Wilco (Wilco.Dijkstra.delete@this.ntlworld.com) on July 15, 2013 3:43 pm wrote:
> David Kanter (dkanter.delete@this.realworldtech.com) on July 15, 2013 1:57 pm wrote:
> > Steve (sberens.Throwaway.delete@this.gmail.com) on July 14, 2013 8:16 pm wrote:
> > > Wilco (Wilco.Dijkstra.delete@this.ntlworld.com) on July 12, 2013 11:59 am wrote:
> > > > AnTuTu apparently fixed their benchmark: http://www.eetimes.com/author.asp?section_id=36&doc_id=1318894&
> > > >
> > > > The RAM score halves on Atom but ARM scores remain the same. Apparently they
> > > > still use ICC, but hopefully AnTuTu will review this given this debacle.
> > > >
> > > > Wilco
> > > >
> > >
> > > Funny how it is only the ICC compiler from Intel that dropped code that tested the RAM.
> > >
> > > It was an Intel problem all along not AnTuTu.
> > >
> > > Intel: opps got caught cheating again so now direct flack to AnTuTu
> >
> > Newsflash: That's a perfectly legal, intelligent and reasonable optimization. Get used to it.
> >
> > It's 100% a benchmark problem that they have crap code lying around.
>
> Sure AnTuTu is a completely crap benchmark and the developers don't seem to have the faintest clue about
> benchmarking. The ARM version is compiled -Os, with all loop optimizations and inlining disabled! You'd
> get a significantly higher score with just changing to -O3 and more when you try a bit harder...
>
> However there is no doubt Intel is cheating here.
>
> The optimization in question was added recently to ICC and only speeds up one particular loop
> in AnTuTu by a huge factor, resulting 300% speedup on the memory score, and a doubling of the
> overall score. Clearly that's not a valid optimization but a benchmark busting trick.
>
> The only compiler used in Android is GCC, I am not aware of the NDK being released with ICC. Using
> the same compiler to build benchmarks means that all architectures benefit from new optimizations
> (including benchmark busting tricks), so it levels the playing field. So why did AnTuTu switch to
> ICC with the AnTuTu busting optimizations? Did Intel pay them so they could double the score?
>
> If other parties haven't been given the same opportunity to improve the AnTuTu score (if only by using a
> newer GCC and better options), then what Intel did is anti-competitive behaviour and completely illegal.

ICC can be used with the NDK.

As someone who works for ARM as compiler writer,
why don't you tell us in more detail how Intel
cheated?