Portability and Parallelism with \ Lightweight
نویسندگان
چکیده
In this paper we discuss the use of cluster based computing as a basis for portable, heterogeneous parallel computing. We describe p4, a cluster based communications library which originates from Argonne National Laboratory, and our implementation of p4 on the transputer architecture. We present an extension to p4, \lightweight p4", which we designed for use in our research into automatic program parallelisation based on optimistic execution. We brieey discuss this research and the underlying techniques which are drawn from parallel discrete event simulation.
منابع مشابه
A lightweight approach to performance portability with targetDP
Leading HPC systems achieve their status through use of highly parallel devices such as NVIDIA GPUs or Intel Xeon Phi many-core CPUs. The concept of performance portability across such architectures, as well as traditional CPUs, is vital for the application programmer. In this paper we describe targetDP, a lightweight abstraction layer which allows gridbased applications to target data parallel...
متن کاملSource-to-source compilation of loop programs for manycore processors
It is widely accepted today that the end of microprocessor performance growth based on increasing clock speeds and instruction-level parallelism (ILP) demands new ways of exploiting transistor densities. Manycore processors (most commonly known as GPGPUs or simply GPUs) provide a viable solution to this performance scaling bottleneck through large numbers of lightweight compute cores and memory...
متن کاملGSHMEM: A Portable Library for Lightweight, Shared-Memory, Parallel Programming
As parallel computer systems evolve to address the insatiable need for higher performance in applications from a broad range of science domains, and exhibit ever deeper and broader levels of parallelism, the challenge of programming productivity comes to the forefront. Whereas these systems (and, in some cases, devices) are often constructed as distributed-memory architectures to facilitate eas...
متن کاملHPVM: A Portable Virtual Instruction Set for Heterogeneous Parallel Systems
We describe a programming abstraction for heterogeneous parallel hardware, designed to capture a wide range of popular parallel hardware, including GPUs, vector instruction sets and multicore CPUs. Our abstraction, which we call HPVM , is a hierarchical dataflow graph with shared memory and vector instructions. We use HPVM to define both a virtual instruction set (ISA) and also a compiler inter...
متن کاملTouchstone { a Lightweight Processor Benchmark Touchstone { a Lightweight Processor Benchmark
Benchmarks are valuable for comparing processor performance. However, porting and running processor benchmarks on new platforms can be di cult. Touchstone, a simple addition benchmark, is designed to overcome portability problems, while retaining performance measurement accuracy. In this paper, we present experimental results that show Touchstone correlates strongly with processor performance u...
متن کامل