[Soot-list] Pointer Assignment Graph for VTA

[Wei L. Zhang]

I figured out the answer for my question:
the difference is the option of on-fly-cg. The default configuration
for points to is on fly call graph construction, but it is not true
for vta. Actually, it seems that in current soot implementation vta
doesn't support on-fly-cg.
-Patrick

On Mar 31, 2005 11:25 PM, patrick zhang <iampatrick.zhang at gmail.com> wrote:
> Hi
> I am using VTA and points to analysis in spark. Interestingly, the
> Pointer Assignment Graph construction time for VTA is MUCH longer than
> the points to analysis, which is counter intuitive imho. Would you
> please tell me that is there anything wrong with the parameters?
> Thanks
> -Patrick
> 
> For example:
> The following comes from the result when I run vta on _201_JESS
> The parameter to the soot is "--app -p cg.spark vta:true -w"
> [Spark] Pointer Assignment Graph in 226.4 seconds.
> [Spark] Type masks in 6.5 seconds.
> [Spark] Pointer Graph simplified in 2.8 seconds.
> [Spark] Propagation in 3.2 seconds.
> [Spark] Solution found in 6.1 seconds.
> 
> The following comes from the result when I run points to analysis on _201_JESS
> The parameter to the soot is "--app -p cg.spark on-fly-cg:true -w"
> [Spark] Pointer Assignment Graph in 10.0 seconds.
> [Spark] Type masks in 1.8 seconds.
> [Spark] Pointer Graph simplified in 0.0 seconds.
> [Spark] Propagation in 99.4 seconds.
> [Spark] Solution found in 99.4 seconds.
> _______________________________________________
> Soot-list mailing list
> Soot-list at sable.mcgill.ca
> http://www.sable.mcgill.ca/mailman/listinfo/soot-list
>