Jakub Polec’s Post

Google to Use RISC-V for Its Custom AI Silicon. SiFive expects to earn $240-$280 million in licensing revenue and new lifetime contracts in 2024, according to Bloomberg. This may indicate that the company is on track to sign another major contract with Google to supply processor core designs for its tensor processing units (TPUs). SiFive pins a lot of hope on a 'second-generation chip for artificial intelligence servers.' Google's current-generation TPU already uses SiFive's X280 general-purpose core. #RISCV #SiFive #CustomSilicon #AI #CustomAIsilicon #Google #TPU #TensorProcessingUnit

GPUs have been all the rage regarding handling AI workloads for training and inference. What often gets overlooked is, that shuttling data within and across data centers is equally important to leverage the true compute potential of new-age GPUs. In a nutshell, your network is the computer and key enabler in scaling. Broadcom has been at the forefront of enabling an open, scalable, and power-efficient approach towards AI infrastructure. Read more - 👇

#io.net and #GAIB have partnered to #democratize access to #GPU computing for #AI, leveraging io.net’s #decentralized GPU infrastructure and GAIB’s expertise in GPU #tokenization. This #collaboration aims to make high-performance GPU resources accessible to a wider #audience by integrating their #technologies and sharing resources. The #partnership will focus on initiatives like compute tokenization, allowing GPU owners to convert #hardware into yield-bearing assets, reducing compute costs for #businesses and #individuals. Plans include technical integration, joint #marketing, and #community engagement to boost #visibility and #adoption of GPU resources. Leaders from both companies, Tausif Ahmed (io.net) and Kony Kwong (GAIB), expressed excitement about expanding access and #innovation in AI compute markets. https://lnkd.in/gbuXzc3c

Link flapping is a developing and costly issue in large GPU clusters – with each link flap event resulting in up to 30 minutes of lost training time and costing up to $200,000! Thanks to Massine Merzouk of X for recognizing how important it is that Credo is taking the lead in addressing this issue. "Cluster reliability is of paramount importance when building the biggest supercomputers in the world with 100,000+ GPUs,” said Massine Merzouk, Network Engineer at X, assisting xAI. “Credo’s HiWire AECs offer the stable transport platform we need to build such massive systems." https://bit.ly/3zPhZUV

Large Language Models (LLMs) are absolute powerhouses when it comes to generating top-notch #GenAI responses. Yet, #LLMs lack that in-depth, domain-specific knowledge that can make all the difference in certain use cases. Additionally, let's not forget the hefty computing resources and expertise required to develop and deploy them. In my latest article, I delve into why Small Language Models (SLMs) are becoming the go-to choice for specific tasks. With their more focused approach, #SLMs are more efficient than LLMs, and have a smaller compute footprint. Thus, they are more suitable to run locally on workstations or on-prem servers. Check out my article 👇 and ‘d love to hear your thoughts about it https://lnkd.in/diRYaNXF #DellTechEMEA #iWork4Dell #AMD, #Intel, #NVIDIA 

In the era of AI, where large amounts of sensitive data are collected and utilized, security is paramount. As discussions about solutions take place across industries and governments, Richard Grisenthwaite sheds light on the significance of foundational security and standards to address emerging challenges. With AI inference workloads at the edge set to grow significantly, Richard provides insights into our security technologies and initiatives in place today that are supporting the world’s AI-based experiences. These include: 🔒 Memory Tagging Extension, which is integrated into our latest v9 CPUs 🛡️ Realm Management Extensions, which forms the basis of the Arm Confidential Compute Architecture 🤝 Our ongoing partnership with PSA Certified to establish standardized frameworks and certifications 🔬 The Morello Program, which explores new ways to deter security breaches through the CPU design These all play a vital role in ensuring that the world's AI-based experiences run #onArm: https://bit.ly/4b9t4Of

Digging into customers’ demand for ML and AI workloads 🤖#theCUBE exclusive at Juniper Networks’ Seize the AI Moment event. Join Raj Yavatkar, the CTO at Juniper Networks, in this #AINativeNow Industry Panel, where he hears from Steve Scott, Corporate Fellow at AMD, on the rising demand for compute consolidation, efficiency, space, power, and flexibility to deal with ML and AI workloads. “AI is just driving a huge appetite for performance. Everybody's constrained by budgets and space and power. One of the things that customers are asking us about is the need to consolidate their general purpose compute to make up space and power for AI,” shares Scott. “A lot of our AI interest is actually on the EPYC side with people using EPYC CPUs to consolidate their older servers because they're very power efficient, and they are seeing a 5x to 8x consolidation possible from older servers,” he furthers. “For the AI infrastructure itself, people are looking for choice, flexibility, and competition. They want better price performance for training, but especially for inference. Most of the inference is done on CPUs, but the high-end inference is all being done on GPUs, and we're pretty consistently hearing that people are looking for a second source for their accelerators. So, there is a lot of interest in the AMD Instinct 30x,” Scott concludes. 📺 Tune in now! https://lnkd.in/g7mscxp4 #theCUBE #SeizetheAIMoment

#supermicro #supercloud #composer enforces #compliance and #governance for a #LCD by insituting emergency shutdown procedures protecting your capital investment. If a leak were detected, all servers associated with a liquid cooled rack would be shutdown, liquid flow rate altered and other safequard measures inacted, sending critical alerts to #administrators to take action. My #supermicro team is committed in building the next #futuristic #innovative #management platform in the industry to support your #enterprise crictical #AI #hpc workloads. Check in with us later because we have a lot more to report on #liquidcooling innovation. We are #supermicro the pioneers and visionaries of the #ai revolution! #intel #nvidia #amd #qualcomm #gpu #cpu #cdu #pdu #llms #ml #dl

🤖 The AI hardware boom is in full swing, the first half of this year witnessed a robust increase in AI server orders. Looking ahead, the latter half of the year promises more to come as NVIDIA’s Blackwell GB200 servers and WoA AI-powered notebooks hit mass production and begin shipping in Q3. This surge is driving ODMs to ramp up their inventory procurement and setting the stage for a spike in orders and shipments of high-capacitance MLCCs. The outcome is a welcome stabilization of market prices and a notable uptick in ASP for suppliers. #MLCCs #chipmarket #AIserver #ODMs

How #NVIDIA's #SuperPod architecture evolves from the #H100 to the #GH200 and #GB200? This article explains innovations in #GPU and #CPU interconnects, including #NVLink and #NVSwitch, which enable #AI supercomputers with up to 576 GPUs to accelerate AI model training, inference, and #HPC workloads. Follow the link to our blog: https://lnkd.in/grVxXYwU