[PATCH] Propagate DAG node ordering during legalization and instruction selection
Jakob Stoklund Olesen
stoklund at 2pi.dk
Mon Mar 18 14:25:04 PDT 2013
On Mar 18, 2013, at 2:02 PM, Justin Holewinski <justin.holewinski at gmail.com> wrote:
> Compile-time impact is negligible for a release build on the unit tests. There is about an 8% impact with assertions enabled.
>
> I'm not sure how CSE factors in here. This propagation happens during legalization. I agree that it should be addressed, but I see it as a separate issue. Unless I'm missing some instance of CSE being performed during legalization?
CSE is continuous in SelectionDAG. Whenever you create a new SDNode, you may actually get an existing node if an identical one happens to exist.
/jakob
More information about the llvm-commits
mailing list