Greg KH a écrit :
On Fri, Aug 13, 2010 at 05:37:04PM +0200, Jean-Christian de Rivaz wrote:
Foster, Dawn M a écrit :
On Aug 12, 2010, at 3:42 AM, David Greaves wrote:
This is the way open source projects are supposed to work. The
people who start the project pick a manageable set of hardware to
get us started just like when Linus only supported 386 with AT
drives in the first version of the Linux kernel because that's
what he was using at the time[1]. [1]
http://www.linux.org/people/linus_post.html
Sorry, but this argument is completely wrong: At the time Linus
started Linux, porting to other hardware was a hug task that
involved years of work. Today, every large distribution routinely
build generic i686 build. Why not Meego ?
Speed. Seriously, go measure it with it turned off, it is very
noticable. And on these tiny netbooks, you need all the speed you can
get.
There's a reason MeeGo is the fastest booting and running of _all_
distros out there at the moment, and this is one of them.
Easy to say, but what real facts can you show to support your claim ?
The booting process has been analyzed by many peoples the last couple of
years. I have not read all of there conclusions, but from what I
remember, reducing the I/O activity (in number and in latency) was a far
biggest point than the SSSE3 set instruction.
If I compare the boot time of for example Debian squeeze i386 and Debian
squeeze amd64 on the same machine (with a AMD or a Intel CPU), I did not
see a noticeable difference.
The SSSE3 instructions will probably only play a noticeable difference
in a codec or stream processing code. In those case, the good practice
is to select at runtime between a generic code and a instruction
specific code. If fact I highly doubt that any upstream project Meego
uses will run only on SSSE3 machine.
Jean-Christian
_______________________________________________
MeeGo-dev mailing list
MeeGo-dev@meego.com
http://lists.meego.com/listinfo/meego-dev
