RE: [OT] GCC vs Intel C++ compiler benchmark

["Scott Robert Ladd" <scott at coyotegulch dot com>]

I've been a professional compiler reviewer (Computer Language,
Dr. Dobb's, for starters), so I have a bit of experience in this
area -- and in my judgment, the Open Magazine article is badly
written, blatantly biased, and poorly researched.

I've been reviewing C++ compilers since the cfront days; you'll
find my two-month-old comparisons of Intel C++ and gcc at:

	http://www.coyotegulch.com/hpc/reviews.html

I haven't had time to do a really thorough analysis of both
compilers, but my initial impressions and benchmarks can be
found at the web site above. My conclusion: in most tests,
Intel's compiler produces faster code.

Does that mean gcc is dead?

NO. NO. NO.

Compared to the entire gcc package, Intel's compiler is a
singleton, a highly-focused tool that accomplishes good things
in a limited arena. That's it. Intel is not cross-platform, it
only supports two languages, and it makes no attempt to support
the current C99 standard. And, while the Intel compiler is
"free" to a limited extent, it is not a tool that most "free"
software advocates will use for their work (see my writings
about the Intel license.)

Clearly, the Open Magazine review is written by someone who sees
only Intel and Microsoft.
The facts and underlying assumptions are just plain wrong -- but
because this is in a "magazine", it will get far too much
attention from the scaremongers and sensationalists.

Now, in my work, Intel's compiler has replaced gcc for two
purposes. First, I need OpenMP support for a project, something
Intel has and gcc does not. Second, I recently had a
non-commercial project that required Fortran 95, something Intel
can provide that is (as yet) unavailable in gcc. The g95 project
may solve that problem, but it's inclusion in standard gcc
appears to be some time off.

I am OPEN TO SUGGESTIONS regarding my benchmarks. If I have
missed some magic gcc switch, let me know, please. I want to
produce accurate information that helps people make wise
choices.

Different tools for different purposes; gcc is in no danger of
being supplanted by Intel's compiler, but the gcc community does
need to recognize that the Intel compiler has some good points,
too. Articles like the one on Open Magazine, however, do nothing
to help anyone make wise choices about compilers.

--
Scott Robert Ladd
Master of Complexity, Destroyer of Order and Chaos
Visit Coyote Gulch, a Free Developer Resource at
http://www.coyotegulch.com


--
Scott Robert Ladd
Master of Complexity, Destroyer of Order and Chaos
Visit Coyote Gulch, a Free Developer Resource at
http://www.coyotegulch.com



> -----Original Message-----
> From: gcc-owner@gcc.gnu.org
> [mailto:gcc-owner@gcc.gnu.org]On Behalf Of
> Claus Fischer
> Sent: Sunday, January 27, 2002 06:48
> To: gcc@gcc.gnu.org
> Subject: [OT] GCC vs Intel C++ compiler benchmark
>
>
>
> The rather biased article on
> http://www.open-mag.com/754088105111.htm
> compares Intel's C++ favourably to GCC, without
> disclosing many
> enlightening details.
>
> I'm not concerned with that but with the notable OS difference
> between Linux (SuSE 7.3) and Windows (XP Pro). IMHO a
> CPU bound
> benchmark should see less than 1 % influence from OS
> and C library.
> The graphics shows roughly 7% better performance on
> Windows for the
> same (Intel) compiler.
>
> Are there any ABI differences which would justify
> such a difference?
> Register use conventions? Other stuff that you could identify?
>
> Unfortunately the benchmark gives no details about
> compiler options,
> so presumably the answer is `the benchmark is flawed'
> and `go ask
> them'. In this case don't bother to answer.
>
> Claus
>
> --
> Claus Fischer <claus.fischer@clausfischer.com>
> http://www.clausfischer.com/
>