China Tunes Neural Networks for Custom Supercomputer Chip

Supercomputing centers around the world are preparing their next generation architectural approaches for the insertion of AI into scientific workflows. For some, this means retooling around an existing architecture to make capability of double-duty for both HPC and AI.

Teams in China working on the top performing supercomputer in the world, the Sunway TaihuLight machine with its custom processor, have shown that their optimizations for theSW26010 architecture on deep learning models have yielded a 1.91-9.75X speedup over a GPU accelerated model using the Nvidia Tesla K40m in a test convolutional neural network run with over 100 parameter configurations.

Efforts on

China Tunes Neural Networks for Custom Supercomputer Chip was written by Nicole Hemsoth at The Next Platform.

Who Controls The Internet?

The title of the paper Who controls the Internet? Analyzing global threats using property traversal graphs is enough to ensnare any Internet researcher. The control plane for a number of attacks, as the paper points out, is the DNS due to the role it plays in mapping names to resources. MX records in the DNS control the flow of mail, CNAME records are used to implement content delivery networks (CDN) services, and TXT records are used to confirm access to and control over a namespace when implementing third party services. This post will cover an interesting case where control is exercised first via the DNS and then using BGP.

Below the DNS, in the depths of internet plumbing, is the lizard brain of internet routing, which is governed by the border gateway protocol (BGP). A common term to describe BGP routing is “hot potato” routing. BGP conversations occur between autonomous systems, ASes, which are identified by their autonomous system number ASN. The ASN represents a system of networks and the policy associated with their routing. ASes are issued regionally by Regional Internet Registries (RIRs), which receive blocks of AS numbers to hand out from the Internet Assigned Numbers Authority Continue reading

33% off iRobot Roomba 652 Robotic Vacuum Cleaner – Prime Day Deal Alert

This is a Prime Day deal good for today only, and reserved for Amazon Prime members (or those with a free trial: sign up here). The Roomba 652 Vacuuming Robot provides a thorough clean at the push of a button. Preset Roomba to clean when it’s convenient for you, so you can keep up with everyday mess. The Roomba 652 is discounted a 33%, so you save a whopping $125 if you buy it today. If you've always wanted a robot to clean your house, see this deal on Amazon.To read this article in full or to leave a comment, please click here

35% off Bose SoundLink Mini Bluetooth Speaker II – Prime Day Deal Alert

Amazon Prime members save a generous 35% ($70) on the highly rated Bose SoundLink Mini Bluetooth Speaker II. This is a Prime Day deal, so good for today only until they run out of stock. The SoundLink Mini Bluetooth speaker II delivers full sound with dramatically deeper lows than you’d expect from a speaker that fits in the palm of your hand. And because it’s wireless and ultra-compact, it’s easy to take anywhere. The built-in speakerphone lets you take calls out loud with clear sound. And voice prompts make Bluetooth pairing easy. Plays unplugged for hours and can be charged from most USB power sources. Head over to Amazon and take advantage of this Prime Day deal. If you're not a Prime member, you can sign up for a free trial here to unlock the deals. To see our picks for today's best Prime Day deals, read our guide (on our sister site PCWorld.com) to the best deals: "Amazon Prime Day 2017: We pick the best electronics, PC, and mobile deals"To read this article in full or to leave a comment, please click here

OpenPower, Efficiency Tweaks Define Europe’s DAVIDE Supercomputer

When talking about the future of supercomputers and high-performance computing, the focus tends to fall on the ongoing and high-profile competition between the United States with its slowly eroding place as the kingpin in the industry and China and the tens of billions of dollars that the government has invested in recent years to rapidly expand the reach of the country’s tech community and the use of home-grown technologies in massive new systems.

Both trends were on display at the recent International Supercomputing Conference in Frankfurt, Germany, where China not only continued to hold the top two spots on the

OpenPower, Efficiency Tweaks Define Europe’s DAVIDE Supercomputer was written by Nicole Hemsoth at The Next Platform.

IPv6 Link-Local Addresses and VLAN Interfaces

One of my readers sent me an email that’s easiest paraphrased into: “Why can’t I have a different IPv6 link-local address (LLA) on every access port connected to a VLAN interface?

