News Posts matching #LLM

Return to Keyword Browsing

Axelera AI Partners with Arduino for Edge AI Solutions

Axelera AI - a leading edge-inference company - and Arduino, the global leader in open-source hardware and software, today announced a strategic partnership to make high-performance AI at the edge more accessible than ever, building advanced technology solutions based on inference and an open ecosystem. This furthers Axelera AI's strategy to democratize artificial intelligence everywhere.

The collaboration will combine the strengths of Axelera AI's Metis AI Platform with the powerful SOMs from the Arduino Pro range to provide customers with easy-to-use hardware and software to innovate around AI. Users will enjoy the freedom to dictate their own AI journey, thanks to tools that provide unique digital in-memory computing and RISC-V controlled dataflow technology, delivering high performance and usability at a fraction of the cost and power of other solutions available today.

NVIDIA Unveils New Jetson Orin Nano Super Developer Kit

NVIDIA is taking the wraps off a new compact generative AI supercomputer, offering increased performance at a lower price with a software upgrade. The new NVIDIA Jetson Orin Nano Super Developer Kit, which fits in the palm of a hand, provides everyone from commercial AI developers to hobbyists and students, gains in generative AI capabilities and performance. And the price is now $249, down from $499.

Available today, it delivers as much as a 1.7x leap in generative AI inference performance, a 70% increase in performance to 67 INT8 TOPS, and a 50% increase in memory bandwidth to 102 GB/s compared with its predecessor. Whether creating LLM chatbots based on retrieval-augmented generation, building a visual AI agent, or deploying AI-based robots, the Jetson Orin Nano Super is an ideal solution to fetch.

Advantech Introduces Its GPU Server SKY-602E3 With NVIDIA H200 NVL

Advantech, a leading global provider of industrial edge AI solutions, is excited to introduce its GPU server SKY-602E3 equipped with the NVIDIA H200 NVL platform. This powerful combination is set to accelerate the offline LLM for manufacturing, providing unprecedented levels of performance and efficiency. The NVIDIA H200 NVL, requiring 600 W passive cooling, is fully supported by the compact and efficient SKY-602E3 GPU server, making it an ideal solution for demanding edge AI applications.

Core of Factory LLM Deployment: AI Vision
The SKY-602E3 GPU server excels in supporting large language models (LLMs) for AI inference and training. It features four PCIe 5.0 x16 slots, delivering high bandwidth for intensive tasks, and four PCIe 5.0 x8 slots, providing enhanced flexibility for GPU and frame grabber card expansion. The half-width design of the SKY-602E3 makes it an excellent choice for workstation environments. Additionally, the server can be equipped with the NVIDIA H200 NVL platform, which offers 1.7x more performance than the NVIDIA H100 NVL, freeing up additional PCIe slots for other expansion needs.

Amazon AWS Announces General Availability of Trainium2 Instances, Reveals Details of Next Gen Trainium3 Chip

At AWS re:Invent, Amazon Web Services, Inc. (AWS), an Amazon.com, Inc. company, today announced the general availability of AWS Trainium2-powered Amazon Elastic Compute Cloud (Amazon EC2) instances, introduced new Trn2 UltraServers, enabling customers to train and deploy today's latest AI models as well as future large language models (LLM) and foundation models (FM) with exceptional levels of performance and cost efficiency, and unveiled next-generation Trainium3 chips.

"Trainium2 is purpose built to support the largest, most cutting-edge generative AI workloads, for both training and inference, and to deliver the best price performance on AWS," said David Brown, vice president of Compute and Networking at AWS. "With models approaching trillions of parameters, we understand customers also need a novel approach to train and run these massive workloads. New Trn2 UltraServers offer the fastest training and inference performance on AWS and help organizations of all sizes to train and deploy the world's largest models faster and at a lower cost."

Microsoft Office Tools Reportedly Collect Data for AI Training, Requiring Manual Opt-Out

Microsoft's Office suite is the staple in productivity tools, with millions of users entering sensitive personal and company data into Excel and Word. According to @nixCraft, an author from Cyberciti.biz, Microsoft left its "Connected Experiences" feature enabled by default, reportedly using user-generated content to train the company's AI models. This feature is enabled by default, meaning data from Word and Excel files may be used in AI development unless users manually opt-out. As a default option, this setting raises security concerns, especially from businesses and government workers relying on Microsoft Office for proprietary work. The feature allows documents such as articles, government data, and other confidential files to be included in AI training, creating ethical and legal challenges regarding consent and intellectual property.

