[comp.compilers] C Compiler Metrics

stevel@qtc.uucp (Steve Lehto) (09/28/89)

As part of an internal project here; I am investigating the importance
of compiler speeds.  Specifically, how fast must a compiler compile in
order to be useful?  Is there a "limit of pain", below which the tool is
considered to be unuseable?  We're most interested in C compilers, but
data about any compiler would be useful.  I would like to collect the
responses and complete a chart which looks something like this:

Language	Compiler	Target	Host	Speed (lines/min)
--------	--------	------	----	----------


If responses are plentiful and meaningful, I'ld be happy to repost the
results for all to use.

Thanks in advance for any and all help.

steve lehto, QTC
[From Steve Lehto <stevel@qtc.uucp>]
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{spdcc | ima | lotus}!esegue.  Meta-mail to compilers-request@esegue.
Please send responses to the author of the message, not the poster.

shankar@hpclscu.HP.COM (Shankar Unni) (09/30/89)

> Language	Compiler	Target	Host	Speed (lines/min)

Compilation speeds can vary enormously (factor of 2 or 3) depending on the
composition of the benchmarks used and the options used. Please insist on a
standard benchmark and a reasonably standard set of options, to make any
meaningful comparison.
-----
Shankar Unni                                   E-Mail: 
Hewlett-Packard California Language Lab.     Internet: shankar@hpda.hp.com
Phone : (408) 447-5797                           UUCP: ...!hplabs!hpda!shankar
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{spdcc | ima | lotus}!esegue.  Meta-mail to compilers-request@esegue.
Please send responses to the author of the message, not the poster.

jesup@cbmvax.commodore.com (Randell Jesup) (09/30/89)

In article <1989Sep27.190232.8318@esegue.segue.boston.ma.us> Steve Lehto <stevel@qtc.uucp> writes:
>  Specifically, how fast must a compiler compile in
>order to be useful?  Is there a "limit of pain", below which the tool is
>considered to be unuseable? 

	How does this give you "the limit of pain"?  It merely tells you
the performance of some compilers.

>Language	Compiler	Target	Host	Speed (lines/min)
>--------	--------	------	----	----------

	An important point you left out: what sort of code?  I have a
compiler that varies by up to 10x in lines/min depending on the source put
into it (it has a global optimizer - _big_ routines annoy it.)
[From jesup@cbmvax.commodore.com (Randell Jesup)]
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{spdcc | ima | lotus}!esegue.  Meta-mail to compilers-request@esegue.
Please send responses to the author of the message, not the poster.