Philippe Marschall wrote:
> 2008/9/27 Joshua Gargus <schwa at fastmail.us>:
>   
>> Philippe Marschall wrote:
>>
>> 2008/9/26, Jason Johnson <jason.johnson.081 at gmail.com>:
>>
>>
>> On Thu, Sep 25, 2008 at 9:15 PM, Igor Stasenko <siguctua at gmail.com> wrote:
>>
>>
>> 2008/9/25 Chris Kassopulo <ckasso at sprynet.com>:
>>
>>
>> http://www.lesser-software.com/en/content/products/lswvst/lswvst.htm
>>
>>
>>
>>
>> It would be nice to look at it (LSWV).
>> So many cool features. Too bad its proprietary. :(
>>
>> --
>> Best regards,
>> Igor Stasenko AKA sig.
>>
>>
>> Is the cost prohibitive?  Dolphin wasn't free, but it was really cheap
>> and a quality product.  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.
>>
>>
>>
>>
>> Let me preface this by saying that I support free software, and use it in
>> preference to proprietary software when possible.  However, I disagree with
>> much of what you say.
>>
>> If you define best best as "longest bar in some benchmark".
>>
>> I define best as "producing the best performance for the program that I am
>> interested in compiling".
>>
>> Intel doesn't manage to charge serious money for their compiler because it
>> is better on some synthetic benchmark but worse in the real world.  Maybe
>> GCC has mostly caught up with 4.3, but there's no doubt that ICC has
>> traditionally generated faster code.  I don't have extensive/varied personal
>> experience with ICC, but if you compile Squeak with it on Intel, you'll see
>> a 20% speed-up on macro benchmarks.
>>
>> If you
>> want to run your software only on Intel chips, no AMD, no Motorola, no
>> Sun, no ....
>>
>> First, it's simply not true that Intel's compiler doesn't work for AMD.
>> After surfing around for 15min or so, I learned:
>> - at various times in the past, AMD chose ICC as the compiler they used to
>> generate SPEC benchmarks.
>> - there are numerous personal accounts where ICC generates the fastest code
>> for AMD for their pet application.
>>
>> Typically, Intel CPUs have a higher percentage improvement by using ICC
>> instead of GCC, but AMD CPUs also benefit.
>>
>> Also, many applications don't target PowerPC, Sparc, or ARM.  For desktop
>> apps, x86 is the name of the game.  I don't feel limited by using an
>> x86-only compiler, especially since I can write my code to compile with a
>> different compiler for other platforms.
>>
>> If you want to put up with all kinds of trouble and
>> issues compiling software that was written with GCC in mind like all
>> the GNU/Linux software.
>>
>> That's an interesting angle.  Many Free software lovers would consider code
>> that uses specific features of Microsoft's C compiler to be broken.  I
>> generally agree with them; unless there is a good reason for it (and there
>> sometimes is, just as it sometimes makes sense to write x86 assembly),
>> compiler-specific code is a bug.  Why should it be any different for
>> GNU-specific code?
>>     
>
> Such things happen automatically if you compile with only one
> compiler. 

Right, but that doesn't make it a good thing.  We do agree that it's not
a good thing to accidentally make choices that limit Squeak to a single
compiler, don't we?  If not, why is taking choice away a good thing?

> You can observe it every time a new GCC version comes out
> that is more standard conformant a lot of software breaks. 

We're not talking about the steady improvements in standard C/C++
compliance.  We're talking about intentional, compiler-specific
extensions that are not part of the C/C++ standards.  That's why it's
called "gnuify" not "standardize" :-).

> You can
> also see it with Squeak. A lot of code has empty statements and temps
> declared in a too wide scope. Once we get a new compiler that supports
> closures or you compile it with a more standard conformant compiler,
> that code has to be updated.
>   

I'm not sure what we're talking about any more.  I initially responded
to a message from you that seemed very dismissive of Jason's point,
namely that in particular circumstances, it may make sense to pay for
your language implementation.  For some reason, I felt compelled to
argue that despite your somewhat glib dismissal, there exist real
situations where it makes sense to use ICC (therefore lending support to
Jason's more general claim).

Given this context (which is of course not necessarily what the
conversation meant to you, subjectively), what are we talking about
here?  :-)

>> To me, it seems like FUD to disparage a compiler for inability to compile
>> non-standard code.
>>
>> Maybe even Squeak thanks to gnuify.
>>
>> No, gnuify transforms the code to make it run faster on GCC, but if you
>> don't apply gnuify, it will compile fine with other compilers.
>>     
>
> But slower. 

Squeak compiled with GCC is faster when gnuify is used.  But Squeak
compiled with ICC is faster still.  On what do you base your assertion
that it is slower?

> You don't actually pay for this, do you?
>
>   

For my personal Squeaking, of course I wouldn't pay for ICC.  But if
someone is working for a company whose product is performance sensitive,
why wouldn't they spend a few hundred dollars to improve performance as
much as $10000s of engineering effort would, especially if it reduces
time-to-market?

Cheers,
Josh


> Cheers
> Philippe
>
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20080927/64eed9f9/attachment.htm