Disabling the feature requires going to: File > Options > Trust Center > Trust Center Settings > Privacy Options > Privacy Settings > Optional Connected Experiences, and unchecking the box. Even with an unnecessary long opt-out steps, the European Union's GPDR agreement, which Microsoft complies with, requires all settings to be opt-in rather than opt-out by default. This directly contradicts EU GDPR laws, which could prompt an investigation from the EU. Microsoft has yet to confirm whether user content is actively being used to train its AI models. However, its Services Agreement includes a clause granting the company a "worldwide and royalty-free intellectual property license" to use user-generated content for purposes such as improving Microsoft products. The controversy raised from this is not new, especially where more companies leverage user data for AI development, often without explicit consent.

Aetina Debuts at SC24 With NVIDIA MGX Server for Enterprise Edge AI

Aetina, a subsidiary of the Innodisk Group and an expert in edge AI solutions, is pleased to announce its debut at Supercomputing (SC24) in Atlanta, Georgia, showcasing the innovative SuperEdge NVIDIA MGX short-depth edge AI server, AEX-2UA1. By integrating an enterprise-class on-premises large language model (LLM) with the advanced retrieval-augmented generation (RAG) technique, Aetina NVIDIA MGX short-depth server demonstrates exceptional enterprise edge AI performance, setting a new benchmark in Edge AI innovation. The server is powered by the latest Intel Xeon 6 processor and dual high-end double-width NVIDIA GPUs, delivering ultimate AI computing power in a compact 2U form factor, accelerating Gen AI at the edge.

The SuperEdge NVIDIA MGX server expands Aetina's product portfolio from specialized edge devices to comprehensive AI server solutions, propelling a key milestone in Innodisk Group's AI roadmap, from sensors and storage to AI software, computing platforms, and now AI edge servers.

Hypertec Introduces the World's Most Advanced Immersion-Born GPU Server

Hypertec proudly announces the launch of its latest breakthrough product, the TRIDENT iG series, an immersion-born GPU server line that brings extreme density, sustainability, and performance to the AI and HPC community. Purpose-built for the most demanding AI applications, this cutting-edge server is optimized for generative AI, machine learning (ML), deep learning (DL), large language model (LLM) training, inference, and beyond. With up to six of the latest NVIDIA GPUs in a 2U form factor, a staggering 8 TB of memory with enhanced RDMA capabilities, and groundbreaking density supporting up to 200 GPUs per immersion tank, the TRIDENT iG server line is a game-changer for AI infrastructure.

Additionally, the server's innovative design features a single or dual root complex, enabling greater flexibility and efficiency for GPU usage in complex workloads.

Q.ANT Introduces First Commercial Photonic Processor

Q.ANT, the leading startup for photonic computing, today announced the launch of its first commercial product - a photonics-based Native Processing Unit (NPU) built on the company's compute architecture LENA - Light Empowered Native Arithmetics. The product is fully compatible with today's existing computing ecosystem as it comes on the industry-standard PCI-Express. The Q.ANT NPU executes complex, non-linear mathematics natively using light instead of electrons, promising to deliver at least 30 times greater energy efficiency and significant computational speed improvements over traditional CMOS technology. Designed for compute-intensive applications such as AI Inference, machine learning, and physics simulation, the Q.ANT NPU has been proven to solve real-world challenges, including number recognition for deep neural network inference (see the recent press release regarding Cloud Access to NPU).

"With our photonic chip technology now available on the standard PCIe interface, we're bringing the incredible power of photonics directly into real-world applications. For us, this is not just a processor—it's a statement of intent: Sustainability and performance can go hand in hand," said Dr. Michael Förtsch, CEO of Q.ANT. "For the first time, developers can create AI applications and explore the capabilities of photonic computing, particularly for complex, nonlinear calculations. For example, experts calculated that one GPT-4 query today uses 10 times more electricity than a regular internet search request. Our photonic computing chips offer the potential to reduce the energy consumption for that query by a factor of 30."

