This is the mail archive of the gcc-patches@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] Teach VRP to derive ranges from SWITCH_EXPR (fix PR21258)


Richard Guenther wrote on 04/12/07 10:58:

> There's a 1% slowdown for compiling tree.i with an unoptimized, checking
> enabled stage1 compiler.
> The patch will be on the C++ tester tonight to gather
> more numbers.  GCC of course has lots of switch statements but also
> benefits most from this optimization.

Well, it's important because last time I tried, the slowdowns didn't
justify the opportunities VRP found.  VRP is already slow.

Transformations for the sake of transformations are rarely a good idea.

> I think that's an orthogonal problem -- will the propagator engine
> be able to handle SWITCH_EXPR if I just touch this piece of VRP?

The only limitation is that you won't be able to tell it that more than
one edge is executable.  It understands ALL, 0 or 1.


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