This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: [GCC 4.2 Project] Omega data dependence test


----Original Message----
>From: Dan Kegel
>Sent: 08 August 2005 16:41

> Sebastian Pop wrote:
>  > [http://gcc.gnu.org/wiki/Omega%20data%20dependence%20test]
>  > ...
> I can't understand a word of the proposal.

  Well, I'll pitch in, because I also wasn't sure at first whether it was
for real and what it was about, but I think I know now.  Did you google
"bill pugh omega solver" and do some background reading?  It didn't take me
too long to get the basic gist of what they're proposing (or at any rate, to
_think_ I had got it!).

  IIUIC, they want to use a linear-algorithm solver to verify the
data-dependence analyses performed by the Bannerjee analyzer by recomputing
the results from an alternative formulation so as to have a 'second opinion'
to compare the output of gcc's current analyses against.[*]

> Mabe you were trying to be funny, but it ended up being obscure.

  Although the last paragraph was purely tongue-in-cheek humour, the rest
looks genuine.  But by the time it gets to the bit about SEB and POP, I
think it's just referring to Seb Pop!

> If the average gcc developer can understand it, then
> it doesn't matter that I can't, but I have a feeling
> others might find it hard to read, too.

  The more esoteric fields of compiler design just _are_ incredibly dense,
theoretical, and hard-to-comprehend without spending a few years at
university studying them.  Unfortunately, that's just the way it is.

    cheers,
      DaveK

[*] - To a first approximation. ;)
-- 
Can't think of a witty .sigline today....


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]