Homework I

Build Andersen’s analysis in SVF

• https://github.com/SVF-tools/SVF

Compare different implementations in SVF

• Runtime and memory overhead
• Dump and compare the points-to results (optional)
• Tip: commands are ./wpa –ander, ./wpa –lander, etc.

Test programs

• Mini benchmarks: https://github.com/SVF-tools/Test-Suite
• SPEC2000 (optional)

Homework II

Run the demand-driven versions of Andersen’s analysis in SVF

• On-demand flow-sensitive
• On-demand flow- and context-sensitive

<aside> 📌 Tip: the binary is “dvf”, not “wpa”

</aside>

Client: function pointers (built-in support in SVF)

Compare the overhead of analyses with different precision

Homework III (Optional)

Customize new clients of the demand-driven analysis