Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Ayar Labs Joins DARPA PIPES Project as Intel Optical IO Provider

Optical startup Ayar Labs has been selected as Intel’s optical I/O solution partner for their recently awarded DARPA PIPES research project. “The goal of PIPES (Photonics in Package for Extreme Scalability) is to develop integrated optical I/O solutions co-packaged with next generation FPGA/CPU/GPU and accelerators in Multi-Chip Packages (MCP) to provide extreme data rates (input/output) at ultra-low power over much longer distances than supported by current technology. In the first phase of the project, the Ayar Labs TeraPHY chiplet will be co-packaged with an Intel FPGA using the AIB (Advanced Interconnect Bus) interface and Intel’s EMIB silicon-bridge packaging.”

DARPA FastNICs Program Looks to Accelerate Application Performance by 100x

DARPA is looking to create new networking approaches to accelerate distributed application performance by 100x with the FastNICs program. “FastNICs seeks to improve network stack performance by a factor of 100 through the creation of clean-slate networking approaches. Enabling this significant performance gain will require a rework of the entire network stack – from the application layer through the system software layer, down to the hardware.”

AMD: Delivering the Future of High-Performance Computing

Dr. Lisa Su from AMD gave this talk at the recent DARPA Electronics Resurgence Initiative Summit. “Optimum system performance requires co-design of silicon chips, system architecture, and software. She presented the example of the Frontier exascale computer system being developed for Oak Ridge National Lab, which should exhibit 1.5 exaflops by 2021. While the highest-performance chips and systems will initially be limited to the most expensive machines, it is expected that similar technology will become available within a few years in data centers, edge computers, and even mobile devices.”

New Funding and DARPA Grant to Propel Optical Interconnects at Ayar Labs

Today Ayar Labs announced that the company has secured additional funding to fuel its growth as it drives to productize its TeraPHY optical I/O chiplets and SuperNova multi-wavelength lasers in 2019. The company aims to disrupt the traditional performance, cost, and efficiency curves of the semiconductor and computing industries by driving a” 1000x improvement” in interconnect bandwidth density at 10x lower power.

DARPA Selects Research Projects for Next-Gen Processing Technology

DARPA (Defense Advanced Research Projects Agency) has announced the academic and industry research groups selected to develop new computing technologies to drive computing performance post Moore’s Law. “These projects hope to overcome one of the fundamental performance bottlenecks facing HPC users the ‘memory bottleneck’. By setting researchers the task of investigating vertical, rather than flat or planar integration of microsystem components—as well as new materials, components, and algorithms capable of closing the gap between memory and logic functions—the program managers leading the 3DSoC and FRANC programs hope to create new means of computing vast amounts of information.”

Radio Free HPC Reviews Lincoln Labs Paper on Spectre/Meltdown Performance Hits

In this podcast, the Radio Free HPC team looks at a new whitepaper from Lincoln Labs focused on the performance hits Spectre/Meltdown mitigations. The news is not good. After that, Shahin point us to the story about how DARPA just allocated $75 Million in awards for thinking-outside-the-box computing innovation. They call it the Electronics Resurgence Initiative and the list of projects funded includes something called Software Defined Hardware.

Intel to Develop New Machine Learning and AI Platform for DARPA

Today Intel announced that it has been selected by DARPA to collaborate on the development of a powerful new data-handling and computing platform that will leverage machine learning and other artificial intelligence (AI) techniques. “By mid-2021, the goal of HIVE is to provide a 16-node demonstration platform showcasing 1,000x performance-per-watt improvement over today’s best-in-class hardware and software for graph analytics workloads,” said Dhiraj Mallick, vice president of the Data Center Group and general manager of the Innovation Pathfinding and Architecture Group at Intel. “Intel’s interest and focus in the area may lead to earlier commercial products featuring components of this pathfinding technology much sooner.”

Video: DARPA’s SyNAPSE and the Cortical Processor

“I will describe a decade-long, multi-disciplinary, multi-institutional effort spanning neuroscience, supercomputing and nanotechnology to build and demonstrate a brain-inspired computer and describe the architecture, programming model and applications. I also will describe future efforts in collaboration with DOE to build, literally, a “brain-in-a-box”. The work was built on simulations conducted on Lawrence Livermore National Laboratory’s Dawn and Sequoia HPC systems in collaboration with Lawrence Berkeley National Laboratory.”