Expanded AMD Instinct GPU Roadmap Now Out

New AMD Instinct MI325X accelerator expected to be available in Q4 2024 with up to 288GB of HBM3E memory; new AMD Instinct MI350 series accelerators based on AMD CDNA 4 architecture expected to be available in 2025 .

New AMD Instinct MI325X accelerator expected to be available in Q4 2024 with up to 288GB of HBM3E memory; new AMD Instinct MI350 series accelerators based on AMD CDNA 4 architecture expected to be available in 2025 .

AMD has unveiled (at the recent Computex 2024) a multiyear, expanded AMD Instinct accelerator roadmap which will bring an annual cadence of artificial intelligence (AI) performance and memory capabilities at every generation.

The updated roadmap starts with the new AMD Instinct MI325X accelerator, which will be available in Q4 2024. After that, the AMD Instinct MI350 series, powered by the new AMD CDNA 4 architecture, is expected to be available in 2025 . Expected to arrive in 2026, the AMD Instinct MI400 series is based on the AMD CDNA “Next” architecture.

“The AMD Instinct MI300X accelerators continue their strong adoption from numerous partners and customers including Microsoft Azure, Meta, Dell Technologies, HPE, Lenovo and others, a direct result of the AMD Instinct MI300X accelerator exceptional performance and value proposition,” says Brad McCredie, corporate vice president, Data Center Accelerated Compute, AMD. 

AMD AI Software Ecosystem Matures

The AMD ROCm 6 open software stack continues to mature, enabling AMD Instinct MI300X accelerators to drive performance for popular LLMs. On a server using eight AMD Instinct MI300X accelerators and ROCm 6 running Meta Llama-3 70B, customers can get better inference performance and token generation compared to the competition2. On a single AMD Instinct MI300X accelerator with ROCm 6, customers can get better inference performance and token generation throughput compared to the competition by 1.2x on Mistral-7B3. AMD also highlighted that Hugging Face, the largest and most popular repository for AI models, is now testing 700,000 of their most popular models nightly to ensure they work out of box on AMD Instinct MI300X accelerators. In addition, AMD is continuing its upstream work into popular AI frameworks like PyTorch, TensorFlow and JAX.

AMD Previews New Accelerators, Reveals Roadmap

During the keynote, AMD revealed an updated annual cadence for the AMD Instinct accelerator roadmap to meet the demand for more AI compute. This will help ensure that AMD Instinct accelerators propel the development of next-generation frontier AI models. The updated AMD Instinct annual roadmap highlighted:

The new AMD Instinct MI325X accelerator, which will bring 288GB of HBM3E memory and 6 terabytes per second of memory bandwidth, use the same Universal Baseboard server design used by the AMD Instinct MI300 series, and be generally available in Q4 2024. The accelerator will have memory capacity and bandwidth.

The first product in the AMD Instinct MI350 Series, the AMD Instinct MI350X accelerator, is based on the AMD CDNA 4 architecture and is expected to be available in 2025. It will use the same Universal Baseboard server design as other MI300 Series accelerators and will be built using advanced 3nm process technology, support the FP4 and FP6 AI datatypes and have up to 288 GB of HBM3E memory.

AMD CDNA “Next” architecture, which will power the AMD Instinct MI400 Series accelerators, is expected to be available in 2026 providing the latest features and capabilities that will help unlock additional performance and efficiency for inference and large-scale AI training.

Sources: Press materials received from the company and additional information gleaned from the company’s website.

More AMD Coverage

Accelerating Electric Vehicle Development with Multidisciplinary Simulation and High-Performance Computing
In this new Making the Case guide, learn how a unified approach to design and multidisciplinary simulation from Dassault Systèmes, combined with high-performance computing powered by AMD EPYC™ processors, can accelerate EV design.
Next-Gen Workstations Meet the Need for Speed
Artificial intelligence, multicore CPUs, GPU acceleration and more continue to drive engineering innovation.
AMD Introduces Versal RF Series Adaptive SoCs
New SoCs offer high compute in a single-chip device with integrated direct RF-sampling converters, AMD reports.
AMD Powers Fast Supercomputer, El Capitan
El Capitan touted as the first exascale-class machine for the National Nuclear Security Administration (NNSA) stands as a computing resource for the NNSA Tri-Labs — LLNL, Los Alamos and Sandia National Laboratories.
New Engineering Design Center for AMD Opens in Serbia
AMD expands in the Balkans with a new design center to improve software and AI capabilities.
Rise of the AI Workstation
Given the rapid interest in artificial intelligence, more workstation vendors are rising to meet demand.
AMD Company Profile

Share This Article

Subscribe to our FREE magazine, FREE email newsletters or both!

Join over 90,000 engineering professionals who get fresh engineering news as soon as it is published.


About the Author

DE Editors's avatar
DE Editors

DE’s editors contribute news and new product announcements to Digital Engineering.
Press releases may be sent to them via DE-Editors@digitaleng.news.

Follow DE
#29004