Monday, May 27, 2024

RISC-V GPU Advances to Manage CPU and NPU Tasks

A new RISC-V microprocessor is capable of handling CPU, GPU, and NPU tasks concurrently.
A novel RISC-V micro processing chip design has been developed by X-Silicon Inc. (XSI) that integrates a RISC-V CPU core, RISC-V GPU acceleration, and vector capabilities onto a single device.

According to Jon Peddie Research, the CPU/GPU hybrid device is open-standard and would presumably be open-source as well. It is intended to perform a range of tasks, such as artificial intelligence (AI), that would typically be performed by dedicated CPUs and GPUs. The problem is that it is meant to accomplish all of this with significantly greater efficiency.

The new hybrid CPU/GPU is intended to be a “jack of all trades” machine. JPR claims that the industry has been looking for an open-standard RISC-V GPU that is adaptable and scalable enough to serve a range of sectors, such as IoT devices, virtual reality, and cars. The goal of this new RISC-V CPU and RISC-V GPU is to give manufacturers an open chip architecture that can handle whatever workload they want.

The chip from X-Silicon is unique from previous architectures since it integrates a GPU and CPU into a single-core architecture. This isn’t like the standard AMD and Intel designs, which have distinct GPU and CPU cores. Rather, the core is built to be able to handle tasks on both the CPU and the GPU. It sounds a lot like Intel’s shelved Lara Bee project in that regard, which tried to employ x86 for workloads such as graphics.

The chip makes use of the C-GPU architecture from X-Silicon, which combines GPU acceleration with a vector CPU core that is RISC-V. A 32-bit FPU and Scalier ALU are located in the RISC-V vector core of the architecture. Thread scheduler, Pasteurizer, Clipping Engine, Texture Unit, Neural Engine, and Pixel Processor are some of its features. The device is designed to handle applications such as 2D and 3D graphics, geometric computation, AI, and high-performance computing (HPC).

Theoretically, X-Silicon’s hybrid chip has numerous benefits due to its ability to process both CPU and RISC-V GPU code within its single core. The chip runs a single instruction stream on the open-standard RISC-V ISA for both the CPU and GPU. This results in higher performance and low-memory footprint execution since there is no data copying between the GPU and CPU memory spaces.

Manufacturers can increase processing power as needed by combining the CPU and RISC-V GPU cores into a multi-core design. A fast fabric is used to connect several cores that are tiled across a chip in a multi-core configuration. This also has fast on-chip SRAM or e DRAM caches, which function as an L2 cache capable of combining data from several cores. When necessary, each core can be scheduled to execute workloads unrelated to the others, such as physics, AI, graphics, video, and HPC.


This architecture may allow X-Silicon’s C-GPU architecture to handle any kind of CPU or GPU workload. It is claimed by X-Silicon that Vulkan graphics API is already functioning with “fused RISC-V GPU acceleration.” This ought to be quite beneficial for its advancement and uptake on Android gadgets.

Unlike x86 and ARM, the new design is based on RISC-V, which means that anyone can use the architecture without having to pay instruction-set license fees. The chips could revolutionize the microprocessor business if they function as planned. Theoretically, what X-Silicon claims to have built is more versatile and competent than the existing conventional designs.

Although workloads from both CPUs and RISC-V GPU can be handled by a new RISC-V micro processor, the following summarizes the main distinctions between them:

Building A RISC-V CPU Core

  • All-purpose: Designed to perform an extensive range of functions, such as multitasking, managing system resources, and executing applications.
  • Sequential processing: Carries out commands one at a time. more effective when dealing with complicated branching and logic tasks.
  • Fewer cores: Usually has fewer cores (four to sixteen) that are better suited to specific activities.

Graphics Processing Unit:

  • Specialized processor: Made to handle graphics and video data quickly, especially for jobs requiring a lot of data to be processed in parallel.
  • Parallel processing: Is perfect for tasks involving repetitive procedures because it can handle multiple calculations at once.
  • More cores: Frequently contains a sizable number of parallel processing-optimized cores, in the hundreds or even thousands.

CPU vs GPU Upgrade

CPU: Comparable to an experienced chef who manages several meals at once, alternating between them and making sure each is prepared to perfection.
GPU: Similar to a group of line cooks who work together to efficiently prepare big amounts of food by assigning each person to a specialized duty on an assembly line.
Even if the RISC-V chip you described is capable of handling both workloads, for best results, it’s critical to grasp the advantages and disadvantages of each type of CPU.

CPU vs GPU comparison

Here is a table summarizing the CPU vs GPU comparison:

Processing StyleSequentialParallel
Core CountFewer (4-16)More (Hundreds/Thousands)
StrengthComplex logic, multitaskingLarge-scale parallel processing
Agarapu Ramesh was founder of the Govindhtech and Computer Hardware enthusiast. He interested in writing Technews articles. Working as an Editor of Govindhtech for one Year and previously working as a Computer Assembling Technician in G Traders from 2018 in India. His Education Qualification MSc.


Please enter your comment!
Please enter your name here

Recent Posts

Popular Post Would you like to receive notifications on latest updates? No Yes