CCL: help needed

Hi Perry,
could you suggest some more refernces where a clear distinction espcially in terms of the diffrences between the architectural differences of windows and lInux is highlighted. thanks,
Arindam

On 10/26/05, Perry E. Metzger perry[#]piermont.com <owner-chemistry_+_ccl.net> wrote:

Sent to CCL by: "Perry E. Metzger" [perry]=[piermont.com]

"Nuno A. G. Bandeira" writes:
> Sent to CCL by: "Nuno A. G. Bandeira" [nuno.bandeira []ist.utl.pt]
> Perry E. Metzger perry^piermont.com wrote:
>
>> Microsoft also has a similar offering that they give away called
>> (inexplicably) "Services for Unix". I've heard they're getting rid of
>> it though.
>> Performance under Cygwin will not rival performance on a real Linux
>> box -- Windows has architectural issues that make it perform pretty
>> badly in general, and Cygwin is just a POSIX scrim over Windows.
>
> This is a false issue. If the program were natively compiled in
> Windows the performance would be different. Problem is not many take
> the trouble to make their programs transferable to microsoft
> platforms. Many have never heard of Visual Studio .NET or even care to
> think of anything non-unix. Tha fault is on them not the OS itself.

Sadly, you are wrong on this. Windows has a number of very serious
design flaws, especially in things like the virtual memory management
subsystem, which seriously compromise performance. For example, it is
difficult to do the sort of "memory for disk I/O" tradeoff that you do
on most Unix systems under Windows because the ability to tune buffer
cache policy is extremely limited. It is easy to put Windows into
situations where it becomes I/O bound even though there is enormous
amounts of memory available for caching.

If you insist on arguing about this with me, I suggest you do it off
the list -- most of the people here aren't going to be interested in
the intimate details. However, the fact here are straightforward. I've
studied the issue very carefully. Windows just doesn't perform well in
a wide variety of cases interesting to people who do computational
chemistry. It is also expensive. There is no good reason to use it on
machines that crunch numbers for a living, and there are excellent
reasons *not* to use it.

Perry



-= This is automatically added to each message by the mailing script =-
To recover the email address of the author of the message, please change look up the X-Original-From: line in the mail header.

E-mail to subscribers: CHEMISTRY_+_ccl.net or use:
      http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST_+_ccl.net or use
      http://www.ccl.net/cgi-bin/ccl/send_ccl_message

Subscribe/Unsubscribe:
      

Before posting check wait time for next message at: http://www.ccl.net

Job advertisements: http://www.ccl.net/jobs

If your mail bounces from CCL with 5.7.1 error, check:
      http://www.ccl.net/spammers.txt

-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+