On Aug 30, 5:06 pm, Roy Schestowitz <newsgro...@xxxxxxxxxxxxxxx>
wrote:
> Race Ubuntu Karmic Alpha 2 vs Windows 7 RC
> http://www.youtube.com/watch?v=ymbB8RT6Aas
> It's not just speed by the way.... GNU/Linux is ahead of Vista 7 at user
> experience:
> Linux native multitouch support
> ,----[ Quote ]
> | This demo made by Mohamed-Ikbel Boulabiar, Stephane Chatty and Sebastien
> | Hamdani from the Interactive Computing Lab at ENAC shows how one can use the
> | multi-touch capabilities that Henrik Rydberg added to the Linux input system.
> `----
Multitouch was originally written or UNIX/X11, porting it to Linux was
trivial.
There may be patent license issues. It may also end up being
"intuitively derived" by some college student based on watching CNN.
It's probably not fair, or even Legal to publish benchmarks between a
production stable version of Linux and a Windows 7 Release Candidate.
The release candidate usually has extra debug flags and some of the
optimization is turned off to make issue resolution and root cause
analysis easier. Meanwhile, Linux is optimized to the hilt by the
time it is production stable.
A more "fair" comparison would be between Linux compiled with all the
"debug" and "nooptimization" flags on gcc, and Windows 7.
Linux has been running circles around Windows pretty much since 2001
when Linux released the 2.4 kernel with the queue based scheduler that
eliminated all the spinlock contention that existed in the 2.2 version
that Microsoft exploited to make it look like NT 4.0 was faster than
Linux (Mindcraft Benchmarks).
Microsoft hasn't permitted the publication of a benchmark since 2.4
and 2.6 is even faster. Microsoft considers such benchmarks to be
"damaging to the brand" and as a result, forbids publicaton of
benchmarks without their prior written approval. Microsoft can also
tell the benchmarking organization to rewrite their results to make it
look like Windows won, even if the "win" was under really bizarre
conditions.
Microsoft did try to "prove" that NTFS was faster than EXT3 by using
capture sizes optimized only for NTFS. File sizes were exact
multiples of NTFS cluster sizes, access was non-sequential, and files
were copied in an order that was intentionally working against the
EXT3 indexing.
Even that disappeared when read-ahead and write-behind caching was
turned on for Linux.
|
|
