Intel Xeon Phi Coprocessor Design

'@Intel Xeon Phi Processor Design.' #BigData #DeepLearning #MachineLearning #DataScience #AI

  • The major functionality of the Intel Xeon Phi coprocessor is a chip that does the heavy computation.
  • To communicate with the host, the Intel Xeon Phi coprocessor resides on the PCI Express bus, and is designed for the Gen 2 x16 standard.
  • In order for the high number of cores that are present on the Intel Xeon Phi coprocessor, a high speed ring based bidirectional on-die interconnect (ODI) is used.
  • Each core has 512 KB of Level 2 cache, which stores data that might be reused.
  • A coprocessor card needs to be designed to physically fit into an existing server must contain a very high number of cores.

A coprocessor card needs to be designed to physically fit into an existing server, yet must contain a very high number of cores. Whereas on a main CPU design, the cores may take up significant real estate, designers must cram in over 50 cores on a smaller motherboard. In addition, the motherboard must be able to hold GDDR5 memory, memory controllers and miscellaneous electronics in order to communicate with the system CPU, but also perform floating point calculations in parallel.
Continue reading “Intel Xeon Phi Coprocessor Design”

Intel acquires deep learning startup Nervana for more than $350 million

Intel acquires deep learning startup Nervana  by @jordannovet

  • Nervana was one of my five deep learning startups to watch in 2015 .
  • Nervana open-sourced its Neon deep learning software last year.
  • While Intel has a running business in high-performance computing (HPC), it has taken a back seat to Nvidia, another HPC supplier, when it comes to creating chips for deep learning, a trendy type of artificial intelligence that involves training artificial neural networks on lots of data and then getting them to make inferences on new data.
  • It’s unclear if Google has joined with Intel in that affair.
  • Today’s deal follows Salesforce’s acquisition of MetaMind , another prominent deep learning startup.

Read the full article, click here.


@VentureBeat: “Intel acquires deep learning startup Nervana by @jordannovet”


Chipmaker Intel today announced that it has acquired Nervana, a startup that has been developing artificial intelligence software and hardware. Terms of the deal weren’t disclosed, but a source familiar with the matter told VentureBeat Intel paid more than $350 million.


Intel acquires deep learning startup Nervana for more than $350 million

ISC High Performance 2016

Intel #XeonPhi is latest technology for high-performance computing & machine learning  #ISC16

  • Now in its 31st year, ISC High Performance is the world’s oldest and Europe’s most important conference and networking event for the high performance computing community.
  • Intel’s first bootable host processor is specifically designed for highly parallel workloads.
  • A bootable x86 CPU, the Intel Xeon Phi processor offers greater scalability and is capable of handling a wider variety of workloads and configurations than accelerator products.
  • It is also the first to integrate both memory and fabric technologies.
  • Among this year’s new products is the Intel® Xeon Phi™ processor.

Read the full article, click here.


@intelnews: “Intel #XeonPhi is latest technology for high-performance computing & machine learning #ISC16”


Now in its 31st year, ISC High Performance is the world’s oldest and Europe’s most important conference and networking event for the high performance computing community. At this year’s show, Intel will introduce and showcase a range of new technologies helping to fuel the path to deeper insight and HPC’s next frontier.


ISC High Performance 2016