Neurxcore provides the SNVDLA line of configurable neural processors, based on an enhanced open‑source NVDLA architecture, that can be tuned for compute, power, and feature requirements across edge, mobile, and data‑center applications. The company also offers the Heracium software suite and integration services to optimize AI inference on embedded systems, supporting multiple OSes, CPUs, and machine‑learning frameworks.
Funding
Funding not disclosed
Founders
Product
Problem
Designing AI inference hardware for edge, mobile, and data‑center devices often requires custom silicon that balances performance, power consumption, and cost, while also supporting diverse operating systems, CPUs, and machine‑learning frameworks. Existing solutions can be inflexible, limiting the ability to tailor neural processors to specific application requirements.
Solution
Neurxcore offers the SNVDLA line of configurable neural processors built on an enhanced version of NVIDIA’s open‑source Deep Learning Accelerator (NVDLA) microarchitecture. The processors are highly tunable and scalable, allowing designers to adjust compute resources, power budgets, and feature sets to match the target workload. Neurxcore also provides the Heracium software environment and dedicated services to integrate, optimize, and validate AI inference on embedded systems, ensuring efficient deployment across a wide range of applications from wearables to data‑center accelerators.
Target Audience
Primary customers are semiconductor companies, system‑integrators, and OEMs developing AI‑enabled chips or embedded platforms for applications such as IoT devices, wearables, AR/VR, robotics, ADAS, surveillance, and data‑center accelerators.
Features
- Configurable and scalable architecture derived from the open‑source NVDLA, enabling customization of compute units and memory hierarchy
- Power‑efficient design capable of operating at milliwatt levels while delivering hundreds of TOPS of inference performance
- Support for multiple operating systems, CPUs, and major machine‑learning frameworks, facilitating seamless integration into existing software stacks
- Silicon‑qualified implementation with demonstration boards available for rapid prototyping and validation
- Heracium software suite for performance tuning, model optimization, and system-level integration of AI workloads
- Dedicated hardware/software services to assist customers in optimizing SNVDLA‑based AI systems