Reliability is set to become a major concern on emergent large-scale architectures. While there are many parallel languages, and indeed many parallel functional languages, very few address reliability. The notable exception is the widely emulated Erlang distributed actor model that provides explicit supervision and recovery of actors with isolated state. We investigate scalable transparent fault tolerant functional computation with automatic supervision and recovery of tasks. We do so by developing HdpH-RS, a variant of the Haskell distributed parallel Haskell (HdpH) DSL with Reliable Scheduling. Extending the distributed work stealing protocol of HdpH for task supervision and recovery is challenging. To eliminate elusive concurrency bugs, we validate the HdpH-RS work stealing protocol using the SPIN model checker. HdpH-RS differs from the actor model in that its principal entities are tasks, i.e. independent stateless computations, rather than isolated stateful actors. Thanks to statelessness, fault recovery can be performed automatically and entirely hidden in the HdpH-RS runtime system. Statelessness is also key for proving a crucial property of the semantics of HdpH-RS: fault recovery does not change the result of the program, akin to deterministic parallelism. HdpH-RS provides a simple distributed fork/join-style programming model, with minimal exposure of fault tolerance at the language level, and a library of higher level abstractions such as algorithmic skeletons. In fact, the HdpH-RS DSL is exactly the same as the HdpH DSL, hence users can opt in or out of fault tolerant execution without any refactoring. Computations in HdpH-RS are always as reliable as the root node, no matter how many nodes and cores are actually used. We benchmark HdpH-RS on conventional clusters and an High Performance Computing platform: all benchmarks survive Chaos Monkey random fault injection; the system scales well e.g. up to 1,400 cores on the High Performance Computing; reliability and recovery overheads are consistently low even at scale.

Graphics Processing Units (GPUs) offer potential for very high performance; they are also rapidly evolving. Obsidian is an embedded language (in Haskell) for implementing high performance kernels to be run on GPUs. We would like to have our cake and eat it too; we want to raise the level of abstraction beyond CUDA code and still give the programmer control over the details relevant to kernel performance. To that end, Obsidian provides array representations that guarantee elimination of intermediate arrays while also using the type system to model the hierarchy of the GPU. Operations are compiled very differently depending on what level of the GPU they target, and as a result, the user is gently constrained to write code that matches the capabilities of the GPU. Thus, we implement not Nested Data Parallelism, but a more limited form that we call Hierarchical Data Parallelism. We walk through case-studies that demonstrate how to use Obsidian for rapid design exploration or auto-tuning, resulting in performance that compares well to the hand-tuned kernels used in Accelerate and NVIDIA Thrust.

The runtime for a modern, concurrent, garbage collected language like Java or Haskell is like an operating system: sophisticated, complex, performant, but alas very hard to change. If more of the runtime system were in the high-level language, it would be far more modular and malleable. In this paper, we describe a novel concurrency substrate design for the Glasgow Haskell Compiler that allows multicore schedulers for concurrent and parallel Haskell programs to be safely and modularly described as libraries in Haskell. The approach relies on abstracting the interface to the user-implemented schedulers through scheduler activations, together with the use of Software Transactional Memory to promote safety in a multicore context.

We compare two inherently different approaches to implement complex process systems in Eden: stable process systems and a compositional approach. A stable process system is characterised by handling several computation stages in each of the participating processes. Often, processes communicate using streams of data, change behaviour with the different computation phases, and more often than not, exactly one process is allocated to each processor element. In contrast, a complex process system can also be achieved by skeleton composition of a number of elementary skeletons, such as parallel transformation, reduction, or special communication patterns. In a compositional implementation, each computation phase leads to a new set of interacting processes. When implementing complex parallel algorithms, skeleton composition is usually easier and more flexible, but has a larger overhead from additional process creation and communication. We present case studies of different parallel application kernels implemented as stable systems and using composition in Eden, including a comprehensive description of Eden's features. Our results show that the compositional performance loss can be alleviated by co-locating processes which directly communicate, and by using Eden's remote data concept to enable such direct communication. Moreover, Eden's parallel runtime system handles communication between co-located processes in an optimised way. EdenTV visualisations of execution traces are invaluable to analyse program characteristics and for targeted optimisations towards better process placement and communication avoidance.

A classic problem in parallel computing is determining whether to execute a thread in parallel or sequentially. If small threads are executed in parallel, the overheads due to thread creation can overwhelm the benefits of parallelism, resulting in suboptimal efficiency and performance. If large threads are executed sequentially, processors may spin idle, resulting again in sub-optimal efficiency and performance. This “granularity problem” is especially important in implicitly parallel languages, where the programmer expresses all potential for parallelism, leaving it to the system to exploit parallelism by creating threads as necessary. Although this problem has been identified as an important problem, it is not well understood—broadly applicable solutions remain elusive. In this paper, we propose techniques for automatically controlling granularity in implicitly parallel programming languages to achieve parallel efficiency and performance. To this end, we first extend a classic result, Brent's theorem (a.k.a. the work-time principle) to include thread-creation overheads. Using a cost semantics for a general-purpose language in the style of lambda calculus with parallel tuples, we then present a precise accounting of thread-creation overheads and bound their impact on efficiency and performance. To reduce such overheads, we propose an oracle-guided semantics by using estimates of the sizes of parallel threads. We show that, if the oracle provides accurate estimates in constant time, then the oracle-guided semantics reduces the thread-creation overheads for a reasonably large class of parallel computations. We describe how to approximate the oracle-guided semantics in practice by combining static and dynamic techniques. We require the programmer to provide the asymptotic complexity cost for each parallel thread and use runtime profiling to determine hardware-specific constant factors. We present an implementation of the proposed approach as an extension of the Manticore compiler for Parallel ML. Our empirical evaluation shows that our techniques can reduce thread-creation overheads, leading to good efficiency and performance.