I've spent too much time caring about no compiler but the C compiler, so I've learned something. Apparently they've split compilers into front-ends (which parse languages), middleware (that does some of the optimization that isn't either language dependent or processor dependent) and a backend (which generates the machine code for the target processor). I had always thought of C as the target of the front end so I snickered at the acronym "Gnu Compiler Collection", but really I wasn't paying attention.
<div><span class="gmail_quote">On 3/16/07, <b class="gmail_sendername">Greg Lindahl</b> <<a href="mailto:firstname.lastname@example.org">email@example.com</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">On Thu, Mar 15, 2007 at 02:14:06PM -0500, Peter St. John wrote:<br><br>> I'd be optimistic for the process of translating F to C, followed by
<br>> compiling the C, to be as effecient as the process of compiling C into<br>> assembler, followed by assmbling into executable.<br><br>I'm not optimistic. The process destroys info. It's not that Fortran
<br>compilers are better at optimizing... most Fortran compilers are also<br>C compilers. It's that Fortran as a language is easier to optimize,<br>and Fortran programmers tend to write easier to optimize programs<br>
(more arrays, less pointers.)<br><br>When you translate to C, automatic or not, you usually lose both<br>properties.<br><br>-- greg<br><br>_______________________________________________<br>Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">
Beowulf@beowulf.org</a><br>To change your subscription (digest mode or unsubscribe) visit <a href="http://www.beowulf.org/mailman/listinfo/beowulf">http://www.beowulf.org/mailman/listinfo/beowulf</a><br></blockquote></div>