I hate to rehash 10 year old dirty laundry, but this guy is bad mouthing 
POV-Ray for something the Intel compiler was doing without the POV 
team's knowledge...

http://www.extremetech.com/computing/193480-intel-finally-agrees-to-pay-15-to-pentium-4-owners-over-amd-athlon-benchmarking-shenanigans


-- 
/*Francois Labreque*/#local a=x+y;#local b=x+a;#local c=a+b;#macro P(F//
/*    flabreque    */L)polygon{5,F,F+z,L+z,L,F pigment{rgb 9}}#end union
/*        @        */{P(0,a)P(a,b)P(b,c)P(2*a,2*b)P(2*b,b+c)P(b+c,<2,3>)
/*   gmail.com     */}camera{orthographic location<6,1.25,-6>look_at a }

Post a reply to this message

On 7/11/2014 00:34, Francois Labreque wrote:
> I hate to rehash 10 year old dirty laundry, but this guy is bad mouthing 
> POV-Ray for something the Intel compiler was doing without the POV 
> team's knowledge...

Thanks for the heads-up. I'm not entirely sure what Joel's point is;
one thing he raised back in 2004 as a possibility was that the source
may have been 'hand-tweaked' to favour Intel, something that is
clearly offensive to the POV-Ray developers and demonstrably untrue
given (as I pointed out at the time) the source was available. If he
had built the code using the same compiler as we did he would have
gotten the same results, and nothing in the intervening ten years or
the Intel settlement changes that.

On the contrary, even if the compiler issues he claims to have
detected do exist (something I have not investigated), the very fact
that it would stem from the compiler and not from manual intervention
on our part demonstrates that he was completely out of line to raise
the possibility we may have hand-tweaked the code in Intel's favor.

-- Chris

Post a reply to this message

Chris Cason <del### [at] deletethistoopovrayorg> wrote:
> On the contrary, even if the compiler issues he claims to have
> detected do exist (something I have not investigated), the very fact
> that it would stem from the compiler and not from manual intervention
> on our part demonstrates that he was completely out of line to raise
> the possibility we may have hand-tweaked the code in Intel's favor.

If I understand correctly, he's referring to the official pre-compiled
executable of POV-Ray 3.6 that was being distributed at the time, and
what he's insinuating is that it was deliberately compiled (perhaps
with some compiler options?) to be slower on AMD processors. That's
what I gather from this:

"The funny thing is, I did do that - but the programmer friend who
helped me with Intel's compiler could never reproduce the results in
POV-Ray 3.6.0, despite compiling six different executables with
different optimization levels in an attempt to do so."

What I think he's trying to say here is that compiling POV-Ray 3.6
normally did not produce a slower executable on AMD, therefore the
official binary had to be compiled on purpose to be slower on AMD
by some other means.

(Of course that's not what happened, at least not on purpose, but
this is what I think he's trying to say.)

-- 
                                                          - Warp

Post a reply to this message

On 06.11.14 14:34, Francois Labreque wrote:
> I hate to rehash 10 year old dirty laundry, but this guy is bad mouthing
> POV-Ray for something the Intel compiler was doing without the POV team's
> knowledge...
>
>
http://www.extremetech.com/computing/193480-intel-finally-agrees-to-pay-15-to-pentium-4-owners-over-amd-athlon-benchmarking-shenanigans

Seems like he is a conspiracy theorist of some sort. Who on their right mind 
keeps a ten year old binary around for such purpose!?! - Usually trying to 
argue with such people is like arguing with any other mentally ill person. 
Any argument will get nowhere as reason isn't anything these people want to 
grasp as it does not fit their image of the world. Best to ignore him. 
Anything reasonable argument will just convince him he is "right" and others 
are going after him because of an imagined conspiracy.

Thorsten

Post a reply to this message

On 06.11.14 16:31, Warp wrote:
> What I think he's trying to say here is that compiling POV-Ray 3.6
> normally did not produce a slower executable on AMD, therefore the
> official binary had to be compiled on purpose to be slower on AMD
> by some other means.
>
> (Of course that's not what happened, at least not on purpose, but
> this is what I think he's trying to say.)



the POV-Ray team was saying then, the 3.6.0 version of the test was compiled 
in a manner that tilted the competitive landscape towards Intel." he clearly 
says we lied back then. In most jurisdictions this would constitute libel. 
But the website is hosted in the US ... and given he clearly has no idea how 
a compiler works (admitting he needed help with that), he probably doesn't 
even know what he is talking about. Any person with a computer science 
background will, as you pointed out, notice that he doesn't use the right 
terminology and hence doesn't know what he is talking about. Still, it is 
frightening that such people get to publish articles these days - 15 years 
ago he wouldn't have been able to connect to internet, now he writes about 
computer technology and thinks he is an expert. Oh well :-(

Thorsten

Post a reply to this message