Viewing posts from October, 2016

Tech Companies Sign On to New Standard for High Performance Server Bus

IBM has unveiled OpenCAPI, an open-standard, high-speed bus interface for connecting devices in servers. The announcement coincides with the formation of a consortium of the same name that will manage the new standard, and which initially includes tech heavyweights Hewlett Packard Enterprise (HPE), Dell EMC, NVIDIA, Mellanox, Micron, Xilinx, and Google. The first OpenCAPI-supported devices and servers are expected to show up in 2017.

Intel ARMs Its FPGAs; Cray Ponders Deep Learning Supercomputer

Episode 144: Addison Snell and Michael Feldman share their thoughts on Intel's latest chip line and Cray's move into deep learning.

Watson, AI Get Spotlight on “60 Minutes”

Artificial intelligence received a nice boost in public exposure last Sunday evening when “60 Minutes,” CBS’s iconic news magazine, aired an in-depth report on the subject. The half-hour segment described how systems like IBM’s Watson supercomputer is changing the nature of medical diagnostics, and how other AI technologies are working their way into our everyday lives.

Intel Expands Its Comfort Zone with New ARM-Powered FPGAs for Datacenters

Intel announced it is sampling its Stratix 10 FPGAs, the latest family of field programmable gate arrays that are designed to accelerate a number of datacenter workloads. The new devices, which Intel is calling “the most significant FPGA innovations in over a decade,” offer advanced features like embedded 64-bit ARM processors, second-generation High Bandwidth Memory (HBM2), and DSP blocks.

Cray Has Deep Learning Supercomputer in the Works

As HPC vendors like IBM and HPE have added deep learning-optimized systems to their product portfolio, Cray has been more circumspect about its plans in this area. We talked with Cray CTO Steve Scott about how they view this new application area of high performance computing and what they may have in the pipeline to serve this burgeoning market.

Adapteva Reaches Thousand Core Epiphany; More GPUs in the Cloud

Episode143: Addison Snell and Michael Feldman discuss Adaptiva's breakthrough technology and cloud computing's more robust offerings.

Berkeley Lab Builds One-Nanometer Transistor

As Moore’s Law reaches its physical limits with silicon-based semiconductor technology, researchers are turning to more exotic materials to extend its life a few more years. A team of scientists at Lawrence Berkeley National Laboratory (aka Berkeley Lab) did just that by inventing a one-nanometer (1nm) transistor.

EPCC Offers Cut-Rate Supercomputing Cycles to Businesses

The Edinburgh Parallel Computing Centre (EPCC) is renting out its new 2,016-core Cirrus supercomputer for businesses in need of on-demand HPC. The system, an SGI ICE XA cluster, is being offered to commercial firms for applications in automotive, aerospace, energy, oil and gas, general engineering, life sciences and financial services for what appears to be a very reasonable price.

Adapteva Tapes Out 1024-Core Superchip

This week, Adapteva CEO Andreas Olofsson announced his company has taped out the Epiphany-V, a 1024-core RISC chip that will likely offer the world’s best performance per watt for double precision arithmetic. This SoC is geared for high performance computing (HPC) in applications such as self-driving cars, autonomous drones and cognitive radio, where performance-per-watt is the most critical attribute.

Microsoft Rolls Out FPGAs Across Cloud; NVIDIA Builds Momentum in Europe

Episode 142: Addison Snell and Michael Feldman dissect the week's top HPC stories.