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: [PATCH] PR/17860: wrong code generated by loop optimizer up to 3.4


> -----Original Message-----
> From: Paolo Bonzini 
> Sent: 12 October 2004 13:25

> property may not be true anymore after CSE.  Uses of vtop 
> like the one I 
> removed need to be either fixed or removed; and I suspect 
> that *all of 
> the uses of vtop* need to.
> 
> Basically, I think that removing a not-very-often-used 
> transformation in 
> an almost dead optimization pass is feasible.  Remember 3.4 
> has Zdenek's 
> dominator-based loop optimizer as well.

  Mmm.  Alas I'm supporting a 3.3 based compiler here, so I won't benefit
from that (but it might be a good reason to jump up a series), but as you
say....

> So removing them can only give pessimization (and less bugs), leaving 
> them in can only give more optimization (and more bugs).

...which is really the sort of tradeoff I need to accept.  When you're
working with experimental hardware, it can become difficult to tell what's
going wrong if you can't trust the hardware _and_ you can't trust the
software.  I'm gonna just chop out the code that emits them.  Thanks for
your advice :)

    cheers, 
      DaveK
-- 
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]