VEO and PyVEO: Vector Engine Offloading for the NEC SX-Aurora Tsubasa

Focht, Erich

doi:10.1007/978-3-030-39181-2_9

Cited by 5 publications

(3 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Vector Engine Offloading (VEO) [8] is a programming model which executes the main program on the VH and offloads kernels to the VEs. While the API somewhat resembles OpenCL, it differs from it due to the SIMD/vector nature of kernels, and due to their ability to execute almost any Linux system call.…”

Section: Background 1sx-aurora Vector Enginementioning

confidence: 99%

Porting and Optimizing Molecular Docking onto the SX-Aurora TSUBASA Vector Computer

Solis-Vasquez

Focht

Koch

2021

JSFI

View full text Add to dashboard Cite

In computer-aided drug design, the rapid identification of drugs is critical for combating diseases. A key method in this field is molecular docking, which aims to predict the interactions between two molecules. Molecular docking involves long simulations running compute-intensive algorithms, and thus, can profit a lot from hardware-based acceleration. In this work, we investigate the performance efficiency of the SX-Aurora TSUBASA vector computer for such simulations. Specifically, we present our methodology for porting and optimizing AutoDock, a widely-used molecular docking program. Using a number of platform-specific code optimizations, we achieved executions on the SX-Aurora TSUBASA that are in average 3.6× faster than on modern 128-core CPU servers, and up to a certain extent, competitive to V100 and A100 GPUs. To the best of our knowledge, this is the first molecular docking implementation for the SX-Aurora TSUBASA.

show abstract

Section: Background 1sx-aurora Vector Enginementioning

confidence: 99%

Porting and Optimizing Molecular Docking onto the SX-Aurora TSUBASA Vector Computer

Solis-Vasquez

Focht

Koch

2021

JSFI

View full text Add to dashboard Cite

show abstract

“…In LLVM this is handled by the libomptarget library, which selects at runtime a target device for the execution of the offloaded code. We developed a corresponding plugin based on the NEC VE Offloading (VEO [15]) interface. However, our current implementation benefits from the new AVEO [16] implementation, which shows a better performance compared to our original library.…”

Section: Openmp Target Device Offloading Designsmentioning

confidence: 99%

“…Besides OpenMP target device offloading, other approaches exist in order to execute computeintensive code parts on a SX-Aurora TSUBASA vector engine. The direct use of the low-level APIs VEO [15], AVEO [16] or VHCall [3] gives the programmer full control of the data transfers and the kernel execution. Noack et al [23] built on top of the portable Heterogeneous Active Messages (HAM) a high-level C++-only framework for SX-Aurora TSUBASA offloading.…”

Section: Related Workmentioning

confidence: 99%

Evaluating the Performance of OpenMP Offloading on the NEC SX-Aurora TSUBASA Vector Engine

2021

JSFI

View full text Add to dashboard Cite

The NEC SX-Aurora TSUBASA vector engine (VE) follows the tradition of long vector processors for high-performance computing (HPC). The technology combines the vector computing capabilities with the popularity of standard x86 architecture by integrating it as an accelerator.To decrease the burden of code porting for different accelerator types, the OpenMP specification is designed to be single parallel programming model for all of them. Besides the availability of compiler and runtime implementations, the functionality as well as the performance is important for the usability and acceptance of this paradigm. In this work, we present LLVM-based solutions for OpenMP target device offloading from the host to the vector engine and vice versa (reverse offloading). Therefore, we use our source-to-source transformation tool sotoc as well as the native LLVM-VE code path. We assess the functionality and present the first performance numbers of real-world HPC kernels. We discuss the advantages and disadvantage of the different approaches and show that our implementation is competitive to other GPU OpenMP runtime implementations. Our work gives scientific programmers new opportunities and flexibilities for the development of scalable OpenMP offloading applications for SX-Aurora TSUBASA.

show abstract

Speeding Up Vector Engine Offloading with AVEO

Focht

2021

Sustained Simulation Performance 2019 and 2020

View full text Add to dashboard Cite

VEO and PyVEO: Vector Engine Offloading for the NEC SX-Aurora Tsubasa

Cited by 5 publications

References 4 publications

Porting and Optimizing Molecular Docking onto the SX-Aurora TSUBASA Vector Computer

Porting and Optimizing Molecular Docking onto the SX-Aurora TSUBASA Vector Computer

Evaluating the Performance of OpenMP Offloading on the NEC SX-Aurora TSUBASA Vector Engine

Speeding Up Vector Engine Offloading with AVEO

Contact Info

Product

Resources

About