GPU 解决方案

GPU 解决方案

单个 GPU 的计算能力相当于 100 个 CPU,因此可帮助科学家解决以前无法解决的问题。Altair 的 GPU 解决方案可帮助客户优化设计,从而改善产品性能和终端用户体验,实施监控和预测性维护,降低运营成本,以及延长产品寿命。GPU 在现代高性能计算 (HPC)、人工智能 (AI) 和机器学习应用程序框架中至关重要。


NVIDIA GPU 提供支持,将人工智能科学推向新高度

NVIDIA 的 Dai Yang 博士和 Michael Liebenthron 讨论了使用全球基于 NVIDIA A100 Tensor Core GPU 构建的 AI 系统 DGX A100,将人工智能和仿真科学推向新高度。




利用 GPU 加快数值仿真速度意味着将能更好地研究和调整设计,并根据更准确的结果更快地做出决策。可以在 GPU 中对多物理场进行准确仿真,在本地或云中实现灵活且多环境的研究和评估方案,显著加快设计工作流。



GPU 能够显著提高速度和吞吐量,从而缩短产品上市时间。客户可以自由部署更节能的计算系统,并能在本地或云中创建灵活的设计和仿真混合环境。



与同等配置的 CPU 相比,GPU 在执行超密集的计算操作时速度提高了六倍,使您的产品能够在现实世界中超过竞争对手一步。




随着渲染越来复杂,质量越来越高,各个行业的设计师、工程师和美工都在努力寻找比以往更快的速度制作视觉上更丰富的内容,同时克服基于 CPU 的低效解决方案带来的局限性。

Altair 基于物理的全局照明渲染器可在 GPU 中提供近乎完备的可扩展性,它具有多个引擎,能够创造出色的逼真场景和复杂动画。Altair 可通过功能强大的 GPU 提供极其高效且完备的渲染解决方案,从桌面上的交互式会话到数据中心的批量渲染全部覆盖。


The Keys to Scalable, Cost-Effective CFD Investment

Fluid mechanics simulation is a critical tool for late-stage failure risk mitigation, as well as a driver of design insights throughout the product development process. Used across all levels of product design and validation, from design engineers seeking to understand fluid and thermal effects on a design proposal to analysts performing advanced aerodynamic modeling, Computational Fluid Dynamics (CFD) serves a broad array of applications and a range of users with varied levels of expertise. The sometimes complex and computationally intensive nature of CFD necessitates careful consideration of the software and hardware investments required to produce accurate solutions and scale them at the speed of a company’s development process.

Technical Document

The Future of HPC

Dr. Bill Nitzberg, CTO, PBS Works at Altair, provides his perspective on the Future of HPC. HPC improves people’s lives – it’s the engine that accelerates solving complex science and engineering problems – making products more sustainable, making cars safer, making dangerous weather more predictable, even making medicines more effective. High Performance Computing has always been about leveraging lots of computing power to solve big problems...quickly. “Big” and “fast” define HPC, with the latest milestone, Exascale computers, capable of a billion billion calculations per second, expected to arrive next year. What is changing is what’s inside these engines of computing, how HPC is being consumed, and who is using HPC. HPC is becoming more and more heterogenous, with the addition of GPUs, vector engines, and high-performance versions of the Arm processors that run on today’s mobile phones, and, eventually, quantum processors. HPC is moving from “owning big, expensive iron” to a combination of buying and renting (via the Cloud). And, HPC is expanding beyond an “engineering experts only” discipline to a more mass-market audience, and also to a broader set of disciplines, especially those leveraging AI, machine learning, and data analytics. The recording is about 11 minutes long, and was presented at the 2020 Global Altair Technology conference.

Global ATC 2020

Accelerate Decision-Making with GPUs

Harnessing the computational capabilities of the graphics processing unit (GPU) is one of the cornerstones of Altair's mission to empower its customers. First and foremost, utilizing GPUs to accelerate numerical simulation delivers significant increases in speed and therefore, throughput. That means more opportunities to explore and fine-tune designs, faster decision-making based on more accurate results, and consequently the prospect of considerable reductions in time to market. In addition, creativity is enhanced, with more vivid, realistic and accurate rendering and visualization put within easier reach.

Product Overview Videos

Artificial Intelligence Taking to the Cloud

This presentation outlines Oracle’s vision for the future of artificial intelligence, including how AI and automation will drive the next wave of business transformation, the proliferation of technologies focused on understanding human language, intent, and context, like natural language processing (NLP), and how they are reshaping our interactions and businesses, as well as the critical role of high-performing, low-latency infrastructure in driving successful AI initiatives. This keynote by Clay Magouyrk, Vice President for Oracle Cloud Infrastructure, was presented at the 2021 Altair HPC Summit, and is over 16 minutes long. View all HPC Summit 2021 Presentations

HPC Summit 2021