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: [tree-ssa] Optimizing silly ELSE clauses


In message <20030814200154.GA27085@redhat.com>, Richard Henderson writes:
 >On Thu, Aug 14, 2003 at 12:08:11PM -0600, law@redhat.com wrote:
 >> And if you do that, then you lose in cases were the conditional isn't
 >> testing "var".  If you try to make it dependent on the condition, then
 >> you're getting into the realm of a gross hack in the out-of-ssa pass.
 >
 >I was thinking more of using the available expression table,
 >and searching it for ssa versions of the LHS that do have the
 >correct value.  I see that our CCP isn't set up to allow that.
To do something like that you'd effectively need to rebuild the available
expression table that the dominator optimizer builds while doing
the out-of-ssa translation.  It's really the dominator walk that is setting
up these constants, not CCP.  CCP does propagate constants, but the dominator
optimizer is the one responsible for the cases I'm looking at.

Jeff


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