Andy Patrizio is a freelance technology writer based in Orange County, California. He's written for a variety of publications, ranging from Tom's Guide to Wired to Dr. Dobbs Journal, and has been on staff at IT publications like InternetNews, PC Week and InformationWeek.
Nvidia’s Grace processor will be one of the first chips to use Arm’s upcoming Neoverse V2 CPU cores.
IBM LinuxONE Emperor 4 server uses the System Z processor and boasts better performance vs. x86 and greater power efficiency.
Benchmarks aren't the same as real world use, but they can give a good idea of what’s to come, and Nvidia's Hopper GPU performance is impressive.
A lawsuit claiming license violations by Qualcomm and Nuvia could hobble production of processors based on Arm technology.
IBM offers a flat-fee subscription for single-core System i servers.
Microsoft adjusts Windows Server fees after threats of lawsuits by European cloud service providers.
New vSphere paired with Nvidia DPUs will speed up data center performance.
Deal with Brookfield Asset Management brings more chip production to the US and frees up more Intel cash.
If Qualcomm is working on a server processor, it likely will be for accelerators such as smartNICs.
Verge-OS virtualization software can now pool and apportion GPU capacity to help create virtual data centers.
Efficiencies and better hardware prompt Azure to extend server-refresh cycles from four years to six.
Compute Express Link can share compute and memory among components and devices, potentially leading to more efficient use of data-center resources.
100Gb+ will be standard for high-performance computing, AI, and cloud providers.
Semtech has been working on IoT technology, and buying Sierra will accelerate that effort.
The small form factor HPE Edgeline EL8000 is designed for AI tasks such as computer vision and natural-language processing.
Enterprise SSD products are separating themselves from their consumer counterparts by getting faster, smarter and denser.
Optane persistent memory technology was never adopted beyond Intel.
PowerStore performance gains include faster Intel processors, greater scalability, and tools to simplify NVMe capacity expansion.
New Power10 servers also include pay-as-you-go consumption options.
The goal is to make it easier to program in quantum computing, which is very different from standard computing.
HPE becomes the first major server vendor to ship servers with the Arm-based processor.
The old SATA interface may be much slower than PCIe but it is still useful in many use cases.
Cloud performance issues can force workloads back on-premises and prevent some from getting to the cloud in the first place.
PCIe continues to double its data rate, targeting 800G Ethernet and cloud computing.
New enterprise storage system can connect NVMe SSDs and HDDs on one platform, which means users don’t have to maintain separate all-flash and hard-disk arrays.
A new AMD accelerated processing unit will deliver 8-times the performance of its current top ACU that powers the world's first exascale supercomputer.
AIRI//S is a scalable AI infrastructure that integrates Pure’s scale-out storage with Nvidia compute systems.
Intel says its reference design will enable accelerator cards to offload security processing from server CPUs without requiring more rack space.
Next generation of Ampere’s server processors features customized Arm cores and is aimed at cloud data centers.
At the International Supercomputer Conference, processor rivals AMD, Intel, and Nvidia talk up recent successes, future plans.
Cloud services provider Vultr is offering fractional Nvidia A100 GPU instances for customers that don’t need the full power of more expensive options.
Nvidia has four new compute-node reference designs for for demanding tasks including AI training, high-performance computing, digital twin modeling, and cloud graphics.
IBM says the ESS 3500 storage devices will improve AI training by up to 70%.
Snowflake has always worked in the cloud, but some data warehouses are on premises, and that needed to be addressed.
Storage giant updates its high-capacity hard drives and SSDs and debuts a new portable storage device.
Cisco says manufacturing errors are to blame for flaws in its 16GB, 32GB, and 64GB dual in-line memory modules (DIMM).
AMD CEO says, ‘You should see us investing a lot more in software.’
Intel CEO linked ongoing chip shortage to a lack of manufacturing equipment and difficulty building semiconductor fabrication plants.
The scale-out servers are descendants of the AS/400 mid-range server dating back to the 80s.
As enterprises move beyond pilot stages of cloud use, they will spend close to $500B worldwide this year, Gartner projects.
Arm's Cortex-M85 micro controller will support internet of things devices while the Corstone-1000 supports AI at the edge.
The Arm processor isn’t setting the world on fire on regular servers, but it does seem to have found a home in the cloud.
A partnership between VMware and Deloitte is designed to help shepherd enterprises through the process of deploying distributed-cloud architectures.
Fujitsu Cloud Service HPC will cater to demanding workloads and provide consulting to tune the service to individual customers.
HPE supercomputer on the International Space Station processes data on site, sorts, and compresses it to save time sending to Earth.
Intel's Blockscale ASIC designed specifically for blockchain hash operations with greater power efficiency is available this fall.
Enterprises may not need supercomputers, but regularly upgrading the computing gear they do use can improve speed and power consumption.
It will take a few years, and a lot of negotiation, but the thawing relations between Intel and Nvidia could help ease processor supply issues.
Nvidia's Grace superchips cater to AI as well as legacy high-bandwidth applications not optimized for GPUs.
Nvidia’s H100 GPU is the first in its new family of Hopper processors that the company claims will underpin the world’s fastest supercomputer.
Sponsored Links