A few observations Re: Clean
Hassett
hassett@acm.org
Fri, 10 Apr 1998 19:25:05 -0500
On Wed 08 Apr, Adrian Hey (evidently not a fan of Ada) wrote
> If you want to see a really slow compiler I suggest
> you try using ADA (yuk!) if you haven't already.
I'm not an expert on the compilation of functional languages (though I
am a fan of functional programming), so I'm curious about compilation
speed.
Is a Clean compiler really much faster than an Ada compiler? The Clean
home page describes the speed of compilers as generally in the thousands
of lines per minute. Here are some Ada compilation speeds that have
been reported to me:
Source Size Optimization Minutes Compiler* Platform
----------- ------------ ------- ---------- ----------------------
20,000 none 8 gnat Pentium II 266MHz
20,000 full 17 gnat Pentium II 266MHz
33,742 full 35 gnat 3.10p "Heavily loaded" UltraSparc
24,297 full 8 AdaMagic SPARCstation 5
(Source sizes are counts of semicolons that are not in comments or
literals, a "traditional" measure of Ada source text that omits all
comments and blank lines, and generally corresponds to a count of the
number of statements, declarations, and similar constructs.)
These numbers range from about 1K to 3K "semicolons" per minute.
Given that compilation speed is subject to many variables (details of
the source text, method of measuring source size, platform speed, OS,
disk access time, system load, degree of optimization, available RAM,
etc.), these seem to be close to the range claimed for Clean compilers.
Does anyone some data points on Clean compilation speed to offer?
- Jim Hassett
* The Ada compilers used:
gnat - GNU Ada, a "free" (GNU GPL) compiler (www.gnat.com
AdaMagic - Intermetrics AdaMagic 2.787 (using gcc as "back end")