Optimizing Chiplet architectures of RISC-V clusters, GPU and DNN using System-Level IP
By Deepak Shankar, Mirabilis Design
Multi-die SoC containing multiple RISC-V clusters, GPU, NPU, accelerators and DNN have considerable benefits for applications in automotive, space and industrial.
Architecture exploration of the chiplet-based SoC requires multiple interconnect protocol models, and multiple coherent and non-coherent compute resources. This Video will demonstrate a methodology for rapid modeling and architecture trade-off using UCIe in modeling and simulating a multi-die SoC in a matter of weeks. These models can be used for UCIe setup, task partitioning, distribution of compute resources across dies, sharing coherency between RISC-V clusters, GPU and AI Engines.
We will show how workloads and application task graphs can be used to optimize the settings to meet the requirements, identify system bottlenecks and optimize for PPA. Results from actual hardware designs with different configurations and allocations will be discussed during the session.
Related Chiplet
- Optical Chiplet
- Reverse Gearbox Chiplet
- FlexGen Multi-Die Smart Network-on-Chip (NoC) IP
- Ncore Multi-Die Interconnect IP
- Integrated voltage regulator (IVR) chiplet
Related Videos
- Paving the Road Ahead: RISC-V and Chiplet Technologies in Modern Automotive and Data Center Architectures
- Exploring the Advancement of Chiplet Technology and the Ecosystem
- MSquare Technology: Enabling Packaging and Interconnect Chiplet IP
- Scaling Open Compute: RISC-V, Chiplets, and the Future of AI and Robotics
Latest Videos
- The European Chip Ecosystem
- Scaling up AI Compute, Memory and Connectivity
- CURIE: A Cycle-Accurate and Synthesizable UCIe Interface for the Agile Design and Co-Simulation of Chiplets
- Federated Chiplet Design and Simulation
- Custom HBM as a System: Co-designing the Base Die, Die to Die Power, and Package Thermals for AI Accelerators