X-Silicon Unveils AI Core with Integrated GPU and RISC-V Vector CPU

San Diego startup X-Silicon has made waves in the tech industry with its groundbreaking IP core that merges a RISC-V processor with an open standard Vulkan GPU, specifically designed for AI and machine learning applications.

The NanoTile "C-GPU" architecture, developed by X-Silicon, integrates GPU acceleration into a RISC-V Vector CPU Core, featuring tightly coupled memory to create a monolithic processor core capable of running the Vulkan open graphics standard.

The C-GPU is available as IP blocks and software for processor companies looking to develop low-power SoCs for the next generation of 3D graphics, compute, and AI/ML-enabled devices such as wearables, AR/VR headsets, auto displays, edge and cloud processing systems, industrial applications, robotics, and various other connected embedded/IoT devices.

X-Silicon has already initiated collaborations with SiFive and Ventana Micro Systems to work on chip designs utilizing the IP. Ventana Micro Systems, known for incorporating GPU cores from Imagination Technologies in the UK for its RISC-V AI chiplet designs, is among the early adopters of X-Silicon's innovative technology.

In a strategic move, X-Silicon has announced plans to open-source its unified RISC-V Vector CPU-with-GPU Instruction Set Architecture (ISA) and provide register-level access to its hardware through a Hardware Abstraction Layer (HAL). This approach aims to empower chip customers, enabling OEMs and content providers to customize drivers and applications to a greater extent than currently possible.

Moreover, X-Silicon has achieved a significant milestone by being the first to introduce Vulkan on RISC-V with integrated GPU acceleration, making it a viable solution for Android devices seeking enhanced graphics capabilities.

The X-Silicon C-GPU is engineered to optimize the rendering and management of dynamic content without relying on traditional triangle-based graphics rendering methods. By leveraging a single C-GPU processor with multiple NanoTiles, this architecture is well-suited for implementing more efficient graphics, AI, and compute algorithms, breaking free from the constraints of existing GPU technologies.

The NanoTile architecture, supported by 14 patents covering AI/ML Compute and GPU rendering using a single processor, focuses on enhancing data transfer efficiency among processing cores and memory, thereby reducing latencies and improving overall compute performance.

Multiple C-GPU cores are distributed across a chip and interconnected via an on-chip fast compositor fabric capable of aggregating outputs from each core dynamically. This aggregation can serve various purposes, such as creating a frame buffer for graphics or pipelined buffers for codec, video effects processing, and AI tasks, with each core programmable for different workloads.

By utilizing fast on-chip SRAM or eDRAM as Level 2 caches to aggregate data from multiple cores, the NanoTile architecture enables in-memory computing, where common operations are executed close to memory, reducing bandwidth requirements and enhancing performance.

Workloads can be executed in parallel or pipelined, allowing tasks to run simultaneously on a core rather than sequentially, as seen in traditional GPU architectures. This flexibility also enables running an operating system directly on a core, expanding the possibilities for edge and cloud processing applications.

X-Silicon's CEO, Dan Nilsson, envisions a future where enhanced visual experiences are seamlessly integrated into industrial and consumer devices, catering to users accustomed to sophisticated mobile phone displays for UI interactions. The company aims to empower edge devices with enhanced intelligence while also enabling low-power visual computing in the cloud, surpassing the capabilities of closed systems.

Industry analyst Jon Peddie commends X-Silicon for its dynamic and scalable C-GPU NanoTile architecture, highlighting its ability to handle multiple tasks concurrently or sequentially, including OS operations, applications, graphics rendering, compute tasks, and AI/ML processing, setting the company apart in the industry.

Collaborating with industry giants like SiFive and Ventana Micro Systems, X-Silicon is poised to revolutionize the RISC-V ecosystem with its Vulkan-enabled graphics rendering solution, offering new possibilities for edge platforms seeking advanced 3D graphics capabilities not available with proprietary GPU solutions.

As X-Silicon prepares to release its software development kits to early development partners later this year, the tech world eagerly anticipates the transformative impact of the company's Open-Standard Vulkan-enabled GPU architecture on the future of AI, graphics, and edge computing.

For more information, visit www.x-silicon.com