Jonathan Reeves

Interesting move and appears a win/win for both companies. That said, may take awhile to implement. I've toured many HPC & AI and enterprise computing datacenters, as well as a number of cryptomining facilities, and the infrastructure is starkly different. HPC & AI datacenters are typically air conditioned, raised floor rooms with direct liquid cooling to chip and high speed (e.g. InfiniBand) interconnect. This is very, very different from the shoeboxes on shelves model, often in former warehouse buildings, with a simple Ethernet connection to the outside world. https://lnkd.in/g-GFNCZ4

12

2 Comments

The unique feature of ULINK DA DriveAnalyzer’s ML system lies in its data collection process. Instead of solely relying on HDD data from a single manufacturer, we aggregate data from multiple HDD manufacturers. #datacenter #hdd #machinelearning #dadriveanalyzer https://lnkd.in/em4ahphp

3

1 Comment

Great article by TIRIAS Research exploring the world of AI PCs. Thanks for the shoutout on how Lattice Semiconductor FPGAs and software are helping today’s top OEMs add intelligence to their latest systems. https://lnkd.in/g_HrsNed

57

The unique feature of DA Drive Analyzer’s ML system compared to the Google model lies in its data collection process. Instead of solely relying on HDD data from a single manufacturer, we aggregate data from multiple HDD manufacturers. #cloud #data #predictiveanalytics https://lnkd.in/eurrwcX3

3

While PCIe fails hard, when you are trying to get everything through the kernels, the new hardware requirements, will make smaller setups to fail horribly without proper RDMA technology at scale. Current NICs are very inefficient, in the transfer scale needed for AI. This calls for another reason to go beyond modern quantization capabilities. This is a problem that will remain persistent, as these new RoCE devices/SuperNics, will have to wait for their economies of scale till they are commodity hardware. NVidia might be coming from Cisco at this rate.... #AI #funding #angelinvesting #Aistartup https://lnkd.in/gUBsheVa

1

#NVMExpress developed the #NVMe Computational Storage feature to enhance NVMe devices and ensure they can manage the exponentially increasing amount of data being processed across #enterprise and #hyperscale #datacenters. Curious to learn how Computational Storage can help you manage data in your data center? Explore the Computational Storage blog post, authored by Bill Martin, Samsung Electronics, and Kim Malone, Intel Corporation: https://bit.ly/3tX4cZs

4

The DA ASUSTOR NAS Application 2.0 is the gateway to AI-driven technology for monitoring the health of storage drives. #asustor #nas #data #storage https://lnkd.in/eiJvjciH

1

Broadcom is stepping up its game in the AI market, challenging NVIDIA's dominance! This move could reshape the landscape of AI technology and innovation. Read about the surge in Broadcom's AI capabilities and what it means for the industry. #AI #ArtificialIntelligence #TechInnovation #Broadcom #Nvidia #TechNews

10

Gen AI has been a boon for infrastructure companies! Broadcom has been one of the beneficiaries of this trend Meta has driven some amazing innovation in Networking for their GPU farms and they leverage merchant silicon from #Broadcom Here are a few bullets from an article in The Register that caught my attention - Hock Tan told investors that revenue from AI-related products "stepped up 280 percent year on year to $3.1 billion, more than offsetting continued cyclical weakness in semiconductor revenue from enterprises and telcos." - #Networking revenue grew 44 percent year on year to reach $3.8 billion, thanks to "strong demand from hyperscalers for both #AI networking and custom accelerators." - "It's interesting to note that as AI datacenter clusters continue to deploy, our revenue mix has been shifting toward an increasing proportion of networking," Tan observed, noting that networking silicon accounted for 53 percent of semiconductor revenue in the quarter. - Part of that growth is due to hyperscalers using Ethernet to connect their GPUs – a trend Tan predicted will mean "Next year, we expect all mega-scaled GPU deployments to be on Ethernet." #GenAI

16

1 Comment

Shedding light on 12 photonics and edge companies you need to know about If you're running large data centers or connecting dispersed locations with high bandwidth connections, photonics are going to be increasing important.  

2

1 Comment

As part of #CRN’s 2024 Storage 100, VirtualZ Computing is featured as a top 50 vendor bringing software capabilities, services and cloud connectivity to storage technology. Joseph Kovar writes, “VirtualZ develops technology that moves data between IBM Z mainframe servers and cloud or on-premises applications and vice versa. The company, which late 2023 exited stealth with $2.2 million in seed funding, is tackling the need for mainframes to run data on the cloud or for cloud applications to run mainframe data without needing custom coding for each instance.” Click here to learn more about Zaac for mainframe to cloud storage: https://lnkd.in/gu78c8EB https://lnkd.in/gaRi-fic #cloudstorage #ibmz #mainframestorage

17

1 Comment

Ioto ™ — IoT at Scale For volume device builders and integrators. If you need to create thousands or millions of connected devices -- Ioto offers (optional) per-device connection to the cloud that can be enabled in the field per device. https://bit.ly/44Cnj9f #aws #iot

4

With Intel’s 4th/5th Gen Xeon processor, Avalue HPC server delivers significant benefits by driving breakthrough discoveries with greater efficiency. Optimized for popular AI models and frameworks, Avalue HPC portfolio helps maximize system performance for new use cases. Key benefits include reducing time to discovery with AI-accelerated HPC, achieving exceptional performance across diverse workloads, accelerating performance and results, and simplifying implementation and scaling of deployments.

19

This is wild technology! During the 56th edition of The IT Press Tour, VMblog met with Michael Buchel of Arc Compute. The company's software suite harnesses low-level #GPU optimizations to accelerate #AI/ML workloads and dramatically reduces #datacenter energy consumption. Learn how their solutions like #ArcHPC Nexus and Oracle can double throughput and slash power usage. https://lnkd.in/gfXdiYPd #GPUs #GPUaaS #ITPT

15

Great article on ArcHPC from VMblog!

7