IBM Expands Its AI Accelerator Offerings; Announces Collaboration With AMD

IBM and AMD have announced a collaboration to deploy AMD Instinct MI300X accelerators as a service on IBM Cloud. This offering, which is expected to be available in the first half of 2025, aims to enhance performance and power efficiency for Gen AI models such as and high-performance computing (HPC) applications for enterprise clients. This collaboration will also enable support for AMD Instinct MI300X accelerators within IBM's watsonx AI and data platform, as well as Red Hat Enterprise Linux AI inferencing support.

"As enterprises continue adopting larger AI models and datasets, it is critical that the accelerators within the system can process compute-intensive workloads with high performance and flexibility to scale," said Philip Guido, executive vice president and chief commercial officer, AMD. "AMD Instinct accelerators combined with AMD ROCm software offer wide support including IBM watsonx AI, Red Hat Enterprise Linux AI and Red Hat OpenShift AI platforms to build leading frameworks using these powerful open ecosystem tools. Our collaboration with IBM Cloud will aim to allow customers to execute and scale Gen AI inferencing without hindering cost, performance or efficiency."

GIGABYTE Launches AMD Radeon PRO W7800 AI TOP 48G Graphics Card

GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of premium gaming hardware, today launched the cutting-edge GIGABYTE AMD Radeon PRO W7800 AI TOP 48G. GIGABYTE has taken a significant leap forward with the release of the Radeon PRO W7800 AI TOP 48G graphics card, featuring AMD's RDNA 3 architecture and a massive 48 GB of GDDR6 memory. This significant increase in memory capacity, compared to its predecessor, provides workstation professionals, creators, and AI developers with incredible computational power to effortlessly handle complex design, rendering, and AI model training tasks.

⁠GIGABYTE stands as the AMD professional graphics partner in the market, with a proven ability to design and manufacture the entire Radeon PRO series. Our dedication to quality products, unwavering business commitment, and comprehensive customer service empower us to deliver professional-grade GPU solutions, expanding user's choices in workstation and AI computing.⁠

Anthropic Develops AI Model That Can Use Computers, Updates Claude 3.5 Sonnet

The age of automation is upon us. Anthropic, the company behind top-performing Claude large language models that compete directly with OpenAI GPT, has today announced updates to its models and a new feature—computer use. The computer use allows Claude 3.5 Sonnet model to access the user's system by looking at the screen, moving the cursor, typing text, and clicking buttons. While only being experimental for now, the system is prone to errors and creating "dumb" mistakes. However, it allows for one very important feature: driving the operating system designed for humans using artificial intelligence.

There is a benchmark that evaluates AI model's ability to use computers like a human does on human-centered operating system. Called OSWorld, the Claude 3.5 Sonnet model has managed to score 14.9% in screenshot-only category, and 22.0% in some other tasks that require more steps. A typical human scores around 72.36% in this testing, which proves to be difficult even for natural intelligence. However, this is only the beginning as these models advance rapidly. Usually, these models work with other types of data, like text and static images, where they process it and do computation based on it. Working on computers designed for human interaction first is a great leap in the capabilities of AI models.

Intel Won't Compete Against NVIDIA's High-End AI Dominance Soon, Starts Laying Off Over 2,200 Workers Across US

Intel's taking a different path with its Gaudi 3 accelerator chips. It's staying away from the high-demand market for training big AI models, which has made NVIDIA so successful. Instead, Intel wants to help businesses that need cheaper AI solutions to train and run smaller specific models and open-source options. At a recent event, Intel talked up Gaudi 3's "price performance advantage" over NVIDIA's H100 GPU for inference tasks. Intel says Gaudi 3 is faster and more cost-effective than the H100 when running Llama 3 and Llama 2 models of different sizes.

