AMD’s RDNA 4 architecture has been developed from the ground up to enable a bevy of enhancements that cater to gamers, such as upgraded ray tracing and AI cores, new compute units, and more. The same architecture was leveraged to launch high-performance products, such as the Radeon RX 9070 XT, with a customized version used for Samsung’s newly announced Exynos 2600. The Xclipse 960 graphics processor, paired with the world’s first 2nm GAA chipset, features a customized AMD RDNA 4 architecture, enabling several performance and feature improvements over the previous-generation Exynos 2500.
The customized AMD RDNA 4 architecture for the Exynos 2600 is called ‘MGFX4,’ allowing for the development of ENSS, or Exynos Neural Super Sampling technology
The Exynos 2500 was developed using AMD’s older MGFX3 architecture, which was a customized version of RDNA 3. The Elec reports that the processor maker’s IP has been scaled down and modified for the Xclipse 960, and is referred to as the MGFX4 architecture. During the official announcement, Samsung stated that the Exynos 2600 not only delivers twice the performance of the Exynos 2500, but its ray tracing performance has improved by 50 percent.
Additionally, utilizing AMD’s MGFX4 architecture allows for the introduction of Samsung’s proprietary upscaling technology called Exynos Neural Super Sampling (ENSS), which behaves similarly to NVIDIA’s DLSS and AMD’s FSR. The Exynos 2600’s specifications include 8 Work Group Processors (WGPs), which are computational block units designed to reduce latency and improve Instructions Per Clock (IPC). One WGP typically comprises of two Compute Units (CUs), with the maximum clock speed operating at 980MHz.
However, compared to the Exynos 2500’s Xclipse 950 GPU, Samsung has slightly decreased the core frequency from 999MHz to 980MHz while retaining the WGP at 8 units. On paper, the Snapdragon 8 Elite Gen 5 is superior to the Exynos 2600, sporting 12 CUs, along with a maximum core clock speed of 1,200MHz. Interestingly, the report states that the Xclipse 960 lags behind the Snapdragon 8 Elite Gen 5’s Adreno 840, with Geekbench 6’s OpenCL and Vulkan tests showing a delta of between 10 percent and 20 percent.
Geekbench 6 OpenCL
- Exynos 2600 - 22,000 points
- Snapdragon 8 Elite Gen 5 - 23,900
Geekbench 6 Vulkan
- Exynos 2600 - 22,800 points
- Snapdragon 8 Elite Gen 5 - 27,600
With the Exynos 2600 reported to be found in the Galaxy S26 and Galaxy S26+, it remains to be seen how big a vapor chamber Samsung will add to give the chipset and GPU a performance advantage. To address overheating concerns, the Korean giant has stated that it has implemented ‘Heat Pass Block’ technology, which improves heat flow and lowers thermal resistance by 16 percent. With the flagships rumored to launch in March 2026, we should have a better idea of the Exynos 2600’s capabilities, so stay tuned for more updates.
News Source: The Elec
About the author: Omar Sohail is a reporter and analyst for Wccftech's mobile section, specializing in the technology and business of the mobile industry. His expertise lies in the intricate hardware supply chain, covering developments in semiconductor manufacturing, chip lithography, and camera sensor technology.
Follow Wccftech on Google to get more of our news coverage in your feeds.
