Andy Patrizio
Author Archives: Andy Patrizio
- Home → Author's archive:
Andy Patrizio
0
IBM leapfrogs everyone with its 2nm chips
As TSMC charges to 5nm transistor designs and Intel struggles for 7nm, IBM has topped them all with the world’s first 2-nanometer node chip.OK, it won’t come to market for four years, according to IBM, and they might not be the first name that comes to mind when you think of processor design, but they are the quiet power in the semiconductor world.[Get regularly scheduled insights by signing up for Network World newsletters.] As far as commercial chips go, IBM makes two: the Power series for its Power line of Unix and Linux servers, and zArchitecture that is used in the z Series of mainframes. But IBM has its IBM Joint Development Alliance which is partnered with just about every semiconductor vendor out there—Intel, AMD, Nvidia, TSMC, Samsung, you name it.To read this article in full, please click here
0
Nvidia competitor Graphcore preps US initiative
A UK-based AI-chip startup is making its first moves into North American to take on Nvidia on its home turf in the enterprise with new channel and reseller partners.Founded in 2016, Graphcore makes what it calls Intelligence Processing Units (IPUs) and shipped its first product—the Colossus GC2 “massively parallel, mixed-precision floating point processor”—in 2018. In July 2020, it released its second-generation processor called GC200, but news of that was drowned out by all the disruption caused by COVID-19.Now see "How to manage your power bill while adopting AI" In addition to chips, the company sells cards and racks. The IPU-M2000 is a 1U blade built around four Colossus GC200 IPU processors, capable of one petaFlop of AI compute. The IPU-POD64 is designed for large-scale deployments and offers the ability to run very large models across as many as 64 IPU processors in parallel.To read this article in full, please click here
0
Nvidia competitor Graphcore preps US initiative
A UK-based AI-chip startup is making its first moves into North American to take on Nvidia on its home turf in the enterprise with new channel and reseller partners.Founded in 2016, Graphcore makes what it calls Intelligence Processing Units (IPUs) and shipped its first product—the Colossus GC2 “massively parallel, mixed-precision floating point processor”—in 2018. In July 2020, it released its second-generation processor called GC200, but news of that was drowned out by all the disruption caused by COVID-19.Now see "How to manage your power bill while adopting AI" In addition to chips, the company sells cards and racks. The IPU-M2000 is a 1U blade built around four Colossus GC200 IPU processors, capable of one petaFlop of AI compute. The IPU-POD64 is designed for large-scale deployments and offers the ability to run very large models across as many as 64 IPU processors in parallel.To read this article in full, please click here
0
Google announces custom video transcoding chip
You know Google has more money than it could ever spend when it invests in a custom chip to do one task. And now they’ve done it for the third time.The search giant has developed a new chip and deployed it in its data centers to compress video content. The chips, called Video (Trans)Coding Units, or VCUs, do that faster and more efficiently than traditional CPUs.In a blog post discussing the project, Jeff Calow, a lead software engineer at Google said the VCU gives the highest YouTube video quality possible on your device while consuming less bandwidth than before.To read this article in full, please click here
0
IBM updates its storage-systems portfolio
IBM announced a pair of additions to its storage portfolio designed to improve the access to and management of data across hybrid-cloud environments and offer faster, higher capacity.The first is container-native software defined storage (SDS) called IBM Spectrum Fusion that’s due out in the second half of 2021. It will initially come in the form of a container-native hyperconverged infrastructure (HCI) system that integrates compute, storage, and networking. Next year, IBM plans to release an SDS-only version of Spectrum Fusion.To read this article in full, please click here
0
IBM updates its storage-systems portfolio
IBM announced a pair of additions to its storage portfolio designed to improve the access to and management of data across hybrid-cloud environments and offer faster, higher capacity.The first is container-native software defined storage (SDS) called IBM Spectrum Fusion that’s due out in the second half of 2021. It will initially come in the form of a container-native hyperconverged infrastructure (HCI) system that integrates compute, storage, and networking. Next year, IBM plans to release an SDS-only version of Spectrum Fusion.To read this article in full, please click here
0
Arm talks 40% and 50% better performance from 2 new server chips
Arm Holdings has disclosed details of its two new server-processor designs, Neoverse N2 and Neoverse V1, as well as an updated high-speed mesh to connect its processors.The two designs were introduced last September but Arm was mum on performance. Now it's talking numbers.The Neoverse V1 is designed for scale-up servers, especially high-performance computing (HPC). It supports for Scalable Vector Extension (SVE) and delivers more than a 50% performance increase over the N1 for HPC machine-learning workloads.To read this article in full, please click here
0
Arm talks 40% and 50% better performance from 2 new server chips
