Intel Iris 5100, Ubuntu 14.04 vs. Windows 8.1

What? Windows wins? Again?

http://www.phoronix.com/scan.php?pag...winlin&n um=1
--
Silver Slimer
Wikipedia & OpenMedia Supporter
GNU/Linux advocates: http://abstrusegoose.com/558

Silver Slimer wrote:
What? Windows wins? Again?

http://www.phoronix.com/scan.php?pag...winlin&n um=1


Let's try another page for a second.

http://www.phoronix.com/scan.php?pag..._win81&nu m=3

In that example they compa

Ubuntu Open Source = Nouveau driver (so-called "untainted", written without NDA info)
Ubuntu Closed Source = Driver written by NVidia staff for Linux
Windows 8.1 = Driver written by NVidia staff for Windows

You'll notice in the top chart, the frame rate results are the same
for Ubuntu Closed Source and Windows 8.1 (both drivers written by NVidia).

*******

Now, let's go back and address the article you picked out. I don't
recollect Intel making drivers for both Windows and Linux. (I don't
think they do a binary blob for Linux, but I could be wrong. It's
not something I would even particularly care about, because who
would "game with Intel hardware" ? )

When I see mention of MESA, that's an emulation of OpenGL done in software.
It's tons slower than calling OpenGL the likes of which is in the
Nvidia closed source (binary blob) driver. At least, that's what
MESA was in the past. I haven't exactly been tracking
progress on MESA. A software emulation is always going to be
slower than one that actually uses the shaders and the like.

Your choice of Intel hardware as a means of comparing the
OSes, wasn't the best you could have done.
Select an article with NVidia or ATI/AMD GPU,
and we'll talk.

You can see in the forum comments on the Phoronix site,
the forum people would prefer to see other benchmarks. There's
certain a lot of room for cherry-picking results, if a person
is dishonest. Such benchmarks have to be done with a
great deal of care. To the point that I wouldn't
trust most of them, unless I did them myself on my
own hardware.

http://phoronix.com/forums/showthrea...IA-Windows-8-1

*******

If you want an example of gross dishonesty, take the
travesty of an article on Tomshardware. I was using my
search engine, looking for a Windows 7 versus Windows 8
gaming comparison. I wanted to see, if Tomshardware was
able to reproduce my test result, where I was noticing
a frame rate drop on a game in Windows 8.

So I go to the site, and the idiots do their test with
an LGA2011 6C 12T processor running overclocked at 4GHz.
In other words, they used as much CPU horsepower as they
could muster, thereby hiding any possible difference
between the two OSes. Such incompetence. No wonder
the frame rates were the same.

There are some more thoughtful sites on the web, like
Xbitlabs or IXBT. If they were to do such a comparison,
they'd do a run with a 6 core machine and also do a
run with a 2 core machine, and then note the differences
seen between the two. And at least show they were
making an effort.

So yeah, I love benchmarks, but there aren't too many
of them that are shining examples of objectivity.

In your quoted article, I would at least want some
expansion on whether the word MESA, means MESA software
emulated OpenGL calls, versus the kind of OpenGL support
that would be put in a native NVidia or native ATI driver.
I wouldn't buy an Intel-only platform, with the plan of
having stellar graphics frame rates. It's like complaining
that two cripples don't run very fast, and one cripple
is slightly faster than the other. A gimp race.

Paul