keywords: Intel 8051 Summary: Only for 'toy' applications In article <1955@hoptoad.uucp>, farren@hoptoad.uucp (Mike Farren) writes: > In article <569@moogvax.UUCP> hank@moogvax.UUCP (Hank Kowal) writes: > > > >Does anyone have any experience and/or info on C compilers for the Intel 8051 > >microcontroller. Preferably it should be MS-DOS based and be compatible with > >Intel's ICE. > > Haven't seen one, and I don't believe that such a beastie exists, or should. But there are several, and you are right, they shouldn't exist. The Archimedes 8051 compiler comes closest to being C, but not full C. Unless your application doesn't care how slow or large the code is, you don't want to even think of using these compilers, at least not until they provide an option to define 'int' as 8 bits. It takes the compiler on the order of 20 instructions just to do: ++i; if you optimize for speed (hah!). Optimizing for space turns the 20 instructions into about 10, with several function calls! Everything gets done in 16 bit arithmetic, which the 8051 doesn't have. None of them would compile the Dhrystone benchmark. One of the ones we tried couldn't hack that much 'C', and one couldn't tolerate more than a handful of locals (guess where locals are stored in the 8051!). If you are still interested, I could get you names, numbers, and prices. Send mail to: ihnp4!polux!rer which is where I'll be when I have easy access to the info. Rick Richardson, PC Research, Inc: (201) 922-1134 ..!ihnp4!castor!pcrat!rick when at AT&T-CPL: (201) 834-1378 ..!ihnp4!castor!polux!rer