Evaluating attainable memory bandwidth of parallel programming models via BabelStream

Abstract

Many scientific codes consist of memory bandwidth bound kernels-the dominating factor of the runtime is the speed at which data can be loaded from memory into the Arithmetic Logic Units, before results are written back to memory. One major advantage of many-core devices such as General Purpose Graphics Processing Units (GPGPUs) and the Intel Xeon Phi is their focus on providing increased memory bandwidth over traditional CPU architectures. However, as with CPUs, this peak memory bandwidth is usually unachievable in practice and so benchmarks are required to measure a practical upper bound on expected performance. We augment the standard set of STREAM kernels with a dot product kernel to investigate the performance of simple reduction operations on large arrays. Such kernels are usually present in scientific codes and are still memory bandwidth bound. The choice of one programming model over another should ideally not limit the performance that can be achieved on a device. BabelStream (formally GPU-STREAM) has been updated to incorporate a wide variety of the latest parallel programming models, all implementing the same parallel scheme. As such this tool can be used as a kind of Rosetta Stone which provides both a cross-platform and cross-programming model array of results of achievable memory bandwidth.

Benchmarking the first generation of production quality Arm‐based supercomputers

“…The number reported in the chart corresponds to the achieved bandwidth for the Triad kernel. To measure the sustained cache bandwidths, we used the methodology described in our previous work, utilizing BabelStream . The Triad kernel from the BabelStream benchmark was run in a tight loop on each core simultaneously, with problem sizes selected to ensure residency in each level of the cache.…”

Intelligent Data Placement on Discrete GPU Nodes with Unified Memory

“…Effective peak single-precision FP performance is obtained from the SHOC benchmark [12]. Effective peak bandwidth is obtained via BabelStream [14]. Copy times are estimated at effective peak bandwidth.…”

Performance Portability across Diverse Computer Architectures

“…The portability of programming models across a range of architectures was explored in the BabelStream benchmark [5], and we include the latest results in this paper. The performance of a number of parallel programming models on GPUs was explored using the TeaLeaf mini-app [6], showing that each model can achieve similar performance.…”

“…For this study, we have selected five mini-apps to represent critical workloads on many of the largest supercomputers in the world: BabelStream [5], CloverLeaf [4], TeaLeaf [11], Neutral [12] and MiniFMM [13]. A short description of each mini-app will be given in Section III.…”

“…The BabelStream benchmark is a re-implementation of the classic McCalpin STREAM benchmark in many parallel programming models [5]. It also includes a dot-product kernel, and employs best-practice implementations for each model.…”