Introduction of SX-Aurora TSUBASA for ISC21 (PCIe card type vector computer)
Presenter: Tsutomu Takeda, Manager, NEC Corporation
Abstract
SX-Aurora TSUBASA is a PCIe card type vector supercomputer and inherits the technology of SX Vector Supercomputer from 1983. The presentation will give the feature of SX-Aurora TSUBASA briefly, and announce two new functions: support of Cloud Bursting on NEC job scheduling SW and NEC LLVM-IR Vectorizer. Staring card business and channel sales will be shared as well. Besides, we will see the value proposition of Vector Engine. The Vector Engine can show high performance for not only traditional HPC applications but also a variety of other applications. Some of BM results will be presented, related to meteorology, manufacture, finance (STAC-A2) and 5G. Finally I will share MPI communication between x86, GPU and Vector Engine for the future of heterogeneous computing.
The talk will also cover the roadmap of SX-Aurora TSUBASA including Vector Engine 3.0 information.

Vector Engine for Geoscience -- From Oil & Gas to Carbon storage --
Presenter: Masashi Ikuta, Manager, NEC Corporation
Abstract
This presentation will focus on how Vector Engine can contribute to geoscience domain. The talk will start from where VE can shine in the Exploration and Production value chain of oil and gas. To prove this, some performance numbers including RTM(Reverse Time Migration), FWI(Full Waveform Inversion) and reservoir related calculation such as spmv will be discussed. This talk will also touch upon CO2 storage and monitoring which is becoming a hot topic today.

OpenMP Target Device Offloading for SX-Aurora TSUBASA
Presenter: Dr. rer. nat. Tim Cramer, IT Center, Group, High Performance Computing, Division, Computational Science and Engineering, RWTH Aachen University
Abstract
Driven by the heterogeneity trend in modern supercomputers, OpenMP provides support for heterogeneous systems since 2013. Having a single programming model for all kinds of accelerator-based systems decreases the burden of code porting to different device types.
The acceptance of this heterogeneous paradigm requires the availability of corresponding OpenMP compiler and runtime environments supporting different target device architectures. The LLVM/Clang infrastructure is designated to extend the offloading features for any new target platform. In this talk we present the opportunities for OpenMP target offloading to the NEC SX-Aurora TSUBASA vector engine based on the LLVM/Clang infrastructure.

Open Source LLVM for SX-Aurora TSUBASA
Presenter: Simon Moll, NEC Deutschland GmbH
Abstract
This talk covers the latest developments in the open source LLVM for SX-Aurora TSUBASA stack: automatic outer-loop vectorization, packed-mode support and OpenMP target offloading.

Delivering Powerful HPC with Altair and NEC on the SX-Aurora TSUBASA System
Presenter: Sujata Patnaik, Senior software engineer, Altair Engineering, Inc.
Abstract
Altair and NEC have teamed up to bring you the NEC SX-Aurora TSUBASA high-performance computing (HPC) architecture — based on NEC’s own Vector Engine (VE) — with fast, powerful workload management by Altair^® PBS Professional^®.

The addition of PBS Professional as NEC’s new solution for job scheduling will simplify access for customers who already use this workload manager to implement SX-Aurora TSUBASA.

In this session we will discuss the technical features that make this integration unique and highly competitive and how this optimized computing will make engineers and organizations more productive.

GigaIO and NEC HPC/AI Solutions: Composable Architectures with Next-Gen Interconnects and Vector Engines
Presenter: Marc Lehrer, VP, Global Sales, GigaIO
Abstract
This presentation is in two parts. 1) an overview of the GigaIO and FabreX™ universal fabric, and 2) the GigaIO/NEC SX-Aurora TSUBASA benchmarking results.

GigaIO has invented the first truly composable cloud-class software-defined universal infrastructure fabric, empowering users to accelerate workloads on-demand, using industry-standard PCIe Express technology. The company’s technology optimizes cluster and rack system performance, and greatly reduces total cost of ownership. With the innovative GigaIO FabreX open architecture, data centers can scale up or scale out the performance of their systems, enabling their existing investment to flex as workloads change over time.