Tech giant AMD quietly revealed a new ray tracing graphical technique that significantly decreases memory usage and allows millions of animated triangles to run at a steady frame rate. However, it remains unclear if or when this graphical research breakthrough from AMD will make a practical difference in games.
As one of the leading faces of the tech industry, AMD is the manufacturer of Ryzen processors and Radeon graphics cards. Additionally, AMD hardware is utilized by PS5 and Xbox Series X/S consoles. Despite AMD’s technological prowess, the PC graphics card market has been largely dominated by its chief competitor, Nvidia. However, Nvidia’s flagship card, the RTX 5090, holds a staggering sale price of over $6,500 from third-party sellers on Amazon and Newegg. By comparison, AMD’s Radeon RX 9070 XT is available directly from retailers starting at $799.
AMD’s Tetrahedral Cage Approach May Reduce GPU Video Memory Usage
-
Image via AMD -
Image via AMD -
Although AMD users already benefit from Microsoft’s Advanced Shader Memory technology, a new ray tracing technique may allow the gap between AMD and Nvidia to close more. AMD Fellow Holger Gruen stated via GPUOpen that AMD’s Tetrahedral Cage Approach would allow developers to ray trace around animated unique cages and reduce the video memory demand from the workload. AMD shared a technical video demonstration of a dense grass field and trees that used tetrahedral cages. While this research demonstration is not a feature already being used in games as of this writing, the technique could allow developers to make more complex and detailed environments without using as much memory.
Find the odd one out before the timer hits zero.
AMD stated that the demo ran at over 60 frames per second (FPS) at 1080p using a Radeon RX 9070 XT, which has 16GB of GDDR6 video memory. While the traditional clustered approach offers developers more precise vertex animation control, it also comes with a steep video memory demand. By comparison, the tetrahedral cage technique used in the video only took up 1.7 GB of video memory. Should the clustered technique have been used instead, it would have required over 80GB of video memory. In short, both the traditional cluster and tetrahedral cage techniques can be used for different purposes, with the latter being potentially preferable for ray tracing foliage and distant characters.
Should game developers adopt AMD’s tetrahedral cage approach, it would also have an effect on current and future console generations. Both Microsoft and Sony stated that AMD hardware will, respectively, be utilized for Project Helix and the PS6. The decision from both companies will continue a trend that started in 2013 during the PS4 and Xbox One eras. AMD stated that the PS6 and Project Helix will use custom AMD processors that will have RDNA 5 graphics and Zen 6 CPU architecture built into both consoles.
With the ongoing DRAM shortage crisis causing the prices of RAM sticks, graphics cards, and solid-state drives to skyrocket, any attempt to reduce video memory usage would be beneficial to gamers running on older or lower-end hardware. It remains to be seen how many developers will utilize AMD’s research heading into the next generation of gaming hardware.