Intel also claims that Gaudi 3 is as power-efficient as the H100 for large language model (LLM) inference with small token outputs and does even better with larger outputs. The company even suggests Gaudi 3 beats NVIDIA's newer H200 in LLM inference throughput for large token outputs. However, Gaudi 3 doesn't match up to the H100 in overall floating-point operation throughput for 16-bit and 8-bit formats. For bfloat16 and 8-bit floating-point precision matrix math, Gaudi 3 hits 1,835 TFLOPS in each format, while the H100 reaches 1,979 TFLOPS for BF16 and 3,958 TFLOPS for FP8.

NVIDIA Fine-Tunes Llama3.1 Model to Beat GPT-4o and Claude 3.5 Sonnet with Only 70 Billion Parameters

NVIDIA has officially released its Llama-3.1-Nemotron-70B-Instruct model. Based on META's Llama3.1 70B, the Nemotron model is a large language model customized by NVIDIA in order to improve the helpfulness of LLM-generated responses. NVIDIA uses fine-tuning structured data to steer the model and allow it to generate more helpful responses. With only 70 billion parameters, the model is punching far above its weight class. The company claims that the model is beating the current top models from leading labs like OpenAI's GPT-4o and Anthropic's Claude 3.5 Sonnet, which are the current leaders across AI benchmarks. In evaluations such as Arena Hard, the NVIDIA Llama3.1 Nemotron 70B is scoring 85 points, while GPT-4o and Sonnet 3.5 score 79.3 and 79.2, respectively. Other benchmarks like AlpacaEval and MT-Bench spot NVIDIA also hold the top spot, with 57.6 and 8.98 scores earned. Claude and GPT reach 52.4 / 8.81 and 57.5 / 8.74, just below Nemotron.

This language model underwent training using reinforcement learning from human feedback (RLHF), specifically employing the REINFORCE algorithm. The process involved a reward model based on a large language model architecture and custom preference prompts designed to guide the model's behavior. The training began with a pre-existing instruction-tuned language model as the starting point. It was trained on Llama-3.1-Nemotron-70B-Reward and HelpSteer2-Preference prompts on a Llama-3.1-70B-Instruct model as the initial policy. Running the model locally requires either four 40 GB or two 80 GB VRAM GPUs and 150 GB of free disk space. We managed to take it for a spin on NVIDIA's website to say hello to TechPowerUp readers. The model also passes the infamous "strawberry" test, where it has to count the number of specific letters in a word, however, it appears that it was part of the fine-tuning data as it fails the next test, shown in the image below.

MSI Unveils AI Servers Powered by NVIDIA MGX at OCP 2024

MSI, a leading global provider of high-performance server solutions, proudly announced it is showcasing new AI servers powered by the NVIDIA MGX platform—designed to address the increasing demand for scalable, energy-efficient AI workloads in modern data centers—at the OCP Global Summit 2024, booth A6. This collaboration highlights MSI's continued commitment to advancing server solutions, focusing on cutting-edge AI acceleration and high-performance computing (HPC).

The NVIDIA MGX platform offers a flexible architecture that enables MSI to deliver purpose-built solutions optimized for AI, HPC, and LLMs. By leveraging this platform, MSI's AI server solutions provide exceptional scalability, efficiency, and enhanced GPU density—key factors in meeting the growing computational demands of AI workloads. Tapping into MSI's engineering expertise and NVIDIA's advanced AI technologies, these AI servers based on the MGX architecture deliver unparalleled compute power, positioning data centers to maximize performance and power efficiency while paving the way for the future of AI-driven infrastructure.

Arm and Partners Develop AI CPU: Neoverse V3 CSS Made on 2 nm Samsung GAA FET

Yesterday, Arm has announced significant progress in its Total Design initiative. The program, launched a year ago, aims to accelerate the development of custom silicon for data centers by fostering collaboration among industry partners. The ecosystem has now grown to include nearly 30 participating companies, with recent additions such as Alcor Micro, Egis, PUF Security, and SEMIFIVE. A notable development is a partnership between Arm, Samsung Foundry, ADTechnology, and Rebellions to create an AI CPU chiplet platform. This collaboration aims to deliver a solution for cloud, HPC, and AI/ML workloads, combining Rebellions' AI accelerator with ADTechnology's compute chiplet, implemented using Samsung Foundry's 2 nm Gate-All-Around (GAA) FET technology. The platform is expected to offer significant efficiency gains for generative AI workloads, with estimates suggesting a 2-3x improvement over the standard CPU design for LLMs like Llama3.1 with 405 billion parameters.