Arm Holdings has disclosed details of its two new server-processor designs, Neoverse N2 and Neoverse V1, as well as an updated high-speed mesh to connect its processors.The two designs were introduced last September but Arm was mum on performance. Now it's talking numbers.The Neoverse V1 is designed for scale-up servers, especially high-performance computing (HPC). It supports for Scalable Vector Extension (SVE) and delivers more than a 50% performance increase over the N1 for HPC machine-learning workloads.To read this article in full, please click here
0
Venerable Linux distro Slackware comes back to life
Being first doesn't guarantee success in the technology industry. Remember the Netscape browser? Still, it can have its advantages, such as a different or unique approach to things.Such is the case with Slackware Linux. Slackware was the first formalized Linux distro, released in 1993, just two years after Linus Torvalds posted the Linux kernel. It was overtaken and overshadowed by Red Hat, SuSe, and Ubuntu, but it never went away. Now it's coming out of the shadows with an upgrade.Slackware creator Patrick Volkerding recently posted a beta version of Slackware 15, the first update to the distro since version 14.2 in 2016. If you think that's ancient, you should see their website.To read this article in full, please click here
0
Siemens and Google Cloud team to deliver AI-based manufacturing solutions
Coming soon to a factory floor near you: Google AI.Google Cloud and Siemens have announced an agreement that calls for Siemens to integrate Google Cloud's AI and machine learning technologies into its factory automation products.Google Cloud's AI/ML capabilities will be combined with Siemens Digital Industries' factory automation portfolio, allowing manufacturers to harmonize their factory data, run cloud-based AI/ML models on top of that data, and deploy algorithms at the network edge. This enables applications such as the visual inspection of products or predicting the wear-and-tear of machines on the assembly line.To read this article in full, please click here
0
Data-center outages: Causes are changing, report says
A new survey by the Uptime Institute found that power issues are becoming less of a problem for data center operators, but networking and software issues are emerging as an increasingly bigger problem.The Uptime Institute's third Annual Outage Analysis notes that while improvements have been made with technology and availability, outages remain a major industry, customer, and regulatory concern. The report also shows that the overall impact and direct and indirect cost of outages continue to grow. When asked about their most recent significant outage, more than half of respondents reported an outage in the past three years and estimated its cost at more than $100,000; among those respondents, almost one-third reported costs of $1 million or above.To read this article in full, please click here
0
Data-center outages: Causes are changing, report says
A new survey by the Uptime Institute found that power issues are becoming less of a problem for data center operators, but networking and software issues are emerging as an increasingly bigger problem.The Uptime Institute's third Annual Outage Analysis notes that while improvements have been made with technology and availability, outages remain a major industry, customer, and regulatory concern. The report also shows that the overall impact and direct and indirect cost of outages continue to grow. When asked about their most recent significant outage, more than half of respondents reported an outage in the past three years and estimated its cost at more than $100,000; among those respondents, almost one-third reported costs of $1 million or above.To read this article in full, please click here
0
Dell reportedly looking to unload its cloud unit
Hot on the heels of its pending spin-off of VMware, Dell is now reportedly looking to sell off its Boomi cloud business for $3 billion. Bloomberg first broke the news, citing people familiar with the matter.Dell is working with a financial adviser on the sale, sources told Bloomberg, and the idea is said to be in the early stages. One thing about Dell, it does not rush these decisions. The VMware spin-off, which has been discussed for years, started last summer and only now is coming to fruition.Boomi provides an integration platform as a service (iPaaS), which enables the connection of applications and data sources. The platform provides API, lifecycle management, and event-driven architecture features for cloud integration. Dell acquired Boomi in 2010.To read this article in full, please click here
0
Verizon launches 5G fixed wireless in parts of 21 more cities
Communications giant Verizon last week launched 5G for Business Internet in 21 new markets, targeting SMBs and enterprises alike. 5G resources What is 5G? Fast wireless technology for enterprises and phones How 5G frequency affects range and speed Private 5G can solve some problems that Wi-Fi can’t Private 5G keeps Whirlpool driverless vehicles rolling 5G can make for cost-effective private backhaul CBRS can bring private 5G to enterprises The fixed-wireless plans provide download speeds of 100Mbps ($69/month), 200Mbps ($99/month), and 400Mbps ($199/month) with no data limits. Upload speeds are slower. Verizon is also offering a 10-year price lock for new customers with no long-term contract required.To read this article in full, please click here
0
Round-up of Nvidia GTC data-center news
