NAS Parallel Benchmarks with CUDA and beyond

Abstract

NAS Parallel Benchmarks (NPB) is a standard benchmark suite used in the evaluation of parallel hardware and software. Several research efforts from academia have made these benchmarks available with different parallel programming models beyond the original versions with OpenMP and MPI. This work joins these research efforts by providing a new CUDA implementation for NPB. Our contribution covers different aspects beyond the implementation. First, we define design principles based on the best programming practices for GPUs and apply them to each benchmark using CUDA. Second, we provide ease of use parametrization support for configuring the number of threads per block in our version. Third, we conduct a broad study on the impact of the number of threads per block in the benchmarks. Fourth, we propose and evaluate five strategies for helping to find a better number of threads per block configuration. The results have revealed relevant performance improvement solely by changing the number of threads per block, showing performance improvements from 8% up to 717% among the benchmarks. Fifth, we conduct a comparative analysis with the literature, evaluating performance, memory consumption, code refactoring required, and parallelism implementations. The performance results have shown up to 267% improvements over the best benchmarks versions available. We also observe the best and worst design choices, concerning code size and the performance trade‐off. Lastly, we highlight the challenges of implementing parallel CFD applications for GPUs and how the computations impact the GPU's behavior.

Assessing Application Efficiency and Performance Portability in Single-Source Programming for Heterogeneous Parallel Systems

“…There is an ongoing effort to create SkePU implementations, and subsequently evaluations, of many benchmark workloads across several benchmark suites. Such suites include Rodinia [10], PARSEC [9] and its parallel derivate P3ARSEC [15], Poly-Bench [34], and NAS Parallel Benchmarks [5,8,27]. The complexity and effort required for benchmarking parallel programming models, interfaces, and frameworks is well-known [32] and examples of ongoing efforts to simplify and standardize parallel benchmark suites are many, including P3ARSEC and Task Bench.…”

“…The original version was written in Fortran and the parallel implementations were in OpenMP and MPI. In recent years, an effort was made to provide parallel versions for C/C++ parallel programming frameworks on multicore systems [26,27] as well as heterogeneous parallel programming on GPUs [5,6,19].…”

“…In addition to the lightweight STREAM workloads, we also considered NPB as a means to select additional possible evaluation points, since these are different computations and are considered standard benchmarks for HPC evaluation. Like STREAM, NPB has not been subject to SkePU parallelization before, and given the recent work on NPB implementations in both C++ parallel CPU frameworks and CUDA for Nvidia GPUs [5,27], we have good reference points for efficiency comparison against SkePU implementations. However, SkePUizing the entirety of NPB will be a future work, because the experience from this initial effort will indicate the viability of such a project.…”

Special issue on new trends in high‐performance computing: Software systems and applications

“…No manual development and no special Field Programmable Gate Array (FPGA) or programming knowledge are required. The logic generated by this improved approach is up to 43 times faster than its hand-optimized High Level Synthesis (HLS) counterpart, depending on the solution method.The third paper titled "NAS Parallel Benchmarks with Compute Unified Device Architecture (CUDA) and Beyond" by Fernandes et al 3 provides a new CUDA implementation for NASA Parallel Benchmark (NPB). The performance results have shown up to 267% improvements over the best benchmark versions available.…”

“…The third paper titled “NAS Parallel Benchmarks with Compute Unified Device Architecture (CUDA) and Beyond” by Fernandes et al 3 provides a new CUDA implementation for NASA Parallel Benchmark (NPB). The performance results have shown up to 267% improvements over the best benchmark versions available.…”