Arm's approach emphasizes the importance of CPU compute in supporting the complete AI stack, including data pre-processing, orchestration, and advanced techniques like Retrieval-augmented Generation (RAG). The company's Compute Subsystems (CSS) are designed to address these requirements, providing a foundation for partners to build diverse chiplet solutions. Several companies, including Alcor Micro and Alphawave, have already announced plans to develop CSS-powered chiplets for various AI and high-performance computing applications. The initiative also focuses on software readiness, ensuring that major frameworks and operating systems are compatible with Arm-based systems. Recent efforts include the introduction of Arm Kleidi technology, which optimizes CPU-based inference for open-source projects like PyTorch and Llama.cpp. Notably, as Google claims, most AI workloads are being inferenced on CPUs, so creating the most efficient and most performant CPUs for AI makes a lot of sense.

MSI Showcases Innovation at 2024 OCP Global Summit, Highlighting DC-MHS, CXL Memory Expansion, and MGX-enabled AI Servers

MSI, a leading global provider of high-performance server solutions, is excited to showcase its comprehensive lineup of motherboards and servers based on the OCP Modular Hardware System (DC-MHS) architecture at the OCP Global Summit from October 15-17 at booth A6. These cutting-edge solutions represent a breakthrough in server designs, enabling flexible deployments for cloud and high-density data centers. Featured innovations include CXL memory expansion servers and AI-optimized servers, demonstrating MSI's leadership in pushing the boundaries of AI performance and computing power.

DC-MHS Series Motherboards and Servers: Enabling Flexible Deployment in Data Centers
"The rapidly evolving IT landscape requires cloud service providers, large-scale data center operators, and enterprises to handle expanding workloads and future growth with more flexible and powerful infrastructure. MSI's new rage of DC-MHS-based solutions provides the needed flexibility and efficiency for modern data center environments," said Danny Hsu, General Manager of Enterprise Platform Solutions.

Western Digital Enterprise SSDs Certified to Support NVIDIA GB200 NVL72 System for Compute-Intensive AI Environments

Western Digital Corp. today announced that its PCIe Gen 5 DC SN861 E.1S enterprise-class NVMe SSDs have been certified to support the NVIDIA GB200 NVL72 rack-scale system.

The rapid rise of AI, ML, and large language models (LLMs) is creating a challenge for companies with two opposing forces. Data generation and consumption are accelerating, while organizations face pressure to quickly derive value from this data. Performance, scalability, and efficiency are essential for AI technology stacks as storage demands rise. Certified to be compatible with the GB200 NVL72 system, Western Digital's enterprise SSD addresses the growing needs of the AI market for high-speed accelerated computing combined with low latency to serve compute-intensive AI environments.

Lenovo Accelerates Business Transformation with New ThinkSystem Servers Engineered for Optimal AI and Powered by AMD

Today, Lenovo announced its industry-leading ThinkSystem infrastructure solutions powered by AMD EPYC 9005 Series processors, as well as AMD Instinct MI325X accelerators. Backed by 225 of AMD's world-record performance benchmarks, the Lenovo ThinkSystem servers deliver an unparalleled combination of AMD technology-based performance and efficiency to tackle today's most demanding edge-to-cloud workloads, including AI training, inferencing and modeling.

"Lenovo is helping organizations of all sizes and across various industries achieve AI-powered business transformations," said Vlad Rozanovich, Senior Vice President, Lenovo Infrastructure Solutions Group. "Not only do we deliver unmatched performance, we offer the right mix of solutions to change the economics of AI and give customers faster time-to-value and improved total value of ownership."

Supermicro Currently Shipping Over 100,000 GPUs Per Quarter in its Complete Rack Scale Liquid Cooled Servers

Supermicro, Inc., a Total IT Solution Provider for Cloud, AI/ML, Storage, and 5G/Edge, is announcing a complete liquid cooling solution that includes powerful Coolant Distribution Units (CDUs), cold plates, Coolant Distribution Manifolds (CDMs), cooling towers and end to end management software. This complete solution reduces ongoing power costs and Day 0 hardware acquisition and data center cooling infrastructure costs. The entire end-to-end data center scale liquid cooling solution is available directly from Supermicro.

