Lies, damn lies and bench... oh wait (was Re: [squeak-dev] Interesting news re Dolphin ST)

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Lies, damn lies and bench... oh wait (was Re: [squeak-dev] Interesting news re Dolphin ST)

Jason Johnson-5
On Sat, Sep 27, 2008 at 2:18 PM, Stéphane Rollandin
<[hidden email]> wrote:

>> Keep in mind that while programming languages
>> have been largely "commoditized"  the best implementations of many
>> languages still cost money.  E.g. you want the best C compiler?  It's
>> not GCC.  Buy the Intel compiler and watch your code run twice as
>> fast.
>
> wrong, according to the people there (from google "icc vs gcc", first
> results page):
> http://blog.alphagemini.org/2008/03/icc-vs-gcc-43.html
> http://www.osnews.com/comments/19462
>
> business is business, quality is quality, saying business => quality is an
> ideological statement that needs yet to be proven. IMHO :)
>
> Stef

Um......  Did you actually read the links you posted?  In the first
one, the ICC bar (program run time) is *less then* half the size of
the gcc one and the blogger mentions that GCC has a long way to go at
the end of the article!

Look, I'm as glad that there is free software out there for me to use
and learn from as the next guy (and I even contribute with a
"libre"-free license).  But there is absolutely nothing wrong with
people who sell software.  They make a strategic choice.  Companies
like RedHat choose to use a software product as a loss leader to drum
up dollars for their support infrastructure.  Places like Cinicom
chose to forgo the potentially risky loss leader strategy.  If someone
thinks one of these strategies is some kind of "guiding light" and the
other is morally bankrupt then that person is, in the best case,
incredibly naive.

Reply | Threaded
Open this post in threaded view
|

Re: Lies, damn lies and bench... oh wait (was Re: [squeak-dev] Interesting news re Dolphin ST)

Stéphane Rollandin

>> http://blog.alphagemini.org/2008/03/icc-vs-gcc-43.html
>> http://www.osnews.com/comments/19462

>
> Um......  Did you actually read the links you posted?  In the first
> one, the ICC bar (program run time) is *less then* half the size of
> the gcc one and the blogger mentions that GCC has a long way to go at
> the end of the article!

Reading again, I see that the blogger's last sentence is "the advantage
of ICC over GCC is negligible and wouldn't justify the time spent in
recompilation and porting."

as for the two sets of bars, they reflect these figures:

GCC-4.1.2: 437.24 sec
GCC-4.2.3: 436.98 sec
GCC-4.3.0: 436.17 sec
ICC 10.1: 429.72 sec

and

GCC-4.1.2: 217.00 sec
GCC-4.2.3: 216.97 sec
GCC-4.3.0: 206.90 sec
ICC 10.1: 191.91 sec

(note that a comment rightly says: 'Your graphs aren't optimal, because
if you look only at the graphs, you think icc is twice as fast as gcc.')

so, to answer your question, yes I actually read the link (twice now),
and not only had a fast glance at the pictures...


> Look, I'm as glad that there is free software out there for me to use
> and learn from as the next guy (and I even contribute with a
> "libre"-free license).  But there is absolutely nothing wrong with
> people who sell software.  They make a strategic choice.  Companies
> like RedHat choose to use a software product as a loss leader to drum
> up dollars for their support infrastructure.  Places like Cinicom
> chose to forgo the potentially risky loss leader strategy.  If someone
> thinks one of these strategies is some kind of "guiding light" and the
> other is morally bankrupt then that person is, in the best case,
> incredibly naive.

sure, I said business => quality is an ideological statement.
open source => quality or business => crap would be other ideological
statements.

my point was: let's not be ideological.


regards,

Stef


Reply | Threaded
Open this post in threaded view
|

Re: Lies, damn lies and bench... oh wait (was Re: [squeak-dev] Interesting news re Dolphin ST)

Edgar J. De Cleene



El 9/28/08 5:35 AM, "Stéphane Rollandin" <[hidden email]> escribió:

> sure, I said business => quality is an ideological statement.
> open source => quality or business => crap would be other ideological
> statements.
>
> my point was: let's not be ideological.
>
>
> regards,
>
> Stef

I add two

open source ~= crap
open source = business

Cheers to you and Jason, being appassionato is not bad.

Edgar



Reply | Threaded
Open this post in threaded view
|

Re: Lies, damn lies and bench... oh wait (was Re: [squeak-dev] Interesting news re Dolphin ST)

Jason Johnson-5
In reply to this post by Stéphane Rollandin
On Sun, Sep 28, 2008 at 10:35 AM, Stéphane Rollandin
<[hidden email]> wrote:
>
> Reading again, I see that the blogger's last sentence is "the advantage of
> ICC over GCC is negligible and wouldn't justify the time spent in
> recompilation and porting."

Well, ok, it was the second to the last paragraph I was referring to.
But as for his porting comment, why does he have to port?  Is it
because the code was written specifically for gcc?

> as for the two sets of bars, they reflect these figures:
>
> GCC-4.1.2: 437.24 sec
> GCC-4.2.3: 436.98 sec
> GCC-4.3.0: 436.17 sec
> ICC 10.1: 429.72 sec
>
> and
>
> GCC-4.1.2: 217.00 sec
> GCC-4.2.3: 216.97 sec
> GCC-4.3.0: 206.90 sec
> ICC 10.1: 191.91 sec
>
> (note that a comment rightly says: 'Your graphs aren't optimal, because if
> you look only at the graphs, you think icc is twice as fast as gcc.')

Yep, missed that.  Ok, so in this particular benchmark from this
person ICC wasn't twice as fast, closer to 5% or so.

> sure, I said business => quality is an ideological statement.
> open source => quality or business => crap would be other ideological
> statements.
>
> my point was: let's not be ideological.

Fair enough.  I just get tired of seeing people poo-pooed for wanting
to have their own business.  The loss leader style of Red Hat, etc.
isn't going to work for everyone.