With a few dozen press releases and blog posts combined, no one can say that Nvidia’s GPU Technology Conference (GTC) is a low-key affair. Like last year’s show it is virtual, so many of the announcements are coming from CEO Jen-Hsun Huang’s kitchen.Here is a rundown of the most pertinent announcements data-center folks will care about.Two Ampere 100 offshoots Nvidia's flagship GPU is the Ampere A100, introduced last year. It is a powerful chip ideal for supercomputing, high-performance computing (HPC), and massive artificial intelligence (AI) projects, but it’s also overkill for some use cases and some wallets.So at GTC the company introduced two smaller scale little brothers for its flagship A100, the A30 for mainstream AI and analytics servers, and the A10 for mixed compute and graphics workloads. Both are downsized from the bigger, more powerful, and more energy-consuming A100.To read this article in full, please click here
0
AT&T picked as top managed-SD-WAN provider for the third year
AT&T, Hughes, and Verizon were selected as the top three SD-WAN providers for the third year in a row in the latest Vertical Systems Group rankings for year-end 2020. Comcast jumped to fourth place.Despite the pandemic, expansion of carrier-managed SD-WAN services in the U.S. increased 39% in 2020. Demand was resilient across bandwidth-intensive markets, but vulnerable for verticals like retail and travel.Vertical Systems Group is an independent market research firm focused on network services. Each year it issues its Carrier Managed SD-WAN Services Leaderboard.The one notable change was Comcast has replaced Lumen Technologies in fourth place, moving up from seventh position in 2019. Lumen is now in sixth, Windstream remains in fifth and Aryaka dipped from sixth to seventh.To read this article in full, please click here
0
Nvidia announces a 2023 launch for an HPC CPU named Grace
Nvidia kicked off its GPU Technology Conference (GTC) 2021 with a bang: A new CPU for high performance computing (HPC) clients--its first-ever data-center CPU--called Grace.Based on the Arm Neoverse architecture, NVIDIA claims Grace will serve up to 10-times better performance than the fastest servers currently on the market for complex artificial intelligence and HPC workloads.But that’s comparing then and now. Grace won’t ship until 2023, and in those two years competitors will undoubtedly up their game, too. But no one has ever accused CEO Jen-Hsun Huang of being subdued.Nvidia made a point that Grace is not intended to compete head-to-head against Intel's Xeon and AMD's EPYC processors. Instead, Grace is more of a niche product, in that it is designed specifically to be tightly coupled with NVIDIA's GPUs to remove bottlenecks for complex AI and HPC applications.To read this article in full, please click here
0
Nvidia announces a 2023 launch for an HPC CPU named Grace
Nvidia kicked off its GPU Technology Conference (GTC) 2021 with a bang: A new CPU for high performance computing (HPC) clients--its first-ever data-center CPU--called Grace.Based on the Arm Neoverse architecture, NVIDIA claims Grace will serve up to 10-times better performance than the fastest servers currently on the market for complex artificial intelligence and HPC workloads.But that’s comparing then and now. Grace won’t ship until 2023, and in those two years competitors will undoubtedly up their game, too. But no one has ever accused CEO Jen-Hsun Huang of being subdued.Nvidia made a point that Grace is not intended to compete head-to-head against Intel's Xeon and AMD's EPYC processors. Instead, Grace is more of a niche product, in that it is designed specifically to be tightly coupled with NVIDIA's GPUs to remove bottlenecks for complex AI and HPC applications.To read this article in full, please click here
0
How to shop for a colocation provider
If you want to move assets out of your data center but for whatever reason can’t shift to the cloud, a colocation, or “colo” for short, is increasingly a viable option.Colo is where the client buys the compute, storage, and networking equipment but instead of putting it into their own data centers, they put them in the data center of a hosting company. They still own and manage the hardware, but they don’t have responsibility for manage the facilities—heating, cooling, lighting, physical security, etcNow see "How to manage your power bill while adopting AI" As such, colocation facilities attract considerable interest from enterprises. IDC puts the 2020 US colocation market at $9 billion, growing to $12.2 billion by 2024 for a compound annual growth rate (CAGR) of 8%. Grand View Research estimates the global data-center colocation market size was valued at $40.31 billion US dollars in 2019 and is expected to grow at a CAGR of 12.9% from 2020 to 2027. Gartner makes the bravest prediction, saying that by 2025, 85% of infrastructure strategies will integrate on-premises, colocation, cloud, and edge delivery options, compared with 20% in 2020.To read this article in full, please click here
0
Samsung demos 512GB DDR5 memory aimed at supercomputing, AI workloads
Samsung Electronics last month announced the creation of a 512GB DDR5 memory module, its first since the JEDEC consortium developed and released the DDR5 standard in July of last year.The new modules are double the max capacity of existing DDR4 and offer up to 7,200Mbps in data transfer rate, double that of conventional DDR4. The memory will be able to handle high-bandwidth workloads in applications such as supercomputing, artificial intelligence, machine learning, and data analytics, the company says. Read more: World's fastest supercomputers: Fugaku still No. 1To read this article in full, please click here