IBM today unveiled what it said is the first chip with 2 nanometer (nm) nanosheet technology. Though still more than two years away from going into production, the company said the 2nm chip will be used to power devices ranging from cell phones to HPC-class data center servers and will deliver nearly 50 percent higher […]
Rice Univ. Researchers Claim 15x AI Model Training Speed-up Using CPUs
April 23, 2021 by 1 Comment
Reports are circulating in AI circles that researchers from Rice University claim a breakthrough in AI model training acceleration – without using accelerators. Running AI software on commodity x86 CPUs, the Rice computer science team say neural networks can be trained 15x faster than platforms utilizing GPUs. If valid, the new approach would be a double boon for organizations implementing AI strategies: faster model training using less costly microprocessors.
Super-expensive Supercomputers: UK Met Office in £1B+ Deal for Microsoft Weather System; $3B System in China on Way
April 22, 2021 by Leave a Comment
As widely reported in February, the UK Met Office and Microsoft have come to an agreement to provision a £1.2 billion (over 10 years) supercomputer the Met Office said will be the world’s most powerful weather and climate forecasting system. The agreement, announced on Earth Day, will be in the top 25 of the Top500 […]
‘Ice Lake’ Launch: Intel Impresses with 3rd Gen Xeon Chips
April 6, 2021 by Leave a Comment
In its cage-match struggle with AMD for data center (and HPC) CPU market superiority, Intel responded this morning to its rival’s vaunted EPYC 7003 Series CPUs, announced two weeks ago, with its 3rd Gen Intel Xeon Scalable data center chips, code named “Ice Lake.” And the early response from a small sampling of industry analysts […]
Silicon’s Successors: 2-D Materials Aim to Carry the Post-Moore’s Law Torch
April 5, 2021 by Leave a Comment
Even as a flurry of new silicon has been launched recently by AMD and Arm, with more to come this week from Intel, silicon’s five decades of microprocessor dominance could start to fade in the relatively near future. The problem is that transistors etched into silicon will soon be shrunk to 5 nanometers, with 3 […]
PLDA Joins ETP4HPC to Share High Speed Interconnect IP with European HPC Ecosystem
March 30, 2021 by Leave a Comment
AIX-EN-PROVENCE, France — PLDA, the leading developer of high-speed interconnect silicon IP, is joining ETP4HPC, an association that gathers leading European players in high performance computing technologies. ETP4HPC – the European Technology Platform (ETP) for High-Performance Computing (HPC) – is a private, industry-led and non-profit association whose main mission is to promote European HPC research […]
Peptone Scales Protein Engineering System with Nvidia and Verne Global HPC Colo
March 30, 2021 by Leave a Comment
March 30, 2021 – Molecular computational physics company Peptone, focused on protein drug discovery and engineering tools, has installed an Nvidia DGX A100 system at Verne Global’s HPC-optimized data center campus in Keflavik, Iceland. The company also announced a collaboration with Nvidia to scale its AI-driven protein engineering system in order to meet increasing demand from […]
DOE Announces $29 Million for Data Analysis Tools
March 26, 2021 by Leave a Comment
March 26, 2021 – The U.S. Department of Energy (DOE) today announced $29 million to develop new tools to analyze massive amounts of scientific information, including artificial intelligence, machine learning and advanced algorithms. All have the potential to reveal critical new insights—and even new discoveries—in research that can help tackle clean energy, climate, and national […]
AiDANT Picks Core Scientific for GPU-as-a-Service
March 25, 2021 by Leave a Comment
BELLEVUE, Wash., March 25, 2021 – Core Scientific, an infrastructure and software solutions provider for artificial intelligence and blockchain, announced it will provide AiDANT, an AI computer vision technology company, full access to high performance NVIDIA RTX servers in an as-a-service model. AiDANT leverages deep learning to notify users when a certain condition is met […]
Samsung: First HKMG-Based 512GB DDR5 Memory Targets Bandwidth-Intensive Applications
March 24, 2021 by Leave a Comment
Samsung today announced it has expanded its DDR5 DRAM memory portfolio with what the company said is the first 512GB DDR5 module based on High-K Metal Gate (HKMG) process technology, designed to support next-gen computing systems, supercomputers and large-scale data centers “Delivering more than twice the performance of DDR4 at up to 7,200 megabits per second […]
