Intel has recently unveiled a patent for a split multi-chipset GPU design, indicating a significant development in the competitive landscape of multi-chip GPUs.
According to WCCF Tech, this patent emphasizes the concept of utilizing multiple specialized small chips rather than relying on the traditional monolithic architecture commonly found in graphics cards. This development highlights Intel’s commitment to exploring multi-chip GPU technology, positioning itself in direct competition with AMD in the graphics card arena.
Multi-Chip GPU: Emerging Future in Graphics Technology
Intel’s new split GPU architecture proposes the segmentation of the GPU into several smaller interconnected chips. Each chip would be assigned a specific function, focusing on areas such as graphics processing, computing, or artificial intelligence (AI). This innovative approach presents several advantages for users, including:
- Enhanced Performance: Each chip can be optimized for its designated task, potentially improving the overall performance of the GPU.
- Energy Efficiency: Smaller chips can be powered down when not in use, leading to significant energy savings.
- Modular Flexibility: The design’s modular nature allows for easier replacement or upgrading of individual chips.
While AMD has previously employed small chip designs in its EPYC CPUs, there currently aren’t any commercially available multi-chip GPUs. Both Intel and AMD are beginning to recognize the substantial potential of this technology, paving the way for the creation of more powerful and efficient graphics solutions.
However, the realization of multi-chip GPUs involves overcoming various challenges associated with manufacturing and interconnect technology. Intel’s new patent signifies a step forward, suggesting that multi-chip GPUs might soon become a practical reality in the market.