Notebookcheck Logo

Think Silicon showcases RISC-V-based NEOX low-power GPU for wearables, AI applications and data centers

RISC-V is making its way to scalable GPUs. (Image Source: Think Silicon)
RISC-V is making its way to scalable GPUs. (Image Source: Think Silicon)
Think Silicon's NEOX GPU is a scalable chip based on the RISC-V architecture that can integrate 4 to 64 cores and can be utilized in a wide array of devices including smartwatches, smart home IoT appliances, electric car dashboards, and even large data centers.
AI GPU Internet of Things (IoT) Wearable

The interest in the RISC-V architecture was rekindled especially after the controversial announcement concerning Nvidia’s ARM buyout. Over the past few years we have seen several RISC-V CPU projects and it looks like this architecture can be adapted to GPUs, as well. Last year, researchers were able to enable Nvidia CUDA support on a RISC-V GPGPU, and just recently, Think Silicon presented a RISC-V GPU that can scale to power wearables and data centers alike.

Think Silicon’s RISC-V-based low-power GPU named NEOX was showcased at the Embedded World 2022 event in Nuremberg, Germany. As detailed on the official site, NEOX “is a parallel multicore and multithreaded GPU architecture based on the RISC-V RV64C ISA instruction set with adaptive NoC (network-on-chip). The number of cores varies from 4 to 64 organized in 1-16 cluster elements, each configured for cache sizes and thread counts. Depending on cluster / core configuration, NEOX compute power is ranging from 12.8 to 409.6 GFLOPS at 800MHz with support for FP16, FP32 and optionally FP64 and SIMD instructions.”

Due to its excellent scalability, the NEOX GPU can be configured to offer processing power for 2D / 3D computer graphics, machine learning / AI applications, computer vision, video coding and decoding, and general purpose compute tasks.

Going deeper into the architecture hardware details, the NEOX GPUs integrate AI specific ISA extensions, SIMD Vector in variable length datatypes including 8-bit and optionally Graphics ISA Extensions / Coprocessors: Unified Shader Architecture, Tile Based Rendering, Color/Vertex, Vector Support and contains dedicated hardware modules, such as rasterizer, texture unit, tile management unit and texture caches. Additionally, the NEOX SDK supports the C/C++ LLVM / GCC compilers, as well as the Posix Threads parallel execution model and developers may also add custom instructions and extensions via the Open Graphics Frameworks.


Buy the StarFive VisionFive kit with JH7100 64-bit RISC-V processors on Amazon

(Image Source: Think Silicon)
(Image Source: Think Silicon)

Source(s)

Think Silicon

static version load dynamic
Loading Comments
Comment on this article
Please share our article, every link counts!
> Expert Reviews and News on Laptops, Smartphones and Tech Innovations > News > News Archive > Newsarchive 2022 06 > Think Silicon showcases RISC-V-based NEOX low-power GPU for wearables, AI applications and data centers
Bogdan Solca, 2022-06-23 (Update: 2022-06-23)