Re: revised PATCH: CALL_EXPR representation changes, part 0/9 (table of contents)

[Sandra Loosemore <sandra at codesourcery dot com>]

Jan Hubicka wrote:

> One thing I find curious still is tree-vn usage:
>
> tree-vn.c:69 (make_value_handle)                    3605256: 1.6%          0: 0.0%          0: 0.0%     400584: 1.3%      50073
> to:
> tree-vn.c:69 (make_value_handle)                    3738816: 1.6%          0: 0.0%          0: 0.0%     415424: 1.3%      51928
>
> I don't see directly how your patch can affect number of value nodes,
> but it might be just some instability of optimizers.

Just to clarify, the number of value nodes went *down* with the CALL_EXPR 
changes, rather than up.  I have not actually instrumented the code to verify 
this, but it looks to me from reading the code that this is probably the result 
of no longer value-numbering the TREE_LIST structure of the argument list in 
phi_translate() in tree-ssa-pre.c.  Specifically, this call went away as part of 
the rewrite of that function to deal with arguments explicitly instead of as a list:

	    if (listchanged)
	      vn_lookup_or_add (newarglist, NULL);

-Sandra