Many-core systems are emerging as the dominant architectural paradigm in high-performance computing. As the number of cores continues to increase, effective resource allocation and scalability across the application and system software stacks becomes paramount. Virtualization can address these challenges while simultaneously increasing productivity in high-performance software development for these platforms. The VT-ASOS project augments virtualization technologies to provide encapsulated and customized hardware execution environments for parallel applications.
We use a model in which the tasks comprising a parallel application are encapsulated in their own runtime execution environment. VT-ASOS consists of user-level runtime libraries, combined with a minimal, custom guest operating system that knows the application''s requirements and implements only the functionality that the application needs. Each guest kernel runs on top of an enhanced virtual machine monitor (VMM), which makes resource management decisions that optimize the assignment of hardware resources to tasks.
VT-ASOS will support application-specific system configuration, application-specific schedulers, and many-core aware memory management policies. A monitoring infrastructure exploits hardware and software counter to guide the decisions made by the infrastructure. The research uses the Xen VMM and Linux as a testbed.
This research will improve virtualization technologies, reducing the overhead of virtualization and improving performance isolation between virtual machines sharing the resources of dense shared-memory multiprocessors. An outcome of this project will be a limited and well-defined set of programmable modules for plugging application-specific policies and mechanisms into virtual machines.