"Supermicro continues to innovate, delivering full data center plug-and-play rack scale liquid cooling solutions," said Charles Liang, CEO and president of Supermicro. "Our complete liquid cooling solutions, including SuperCloud Composer for the entire life-cycle management of all components, are now cooling massive, state-of-the-art AI factories, reducing costs and improving performance. The combination of Supermicro deployment experience and delivering innovative technology is resulting in data center operators coming to Supermicro to meet their technical and financial goals for both the construction of greenfield sites and the modernization of existing data centers. Since Supermicro supplies all the components, the time to deployment and online are measured in weeks, not months."

Intel Updates "AI Playground" Application for Local AI Models with "Lunar Lake" Support

Intel has announced the release of an updated version of its AI Playground application, now optimized for the new Intel Core Ultra 200V "Lunar Lake" series of processors. This latest iteration, version 1.21b, brings a host of new features and improvements designed to make AI more accessible to users of Intel's AI-enabled PCs. AI Playground, first launched earlier this year, offers a user-friendly interface for various AI functions, including image generation, enhancement, and natural language processing. The new version introduces several key enhancements. These include a fresh, exclusive theme for 200V series processor users, an expanded LLM picker now featuring Phi3, Qwen2, and Mistral models, and a conversation manager for saving and revisiting chat discussions. Additionally, users will find adjustable font sizes for improved readability and a simplified aspect ratio tool for image creation and enhancement.

One of the most significant aspects of AI Playground is its ability to run entirely locally on the user's machine. This approach ensures that all computations, prompts, and outputs remain on the device, addressing privacy concerns often associated with cloud-based AI services. The application is optimized to take advantage of the Xe Cores and XMX AI engines found in the Intel Core Ultra 200V series processors, allowing even lightweight devices to perform complex AI tasks efficiently. Intel has also improved the installation process, addressing potential conflicts and providing better error handling. The company encourages user engagement through its Intel Insiders Discord channel, helping the community around AI Playground's development and use. Although the models users can run locally are smaller in size, usually up to 7 billion parameters with 8/4-bit quants, having a centralized application to help run them locally is significant for slowly embedding AI in all aspects of personal computing.

Advantech Launches AIR-310, Ultra-Low-Profile Scalable AI Inference

Advantech, a leading provider of edge computing solutions, introduces the AIR-310, a compact edge AI inference system featuring an MXM GPU card. Powered by 12th/13th/14th Gen Intel Core 65 W desktop processors, the AIR-310 delivers up to 12.99 TFLOPS of scalable AI performance via the NVIDIA Quadro 2000A GPU card in a 1.5U chassis (215 x 225 x 55 mm). Despite its compact size, it offers versatile connectivity with three LAN ports and four USB 3.0 ports, enabling seamless integration of sensors and cameras for vision AI applications.

The system includes smart fan management, operates in temperatures from 0 to 50°C (32 to 122°F), and is shock-resistant, capable of withstanding 3G vibration and 30G shock. Bundled with Intel Arc A370 and NVIDIA A2000 GPUs, it is certified to IEC 61000-6-2, IEC 61000-6-4, and CB/UL standards, ensuring stable 24/7 operation in harsh environments, including space-constrained or mobile equipment. The AIR-310 supports Windows 11, Linux Ubuntu 24.04, and the Edge AI SDK, enabling accelerated inference deployment for applications such as factory inspections, real-time video surveillance, GenAI/LLM, and medical imaging.

AMD Instinct MI300X Accelerators Available on Oracle Cloud Infrastructure

AMD today announced that Oracle Cloud Infrastructure (OCI) has chosen AMD Instinct MI300X accelerators with ROCm open software to power its newest OCI Compute Supercluster instance called BM.GPU.MI300X.8. For AI models that can comprise hundreds of billions of parameters, the OCI Supercluster with AMD MI300X supports up to 16,384 GPUs in a single cluster by harnessing the same ultrafast network fabric technology used by other accelerators on OCI. Designed to run demanding AI workloads including large language model (LLM) inference and training that requires high throughput with leading memory capacity and bandwidth, these OCI bare metal instances have already been adopted by companies including Fireworks AI.

