Basemark Releases Breaking Limit Cross-Platform Ray Tracing Benchmark

Basemark announced today the release of a groundbreaking cross-platform ray tracing benchmark, GPUScore: Breaking Limit. This new benchmark is designed to evaluate the performance of the full range of ray tracing capable devices, including smartphones, tablets, laptops and high-end desktops with discrete GPUs. With support for multiple operating systems and graphics APIs, Breaking Limit provides a comprehensive performance evaluation across various platforms and devices.

As ray tracing technology becomes increasingly prevalent in consumer electronics, from high-end desktops to portable devices like laptops and smartphones, there is a critical need for a benchmark that can accurately assess and compare performance across different devices and platforms. Breaking Limit addresses this gap, providing valuable insights into how various devices handle hardware-accelerated graphics rendering. The benchmark is an essential tool for developers, manufacturers, and consumers to measure and compare the performance of real-time ray tracing rendering across different hardware and software environments reliably.

“As ray tracing continues to gain traction on all device categories, it was a natural progression for us to develop a benchmark that reflects the current and future landscape of this technology,” said Mircea Cristea, Product Owner at Basemark. “We are excited to introduce Breaking Limit, a unique benchmark which enables face-to-face comparison across all the types of ray tracing supporting devices.”

Breaking Limit comes with two distinct workloads: Breaking Limit and Breaking Limit Ultra. Breaking Limit is ideal for comparing mobile devices, laptops, and desktops equipped with ray tracing supporting GPUs. Breaking Limit Ultra features an exceptionally demanding workload designed specifically for desktops with high-end GPUs. It also includes performance evaluation tests for the most known AI upscaling techniques: FSR (FidelityFX Super Resolution) and DLSS (Deep Learning Super Sampling).