Breaking the Memory Wall: How d-Matrix Is Redefining AI Inference with Chiplets
By Maurizio Di Paolo Emilio, embedded.com | April 23, 2025
As AI workloads push the limits of performance, power efficiency, and memory bandwidth, chiplets are rapidly emerging as the architectural solution of choice. In this interview, Sree Ganesan, Vice President of Product at d-Matrix, dives deep into how its pioneering chiplet-based platform is revolutionizing AI inference. From solving the memory wall with Digital In-Memory Computing (DIMC) to enabling seamless multi-chiplet communication via custom interconnects, d-Matrix reveals how its innovations are unlocking 10x faster token generation, 3x better energy efficiency, and a scalable roadmap for generative AI.
To read the full article, click here
Related Chiplet
- 12nm EURYTION RFK1 - UCIe SP based Ka-Ku Band Chiplet Transceiver
- Interconnect Chiplet
- Bridglets
- Automotive AI Accelerator
- Direct Chiplet Interface
Related News
- EdgeCortix Awarded New 3 Billion Yen NEDO Project to Develop Advanced Energy-Efficient AI Chiplet for Edge Inference and Learning
- Alphawave Semi Tapes Out Industry-First, Multi-Protocol I/O Connectivity Chiplet for High-Performance Compute and AI Infrastructure
- YorChip announces Low Latency 200G Chiplet for edge AI
- Tenstorrent Licenses Baya Systems’ Fabric into next-generation AI and Compute Chiplet Solutions
Latest News
- Multi-Die Assemblies Complicate Parasitic Extraction
- Alphawave Semi at the Forefront of PCIe® 7.0 Specification: Showcasing Next-Gen Chiplet Interoperability and Optical PCIe Technology at PCI-SIG® Developers Conference 2025
- Nordson Electronics Solutions Develops Panel-Level Packaging Solution for Powertech Technology, Inc. that Achieves Yields Greater than 99% for Underfilling During Semiconductor Manufacturing
- QuickLogic Joins Intel Foundry Chiplet Alliance
- Qualitas Semiconductor Successfully Develops UCIe v2.0-Compliant PHY IP, Enabling Up to 512Gbps Die-to-Die Connectivity