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
- Untether AI Enters Into a Strategic Agreement with AMD
- Alphawave Semi Tapes Out Breakthrough 36G UCIe™ IP on TSMC 2nm, Unlocking Foundational AI Platform IP on Nanosheet Processes
- How Secure Are Analog Circuits?
- Sarcina Technology advances photonic package design to address key data center challenges
- Imec demonstrates 16nm pitch Ru lines with record-low resistance obtained using a semi-damascene integration approach