You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Tracing an oracle pointsto query shows that 3252ms out of 4583ms is spent doing pointer analysis and it is done entirely in one thread. Making some of the passes in the pointer analysis run in parallel should help with this.
The text was updated successfully, but these errors were encountered:
Andersen's analysis is not easily parallelizable. There has been some recent work by Pingali's group at UT Austin on parallel (2010) and GPU-based (2012) implementations, but the implementation complexity is fearsome.
Tracing an oracle pointsto query shows that 3252ms out of 4583ms is spent doing pointer analysis and it is done entirely in one thread. Making some of the passes in the pointer analysis run in parallel should help with this.
The text was updated successfully, but these errors were encountered: