Fujitsu and AMD partner to develop energy-efficient HPC/AI platformsPartnership aims to broaden access to AI, support open-sourceMonaka chip features 288 cores, 2nm process, Armv9-A architecture
Fujitsu and AMD have announced a new strategic partnership focused on developing HPC and AI platforms.
This collaboration will combine Fujitsu’s ARM-based processor technology with AMD’s GPU expertise, aiming to build energy-efficient and open-source solutions addressing the growing demand for diverse, cost-effective computing architectures.
The partnership, formalized in a memorandum of understanding, covers joint efforts in technology development, commercialization, and ecosystem expansion, with a goal of creating powerful computing platforms by 2027.
Instinct accelerators
Central to this partnership is Fujitsu’s next-generation Arm-based processor, the Monaka chip, set for release in 2027.
As we first reported in July 2024, Monaka will feature Armv9-A architecture, a 2nm process for high performance and power efficiency, and a 288-core structure (144 cores per socket). Notably, Monaka moves away from high-bandwidth memory in favor of PCIe 6.0 (CXL3.0), enhancing scalability and connectivity. This processor is expected to build upon Fujitsu’s work with the A64FX chip, used in the Fugaku supercomputer, and could support the upcoming FugakuNEXT project planned for 2030.
AMD will support Monaka with its Instinct accelerators, providing customers with flexible options for handling massive AI workloads while optimizing data center costs. The collaboration will also make use of AMD’s ROCm software stack and Fujitsu’s proprietary software, developing an open source ecosystem to accelerate the development of AI and HPC applications.
This joint venture will also involve global marketing efforts and customer engagement, as well as a shared customer center to support the development and implementation of AI technologies.
“By combining AMD’s innovative GPU technology with Fujitsu’s low-power/high-performance processor Fujitsu-Monaka, we seek to create an environment in which more companies will be able to utilize AI while reducing the power consumed by data centers,” noted Vivek Mahajan, CTO of Fujitsu.
You might also like
Fujitsu is planning the most powerful CPU ever for its supercomputerAMD head outlines the challenges for data centersAMD launches a new single-slot GPU for data centers
