[VIVEKFEST24] Vivek Sarkar (Georgia Tech)
summary
History of compiler and parallel language work with concrete outcomes: a production middle-end optimizer (AXA/AXAI) ships in Fortran compilers since 1996; a JVM implemented in Java proves it scales better than licensed native JVMs on SPECjbb-style benchmarks; an X10-style async/finish/futures model yields provable guarantees like deadlock freedom and data-race-implies-determinism; a practical SSA/array-SSA extension adds timestamps for executable semantics; group produces about 25 PhD graduates and multiple long-lived industrial transfers.
AI-written summary · about
chapters (13)
- 0:00Welcome and reflections
- 3:32Early life influences
- 10:21Fran Allen mentorship
- 12:05IBM tech transfer
- 15:12SSA form collaboration
- 19:07Jikes RVM breakthroughs
- 21:42Founding X10
- 22:03Rice research growth
- 25:15Medical research applications
- 26:20Georgia Tech adaptations
- 27:44Current research focus
- 30:55PhD advising legacy
- 33:03Closing thoughts
related talks
01h 06mWhy are partial evaluation and supercompilation still not widely used in practic
39mAsk Me Anything with Alex Aiken
22mCompilers should get over themselves and respect semantic dependencies!
01h 16m“What Has My Compiler Done for Me Lately? Unbolting the Compiler's Lid”
28mSSA CFG Yul A new viaIR backend
09mCompilers for nothing, executables for free!
59mCompilers can be surprising
17mRecent Developments in Optimizing Compilers for RISC-V