it is never the benchmarker's job to apologized for the inadequatices of
the benchmarked products.
now I will give to the point that they compared apples and oranges, at
the level of detail we understand, but as the market/marketing seems to
imply the products are competitive and comparable...
many products in the market today do not meet the market/marketing
expectations
rich
Martin Bochnig wrote:
On Tue, Feb 17, 2009 at 5:01 PM, paul <p...@subsignal.org> wrote:
Joerg Schilling schrieb:
The concern is why so slow 'here and here' and how do we fix it?
As long as there is no evidence that the speed difference is a result of
using OpenSolaris, I would asume that it is caused by different
optimization.
Are you saying the phoronix guys have somehow manipulated the shipped
product (opensolaris 2008.11) to make it look bad?
cheers
Paul
:::GraphicsMagick::: part of the results ("mamy times slower than linUX ...")
Look: They compared an operating system distro with a 1.5.x based Xorg
with native driver support (not vesa), glx and drm enabled,
versus
an operating system with Xorg server 1.3, ***may***be native driver
support or not, and drm and glx maybe not enabled.
While it is true that Sun should ensure, that Indiana always ships
with the very latest Xorg, the most recent drivers / hw support and so
on (what they now exactly did and delivered in an update), this can
only be one part of the matter. In their test-report they could have
written something like "unfortunately, 2008.11 only ships with
obsolete xserver 1.3 bla bla, we will be happy to repeat the test
under EQUAL conditions".
INSTEAD they compared a car with 2 cylinders defunct with another
vendor's car with all 8 cylinders active.
When you are going to review different vendor's lamps, you should
ensure that both lamps get electricity. And in case one of them
doesn't, at the minimum they have to mention it. Not doing so and
simply publishing the "results" makes them appear not trustworthy and
throws a shadow over them, rather then over the loser of their
"benchmark". What they did: They compared a Intel cpu with all caches
enabled with an AMD cpu with L2 disabled. Then they claimed "Intel is
twice as fast in all but one tests bla nonsense bla" !
%m
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
