RE: Does anyone run test win vs. linux and can share how muchfaster is linux than windows, please?

Cary Millsap wrote:

> The answer is "may not be." Imagine some syscall that's slower
> on operating system A than it is on operating system B. It's of
> course possible to construct an app on system A that uses this
> syscall so many times that the syscall's contribution to response
> time dominates the user's experience with the app.

Exactly. And that could well be the wrong type of syscall to use. For example, how do you implement traditional Unix IPC message queues on Windows..? There are several options (from WM_USER messages to DDE) and some will be simply a bad choice under certain circumstances.

> I'd like to insert the perspective, "Why bank on 'odds?'" No matter
what the
> odds, it might be affecting you; it might not. If you learn how to
measure
> response time accurately, then you'll KNOW whether you're affected,
and you'll
> be able to make an informed decision about what to do next.

Right. My take is however, with my software developer hat on, that you should not get to that stage if you know what the heck you're doing in the first place on that o/s. That's the point I'm trying to make. You cannot take your Unix savvy and apply just like that to Windows. Or vice versa. Yet, this is exactly what is being done (at sysadmin and developer and dba levels) and is the prime reason why the perception persists that "this o/s is slow and bad and sucks".

Is it frustrating to go from one o/s to another? Hell yes. But recognise the fact that the new o/s is *different*.

The old story of emptying your cup of coffee before pouring tea into it, as that's the only way you can really taste and enjoy the tea.

--
Billy

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This e-mail and its contents are subject to the Telkom SA Limited
e-mail legal notice available at
http://www.telkom.co.za/TelkomEMailLegalNotice.PDF
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
--
http://www.freelists.org/webpage/oracle-l