Products
Explore, optimize, and deploy scalable interconnect architectures through a unified workflow for physically aware design, modular implementation, and faster time to market.
WeaverPro is Baya’s foundational software platform for cache and fabric architecture, enabling engineering teams to explore, optimize, and deploy scalable interconnect systems for advanced SoCs, AI infrastructure, and chiplet-based architectures. Built around workload-driven analysis and KPI-focused optimization, WeaverPro helps architects evaluate real system behavior earlier in the design process, reducing bottlenecks, accelerating architecture decisions, de-risking deployment, and enabling faster time to market. The platform connects cache hierarchy exploration and fabric microarchitecture optimization through a unified architecture-to-deployment workflow, linking system specification, workload-driven analysis, physically aware optimization, modular implementation, and deployment readiness.
Through fast simulation, static analysis, and physically aware design exploration, WeaverPro enables teams to analyze critical metrics such as bandwidth, latency, congestion, miss rates, and topology efficiency while operating within physical implementation constraints. Its correct-by-construction methodology validates architectural correctness before simulation and guarantees deadlock freedom through built-in algorithms, helping teams reduce design risk, improve predictability, and move toward implementation with greater confidence. By replacing fragmented workflows with a connected platform for exploration, optimization, and deployment, WeaverPro helps engineering teams accelerate scalable system design for modern compute architectures.
Analyze real system behavior early in the design process using fast simulation and KPI-focused exploration to optimize bandwidth, latency, congestion, and overall architecture efficiency.
Validate architectural correctness before simulation with built-in static analysis and deadlock-free algorithms that help reduce design risk and accelerate deployment readiness.
Explore cache and fabric architectures within physical implementation constraints, enabling scalable, modular implementation and faster time to market for advanced compute systems.
CacheStudio accelerates cache and memory architecture exploration through workload-driven analysis, enabling engineering teams to optimize bandwidth, latency, cache efficiency, and data movement for advanced SoCs and scalable chiplet-ready systems.
FabricStudio enables data-driven fabric microarchitecture exploration and optimization through static analysis, dynamic simulation, and physically aware design planning, helping teams create scalable interconnect architectures with modular implementation and de-risked deployment.
Move toward implementation with greater confidence through early validation, predictable architecture behavior, and correct-by-construction design methodologies.
Replace fragmented workflows with a connected architecture-to-deployment platform that accelerates exploration, optimization, and implementation readiness.
Optimize cache, memory, and fabric architectures for advanced SoCs, AI infrastructure, and chiplet-based systems while balancing bandwidth, latency, power, and physical constraints.
Featured Case Study
Baya Systems and Tenstorrent collaborated on a scalable RISC-V compute subsystem reference design for AI and high-performance computing applications. Using WeaverPro FabricStudio, the teams explored workload-driven NoC optimization, cache hierarchy design, and physically realizable interconnect architectures to evaluate demanding bandwidth, latency, and scalability requirements across multiple workload scenarios. The resulting reference platform demonstrated how workload-driven exploration and unified coherent and non-coherent interconnect design can support scalable compute architectures while enabling modular implementation and faster time to market for advanced SoCs and chiplet-based systems.
Resources
Bridging Compute and Fabric IP_ A New Era for RISC-V System Design (with Baya Systems & Tenstorrent)
Explore how WeaverPro brings cache and fabric architecture development into a unified, workload-driven design platform. The brief outlines its Python API and graphical interface, end-to-end design flow, static correctness analysis, system-level simulation, KPI-based optimization, and integration between CacheStudio, FabricStudio, and Baya’s fabric IP portfolio.
White paper
As demand grows for scalable, efficient, and customized compute, more companies are turning to RISC-V as the preferred architecture for high-performance computing. Tenstorrent and Baya Systems have designed a compute subsystem combining IP from both companies designed to enable AI and HPC use cases. The solution leverages Tenstorrent’s Ascalon™ processor and Baya’s advanced interconnect technology. This paper showcases a compute subsystem that has been designed to deliver strong performance, power efficiency, flexibility and scalability, making it suitable for cloud, server, data center, and edge computing applications to be used in SoCs and chiplets. This collaboration also enables rapid customization and extension of well-tested compute subsystems.
Webinar
The future of AI-driven compute depends on how well data can move. Traditional network-on-chip (NoC) architectures are hitting their limits, and the next generation of RISC-V subsystems demands faster, more intelligent interconnects.
In this live engineer-focused session, Baya Systems and Tenstorrent will explore why collaboration between compute and fabric IP is now critical for performance and scalability in modern SoC design. You’ll see how combining Ascalon processors and Tenstorrent System IP with Baya’s RISC-V Fabric IP enables more efficient data movement, stronger interoperability, and streamlined subsystem integration.
The session will include a live interoperability demo showing the Ascalon + Baya Fabric IP reference design in action.
What You’ll Learn
CacheStudio is a rapid architecture and design platform for chiplet based cache coherent systems. Cache hierarchies and parameters are specified in minutes with an abstract Python frontend, and accurate workload driven simulations are performed at Millions of instructions per second to uncover stateful performance indicators that require long runtimes. The results are presented interactively at multiple abstraction levels and organized structurally with the cache hierarchy and the chiplet organization, for intuitive comprehension and analysis.