"AMD Instinct MI300X and ROCm open software continue to gain momentum as trusted solutions for powering the most critical OCI AI workloads," said Andrew Dieckmann, corporate vice president and general manager, Data Center GPU Business, AMD. "As these solutions expand further into growing AI-intensive markets, the combination will benefit OCI customers with high performance, efficiency, and greater system design flexibility."

SK hynix Presents Upgraded AiMX Solution at AI Hardware and Edge AI Summit 2024

SK hynix unveiled an enhanced Accelerator-in-Memory based Accelerator (AiMX) card at the AI Hardware & Edge AI Summit 2024 held September 9-12 in San Jose, California. Organized annually by Kisaco Research, the summit brings together representatives from the AI and machine learning ecosystem to share industry breakthroughs and developments. This year's event focused on exploring cost and energy efficiency across the entire technology stack. Marking its fourth appearance at the summit, SK hynix highlighted how its AiM products can boost AI performance across data centers and edge devices.

Booth Highlights: Meet the Upgraded AiMX
In the AI era, high-performance memory products are vital for the smooth operation of LLMs. However, as these LLMs are trained on increasingly larger datasets and continue to expand, there is a growing need for more efficient solutions. SK hynix addresses this demand with its PIM product AiMX, an AI accelerator card that combines multiple GDDR6-AiMs to provide high bandwidth and outstanding energy efficiency. At the AI Hardware & Edge AI Summit 2024, SK hynix presented its updated 32 GB AiMX prototype which offers double the capacity of the original card featured at last year's event. To highlight the new AiMX's advanced processing capabilities in a multi-batch environment, SK hynix held a demonstration of the prototype card with the Llama 3 70B model, an open source LLM. In particular, the demonstration underlined AiMX's ability to serve as a highly effective attention accelerator in data centers.

SambaNova Launches Fastest AI Platform Based on Its SN40L Chip

SambaNova Systems, provider of the fastest and most efficient chips and AI models, announced SambaNova Cloud, the world's fastest AI inference service enabled by the speed of its SN40L AI chip. Developers can log on for free via an API today — no waiting list — and create their own generative AI applications using both the largest and most capable model, Llama 3.1 405B, and the lightning-fast Llama 3.1 70B. SambaNova Cloud runs Llama 3.1 70B at 461 tokens per second (t/s) and 405B at 132 t/s at full precision.

"SambaNova Cloud is the fastest API service for developers. We deliver world record speed and in full 16-bit precision - all enabled by the world's fastest AI chip," said Rodrigo Liang, CEO of SambaNova Systems. "SambaNova Cloud is bringing the most accurate open source models to the vast developer community at speeds they have never experienced before."

NVIDIA Blackwell Sets New Standard for Generative AI in MLPerf Inference Benchmark

As enterprises race to adopt generative AI and bring new services to market, the demands on data center infrastructure have never been greater. Training large language models is one challenge, but delivering LLM-powered real-time services is another. In the latest round of MLPerf industry benchmarks, Inference v4.1, NVIDIA platforms delivered leading performance across all data center tests. The first-ever submission of the upcoming NVIDIA Blackwell platform revealed up to 4x more performance than the NVIDIA H100 Tensor Core GPU on MLPerf's biggest LLM workload, Llama 2 70B, thanks to its use of a second-generation Transformer Engine and FP4 Tensor Cores.

The NVIDIA H200 Tensor Core GPU delivered outstanding results on every benchmark in the data center category - including the latest addition to the benchmark, the Mixtral 8x7B mixture of experts (MoE) LLM, which features a total of 46.7 billion parameters, with 12.9 billion parameters active per token. MoE models have gained popularity as a way to bring more versatility to LLM deployments, as they're capable of answering a wide variety of questions and performing more diverse tasks in a single deployment. They're also more efficient since they only activate a few experts per inference - meaning they deliver results much faster than dense models of a similar size.
Return to Keyword Browsing
Dec 30th, 2024 12:22 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts