This is the mail archive of the
mailing list for the GCC project.
RE: Improving Alias Analysis
- From: "Sanjiv Kumar Gupta, Noida" <sanjivg at noida dot hcltech dot com>
- To: Richard Henderson <rth at redhat dot com>
- Cc: gcc at gcc dot gnu dot org, Diego Novillo <dnovillo at redhat dot com>
- Date: Fri, 18 Oct 2002 16:47:10 +0530
- Subject: RE: Improving Alias Analysis
On Thu, Oct 17, 2002 at 05:48:09PM +0530, Sanjiv Kumar Gupta, Noida wrote:
>> This is primarily because it does not record sufficient
>> info for alias analysis, by which, it can know what
>> values various registers are supposed to contain
>> at a particular program point.
>This is because intended usage is not "at a particular point",
>but function-wide. If you want the former, you'll have to
>come up with some different infrastructure.
True, that means this case fails because it's not covered by the
design. But,IMHO, we will need to have a mechansim which tells
us the pointers contents at a desired point considering the
address arithmatic done on them so far. please correct me if you
do not agree. For that we need to save the address arithmatic. This
could be done by calculating possible symbolic value for each register
at each program statement but I am still looking for good ideas here.