Continuous Program Optimization 
Gheorghe Almasi

As the complexity of both machine architecture and software systems
increases, it is likely that neither current static nor current
dynamic compilation techniques are going to be adequate to handle
emerging workloads. The static analysis algorithms in current
compilers are limited by a lack of runtime information.
Dynamic compilers cannot afford expensive optimization.
algorithms.

In this talk we present a vision for next generation compilation
systems that combines static analysis and optimization with
runtime performance and environment monitoring at and
dynamic code generation/optimization. Performance data is
collected at runtime and used by the dynamic compiler as well as
stored externally for post-run analysis and profile directed
optimization in the static compiler.