I wasn't sure quite where to put this article. If anyone believes it should be in the benchmarking forum instead just let me know and I'll move this thread. 
 
https://www.overclock3d.net/news/software/variable_rate_shading_is_coming_to_3dmark/1
 
Today, both Nvidia (Turing and Newer) and Intel (Gen 11 and Newer) have integrated hardware support for a new feature called Variable Rate Shading (VRS), a technique which is designed to improve system performance by selectively reducing the level of detail of parts of a frame that will have a minimal impact on perceivable image quality. High shading rates will increase accuracy at the cost of performance, while lower shading rates will boost performance, but at the cost of graphical quality. 

As the name suggests, Variable Rate Shading allows the shading rate of a frame to vary, allowing developers to lower the shading rate of unimportant sections of a frame, increasing performance while delivering an almost unnoticeable drop in image quality. This promise of improved performance is why VRS has been integrated into the DirectX 12 API. 

UL Benchmarks has today promised to bring a VRS feature test to its 3DMARK gaming benchmark utility, showcasing the feature to both gamers and developers alike while enabling reviewers and analysts to judge the performance and quality impact of the effect. 
 
3DMARK's VRS feature test is due to arrive on August 26th as a free update to the utility. The feature test will first play the forest scene pictured above without VRS and then with VRS, enabling gamers to judge both the performance and image quality impacts of VRS. 

Why have it as a test? It really seems like VRS would depend on the game and what is rendered to really give a sense of how well it would work. It could work very well in the 3Dmark test but not be very good in a game. In fact it could cause issues with games like we saw people complain about with DLSS before it was patched.