AMD Unveils CPU & GPU Solutions in AI Leadership Bid
As demand for specialised AI hardware continues to surge, driven by the widespread adoption of generative AI technologies across industries, Advanced Micro Devices (AMD) has launched a new range of artificial intelligence (AI) accelerators and networking solutions, intensifying its competition with market leader Nvidia in the rapidly growing AI chip sector.
The global AI chip market, valued at US$123bn in 2023, is projected to reach US$311bn by 2028, according to research firm MarketsandMarkets. Nvidia currently dominates this space, with its graphics processing units (GPUs) powering much of the world’s AI infrastructure. However, AMD, long known for its central processing units (CPUs) and GPUs for personal computers and data centres, is making a concerted effort to capture a larger share of this lucrative market.
Announced at its 2024 Advancing AI event, AMD’s new offerings – which include high-performance accelerators, networking components and the availability of its latest AMD EPYC processors – are designed to address the growing need for more powerful and efficient AI computing solutions. These products aim to compete directly with Nvidia's data centre GPUs, which have become the de facto standard for training and running large language models that underpin popular AI applications like ChatGPT.
Having recently celebrated its 55th birthday, AMD is looking ahead at the opportunities the new AI era presents, believing that AI’s full potential will be realised when the technology is pervasive, announcing recent acquisitions of ZT Systems for US$4.9bn and Finnish startup Silo AI for US$665m as part of a major push into the AI infrastructure market.
“The data centre and AI represent significant growth opportunities for AMD, and we are building strong momentum for our EPYC and AMD Instinct processors across a growing set of customers,” said AMD Chair and CEO Dr Lisa Su. “Looking ahead, we see the data center AI accelerator market growing to US$500bn by 2028. We are committed to delivering open innovation at scale through our expanded silicon, software, network and cluster-level solutions.”
Advances in AI accelerators
The AMD Instinct MI325X accelerators, built on the company's CDNA 3 architecture, are designed for AI tasks such as training and inference of large language models. With its MI300 series accelerators powering the most powerful Gen AI platforms from Microsoft, OpenAI and Meta, AMD claims the MI325X provides superior performance compared to competing products.
According to Forrest Norrod, Executive Vice President and General Manager of AMD's Data Center Solutions Business Group: “AMD continues to deliver on our roadmap, offering customers the performance they need and the choice they want, to bring AI infrastructure, at scale, to market faster.”
The company plans to begin production shipments of the MI325X in the fourth quarter of 2024, with system availability from partners including Dell Technologies, Hewlett Packard Enterprise and Lenovo expected in the first quarter of 2025.
AMD also previewed its next-generation Instinct MI350 series accelerators, based on the CDNA 4 architecture. These are slated for release in the second half of 2025 and are projected to offer significant performance improvements over the current generation – with AMD quoting a 35x improvement in inference performance compared to AMD’s CDNA 3-based accelerators.
Networking solutions
To address the networking requirements of AI infrastructure, AMD introduced two new products: the Pensando Salina Data Processing Unit (DPU) and the Pensando Pollara 400 Network Interface Card (NIC).
The Pensando Salina DPU, designed for the front-end of AI networks, supports 400 gigabit per second throughput. The Pensando Pollara 400, described as the industry's first Ultra Ethernet Consortium (UEC) ready AI NIC, is intended for back-end networks to facilitate accelerator-to-accelerator communication.
Both networking products are currently being sampled by customers and are expected to be available in the first half of 2025.
CPUs enabling AI leadership
AMD also announced the availability of the fifth generation of its EPYC processors, formerly codenamed ‘Turin,’ describing it as the world’s best server CPU for enterprise, AI and cloud.
Using its Zen 5 core architecture, the AMD EPYC 9005 Series processors extend the record-breaking performance and energy efficiency of the previous generations with the top of stack 192 core CPU, with AMD figures stating this chip is capable of delivering up to 2.7 times the performance compared to its competitors.
New to the AMD EPYC 9005 Series CPUs is the 64 core AMD EPYC 9575F, tailor made for GPU-powered AI solutions that need the ultimate in host CPU capabilities. Boosting up to 5GHz, compared to the 3.8GHz processor of the competition, it provides up to 28% faster processing needed to keep GPUs fed with data for demanding AI workloads.
“From powering the world’s fastest supercomputers, to leading enterprises, to the largest hyperscalers, AMD has earned the trust of customers who value demonstrated performance, innovation and energy efficiency,” comments Dan McNamara, SVP and General Manager of AMD’s server business. “With five generations of on-time roadmap execution, AMD has proven it can meet the needs of the data centre market and give customers the standard for data centre performance, efficiency, solutions and capabilities for cloud, enterprise and AI workloads.”
******
Make sure you check out the latest edition of Technology Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
******
Technology Magazine is a BizClik brand
- Gartner: Why CIOs Struggle With Digital Initiative SuccessDigital Transformation
- How Palantir's AI Growth is Fuelling Investor ConfidenceAI & Machine Learning
- IAG: How AI is Impacting the Aviation IndustryAI & Machine Learning
- Meta's AI Plans Face Environmental Challenges: ExplainedAI & Machine Learning