|question on register spill behavior email@example.com (BX) (2008-08-27)|
|Re: question on register spill behavior firstname.lastname@example.org (2008-08-28)|
|Re: question on register spill behavior email@example.com (cr88192) (2008-08-29)|
|Re: question on register spill behavior firstname.lastname@example.org (BX) (2008-09-08)|
|Date:||Mon, 8 Sep 2008 10:52:12 -0700 (PDT)|
|Posted-Date:||10 Sep 2008 06:42:53 EDT|
Thank you both.
Just to make this message a little bit more than a waste of your eye-
ball movement, I just want to give a little background on my
question. I was doing a project on dependence analysis. In this
context, the example I gave above will give me a sort of spurious
control dependence for the last use of 'x' value after the branch (if
loaded from spill register in the branch). So, if these kind of spill
are common, then it will definitely make a negative impact on the
dependence results. One can probably work around this, but it's kind
of interesting to see how compilers optimizations and architecture-
level conventions can slightly "change" the program semantics, or at
least give headaches to writing program analysis.
Return to the
Search the comp.compilers archives again.