Bio-Data
Professor, Department of Computer
Science & Engg., IIT Bombay
Research
-
Areas of Interest
Have been working in the area of compilers and program
analysis since 1989. Current broad area of research is precise
and scalable program analysis. topics of interest are all
aspects of Data Flow Analysis (eg. interprocedural analysis,
heap analysis, pointer analysis etc.).
-
We have an active group of people working in the area of
programming systems. Current list of active collaborators
includes a couple of faculty members in CSE (Amitabha Sanyal ,
Supratim Biswas ),
several students and project
engineers. The names of the project engineers can be found on
the GCC Resource
Center web page.
-
Publications and other technical writing
-
Current Research
(Last updated on 14 Dec 2022)
There are three major strands in my current work on precise
and scalable program analysis. Please see this presentation to get
an overview.
- The first strand explores precise pointer analysis of
C/C++ programs in its full glory where precision is with
respect to non-deterministic branching in programs. While
the rest of the world tried to begin with scalable and
imprecise methods and tries to increase their precision
heuristically (and hence without precision guarantees), we
begin with precise but non-scalable methods and try to
increase their scalability heuristically (thereby
preserving and hence guaranteeing, precision modulo
non-deterministic branching).
- The second strand explores precise and scalable
interprocedural analysis where precision requires
maintaining full flow- and context-sensitivity.
- The third strand of my research proposes CoS-SSA
(context-sensitive SSA) which extends the default
intraprocedural SSA (which is restricted to local
non-address taken variables) to global variables at the
interprocedural level without any compromise on
context-sensitivity even in presence pointers and
recursion.
-
Past Research
A Generalised Theory of Exhaustive and Incremental Data Flow
Analysis. This work has provided
- First ever theoretical framework for handling
bidirectional data flow problems.
- A unified framework for handling unidirectional and
bidirectional data flow problems.
- A functional model for unidirectional and bidirectional
incremental data flow analysis.
- Efficient, simple, and generic algorithms for
exhaustive and incremental analysis of unidirectional and
bidirectional flows.
- First ever complexity bounds for bidirectional data
flows
- More precise and more general unified complexity bounds
for unidirectional and bidirectional data flows.
Back to
top of the page
Back to the main page