There’s probably nothing stopping someone from implementing such an approach, but it would go against the usual understanding of how bridging and routing interact in L2+L3 switches.

Read more ...

Interface basics on the Juniper MX

I’ve been spending more time on the MX recently and I thought it would be worthwhile to document some of the basics around interface configuration.  If you’re like me, and come from more of a Cisco background, some of configuration options when working with the MX weren’t as intuitive.  In this post, I want to walk through the bare bone basic of configuring interfaces on a MX router.

Basic L3 interface

ge-0/0/0 {
    unit 0 {
        family inet {
            address 10.20.20.16/24;
        }
    }
}

The most basic interface configuration possible is a simple routed interface. You’ll note that the interface address is configured under a unit. To understand what a unit is you need to understand some basic terminology that Juniper uses. Juniper describes a physical interface as an IFD (Interface Device). In our example above the IFD would be the physical interface ge-0/0/0. We can then layer one or more IFL (Interface Logical) on top of the IFD. In our example the IFL would be the unit configuration, in this case ge-0/0/0.0. Depending on the configuration of the IFD you may be able to provision additional units. These additional units (Logical interfaces (IFLs)) Continue reading

Docker for the SysAdmin Webinar Q&A

On June 27th I presented a webinar on “Docker for the SysAdmin”.  The webinar was driven by a common scenario I’m seeing: A sysadmin is sitting at her desk minding her own business when a developer walks in and says “here’s the the new app, it’s in a Docker image. Please deploy it ASAP”. This session is designed to help provides some guidance on how sysadmins should think about managing Dockerized applications in production.

In any case, I was a bit long-winded (as usual), and didn’t have time to answer all the Q&A during the webinar (and there were quite a few).

So, as promised, here are all the questions from that session, along with my answers.  If you need more info, hit me up on Twitter: @mikegcoleman

————

Q: I am planning an application deployment and want to use Docker. What cloud would you recommend at the moment? I have GCP, Azure, AWS under my belt. 1) TCO 2) Performance ?

A: Answering that would require me to understand your application on a pretty deep level, so I can’t really provide a specific response. I will say that if you choose one cloud provider today, and realize that Continue reading

Google, IBM look to mimic the human brain

Several years ago, there were reports that an IBM artificial intelligence (AI) project had mimicked the brain of a cat. Being the smartass that I am, I responded on Twitter with, “You mean it spends 18 hours a day in sleep mode?”That report was later debunked, but the effort to simulate the brain continues, using new types of processors far faster and more brain-like than your standard x86 processor. IBM and the U.S. Air Force have announced one such project, while Google has its own.+ Also on Network World: Machine learning proves its worth to business + Researchers from Google and the University of Toronto last month released an academic paper titled “One Model To Learn Them All,” and they were pretty quiet about it. What Google is proposing is a template for how to create a single machine learning model that can address multiple tasks.To read this article in full or to leave a comment, please click here

Google, IBM look to mimic the human brain

Several years ago, there were reports that an IBM artificial intelligence (AI) project had mimicked the brain of a cat. Being the smartass that I am, I responded on Twitter with, “You mean it spends 18 hours a day in sleep mode?”That report was later debunked, but the effort to simulate the brain continues, using new types of processors far faster and more brain-like than your standard x86 processor. IBM and the U.S. Air Force have announced one such project, while Google has its own.+ Also on Network World: Machine learning proves its worth to business + Researchers from Google and the University of Toronto last month released an academic paper titled “One Model To Learn Them All,” and they were pretty quiet about it. What Google is proposing is a template for how to create a single machine learning model that can address multiple tasks.To read this article in full or to leave a comment, please click here

What’s the difference between SDN and NFV?

SDN, NFV & VNF are among the alphabet soup of terms in the networking industry that have emerged in recent years.Software defined networking (SDN), network function virtualization (NFV) and the related virtual network functions (VNF) are important trends. But Forrester analyst Andre Kindness says vague terminology from vendors has created a complicated marketplace for end users evaluating next-generation networking technology. “Few I&O pros understand (these new acronyms), and this confusion has resulted in many making poor networking investments,” he says.So what’s the difference between SDN, NFV and VNF?SDN: Software defined networking To read this article in full or to leave a comment, please click here