Advanced Computing Solutions

高级计算解决方案提供业界领先的高性能计算技术. Built on FPGA technologies and targeted at PCI Express® (PCIe®)-based HPC or stand-alone embedded applications, 我们的模块化和高度可伸缩的架构提供了数量级的性能提升, reduced energy costs, small form factors and simplified application designs.

Overview

为了将今天的数据洪流转化为洞察力和竞争优势,我们需要具有惊人速度和能源效率的计算系统. 实现这一目标的一种强大的技术是将计算密集型的工作负载转移到加速器上,加速器是一种特殊用途的设备,用于比通用计算系统更快地执行特定任务.

美光的紧凑和高度可扩展的基于fpga的加速器解决方案提供卓越的性能,并为广泛的先进算法提供简单的应用集成. 

Micron Announces Comprehensive AI Development Platform


美光最近宣布收购FWDNXT,为深度学习应用提供一套强大的高性能硬件和软件工具, a software and hardware start-up based in Indiana. FWDNXT’s AI hardware and software technology, when combined with advanced Micron memory, 使美光能够探索数据分析所需的深度学习解决方案, particularly in IOT and edge computing. With this acquisition, Micron is integrating compute, memory, 工具和软件成为一个全面的人工智能开发平台,以提供探索为人工智能工作负载优化的创新记忆所需的关键构建块.   

Learn more

Deep Learning Accelerators

深度学习使用神经网络来执行机器视觉等任务, voice recognition and natural language processing. 训练有素的模型通常被部署在许多设备上,比如手机,以从新数据中得出见解——这一过程被称为推理.

美光的深度学习加速器使推理任务能够更快地执行, using much less energy, than general-purpose computers.

开发人员可以轻松利用我们的技术,使用广泛的流行开源框架探索和部署深度学习模型,以解决从边缘到云端之前难以解决的数据分析问题. 

Learn more

AI/DL Platform Overview

Fully Integrated SDK

  • Broad deep learning framework support
  • Easy to use toolset; software programmability
  • Ability to run on multiple neural networks simultaneously

Deep Learning Compiler

  • 直接编译经过训练的网络到推理机机器码
  • Underlying hardware is abstracted away
  • Feature rich & designed to optimize for specific use cases

Advanced Accelerators

  • 具有高性能内存的FPGA解决方案可用于原型开发或早期部署
  • Scale with multiple cores
  • Integrated with SDK to fully utilize silicon resources

Advanced Accelerators

Our modular and massively scalable architecture, 基于先进的内存和现场可编程门阵列(FPGA)技术, brings industryleading performance, high-compute density, 将节能和简化应用设计提升到一个全新的水平.

Advanced Accelerator Products

AC-511 module uses a Xilinx Virtex Ultrascale+ FPGA. It is performance and bandwidth together.

AC Series HPC Modules

Our business card-sized, 高性能计算(HPC)模块是基于PCI Express的真正可伸缩系统的基本构建块® (PCIe®) standard.
Learn more >
+
EX-750背板为安装在其上的每个AC系列模块提供独立的PCIe x8 Gen3总线.

HPC Backplanes

Modular, scalable performance delivered. 我们的两类HPC背板旨在产生最高的性能和计算密度,同时简化设计和部署,以适应特定的计算需求.
Learn more >
+
SB-852 is full-height, GPU-length. PCIe x16 Gen3 board with a Xilinx�Virtex Ultrascale+ FPGA.

HPC Single-Board Accelerators

Highest performance. 我们的单板计算解决方案结合了基于fpga的计算能力和我们最快的内存技术,以满足最具挑战性的系统需求.
Learn more >
+
SC6-4U data center system

HPC Data Center Systems

Solving the world’s most difficult computing challenges. Our industrial-grade, FPGA-accelerated, 机架式HPC数据中心系统比传统服务器快上几个数量级,同时保持较小的物理足迹和消耗少量的能量.
Learn more >
+
SC6-Mini desktop system

HPC Desktop Systems

High performance for desktops. Our HPC desktop solution is a flexible, high-performance supercomputer in a low-power, 占地面积小,为升级提供了无障碍路径, is easy to use when coupled with OpenCL and the PicoFramework, 并且可以配置任何基于fpga的PCIe加速板.
Learn more >
+

HPC Development Environment

PicoFramework

The highly efficient, 基于linux的picofframework在主机上运行的应用程序软件和硬件算法之间提供了一个重要的链接, or firmware, implemented on the Micron Advanced Computing series FPGA modules. The FPGA becomes a customized, efficient, 强大的和可扩展的处理器-针对您的特定工作负载进行优化. picframe提供了一个强大但不可见的主动智能,用于管理FPGA设计的板级实现, as well as data flow, memory management, system communication and monitoring/debug. 

OpenCL Standard

作为我们专注于简化和加速基于fpga的高性能计算系统设计的一部分, Micron Advanced Computing Solutions support the OpenCL standard, 一个并行编程框架,它将类c代码编译到各种计算和处理平台, including FPGAs.

For more information on the OpenCL standard please visit OpenCL Devices and FPGAs or Intel® SDK for OpenCL™ Applications.

Find out more about Advanced Computing Solutions

Contact us
+