Date: Wed, 26 Feb 1997 10:37:03 +0800 (GMT)
From: Orlando Andico <orly AT gibson DOT eee DOT upd DOT edu DOT ph>
To: Geir Thomassen <geirt AT powertech DOT no>
cc: djgpp AT delorie DOT com
Subject: Re: gcc as a cross-compiler -- offtopic
In-Reply-To: <5erhjg$ep0$1@troll.powertech.no>
Message-ID: <Pine.SGI.3.93.970226103304.9213E-100000@gibson.eee.upd.edu.ph>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Mon, 24 Feb 1997, Geir Thomassen wrote:

> 
> Agree. I have tried the 8051 compiler from Dunfield. It is usable, 
> BUT it is not near ANSI, and generates about 20 % more code than
> the IAR 8051 compiler (www.iar.com). The IAR compiler is fairly
> expensive, and very buggy, at least v5.12. It is in my opinion
> not worth the money, because of the bugs.
> 
> A good compiler to port to other architectures should have extensive
> support for optimisation in the early stages of the compiler (low
> level opt. is of course architecture dependent), correct error
> messages (i.e. a clever parser), and be (near) ANSI.
> 

I'm using the GCC-68hc11 here in school for my microprocessor interfacing
subject. I've no idea how "good" it is because so far we've been doing
fairly simple stuff (i.e. blink-the-LED) and my hardware has 8K (!!) of
memory. Having to support the overhead of GCC is sucky though, the
compiler has to use memory as registers.. and the 68hc11 is "easier"  to
port to than the 8051 for example because it has two or three 16-bit
registers that can hold the result of a multiply..

.-----------------------------------------------------------------.
| Orlando Andico                email: orly AT gibson DOT eee DOT upd DOT edu DOT ph |
| IRC Lab/EE Dept/UP Diliman   http://gibson.eee.upd.edu.ph/~orly |
|  "through adventure we are not adventuresome" -- 10000 Maniacs  |
`-----------------------------------------------------------------'