As the complexity and scale of modern parallel machines continue to grow, programmers increasingly rely on composition of software libraries to encapsulate and exploit parallelism. However, many libraries are not designed with composition in mind and assume they have exclusive access to all resources. Using such libraries concurrently can result in contention and degraded performance. Prior solutions involve modifying the libraries or the OS, which is often infeasible.

Python's ease of use and rich collection of numeric libraries make it an excellent choice for rapidly developing scientific applications. However, composing these libraries to take advantage of complex heterogeneous nodes is still difficult. To simplify writing multi-device code, we created Parla, a heterogeneous task-based programming framework that fully supports Python's scientific programming stack. Parla's API is based on Python decorators and allows users to wrap code in Parla tasks for parallel execution. Parla arrays enable automatic movement of data between devices. The Parla runtime handles resource-aware mapping, scheduling, and execution of tasks. Compared to other Python tasking systems, Parla is unique in its parallelization of tasks within a single process, its GPU context and resource-aware runtime, and its design around gradual adoption to provide easy migration of and integration into existing Python applications. We show that Parla can achieve performance competitive with hand-optimized code while improving ease of development.

Applications are migrating en masse to the cloud, while accelerators such as GPUs, TPUs, and FPGAs proliferate in the wake of Moore's Law. These trends are in conflict: cloud applications run on virtual platforms, but existing virtualization techniques have not provided production-ready solutions for accelerators. As a result, cloud providers expose accelerators by dedicating physical devices to individual guests. Multi-tenancy and consolidation are lost as a consequence.

Applications are migrating en masse to the cloud, while accelerators such as GPUs, TPUs, and FPGAs proliferate in the wake of Moore's Law. These technological trends are incompatible. Cloud applications run on virtual platforms, but traditional I/O virtualization techniques have not provided production-ready solutions for accelerators. As a result, cloud providers expose accelerators by using pass-through techniques which dedicate physical devices to individual guests. The multi-tenancy that drives their business is lost as a consequence.

Concurrent and parallel programming environments must balance three competing goals: performance, productivity, and generality. Most current environments take a sequential specification and parallelize it through a series of program analyses and transformations. Programmer productivity is determined by how much the programmer is involved in specifying the transform. A programming environment’s generality is determined by how much it restricts its programming model to enable the transform to be more automatic or provide better performance. Productivity and generality impact the performance of the resulting code, giving rise to a performance–productivity–generality trade-off space.

The majority of modern programming languages provide concurrency and object-orientation in some form. However, object-oriented concurrency remains cumbersome in many situations. We introduce the language OrcO, Orc with concurrent Objects, which enables a flexible style of concurrent object-oriented programming. OrcO extends the Orc programming language by adding abstractions for programming-in-the-large; namely objects, classes, and inheritance. OrcO objects are designed to be orthogonal to concurrency, allowing the concurrent structure and object structure of a program to evolve independently. This paper describes OrcO's goals and design and provides examples of how OrcO can be used to deftly handle events, object management, and object composition.