OpenVR Benchmark vs vkmark

OpenVR Benchmark is a Windows game that measures VR performance by rendering a repeatable test inside a VR headset. A run takes about 60 seconds and produces a score that is automatically compared with results from users who have the same GPU and headset. It ignores headset tracking and renders the same view each time, while shared SteamVR performance settings such as supersampling still apply. The benchmark supports SteamVR-compatible headsets, and controllers are not required. It is controlled with a mouse at a desktop monitor, and wearing the headset during a run is optional. Results can be exported as PNG images or copied as formatted text. The core benchmark is free. An optional Leaderboard DLC costs 329.00 EUR once and adds a full results leaderboard with hardware details, rank, and the option to submit a name. Listed minimum requirements include Windows 7 or newer, a 64-bit system, an AVX CPU, 4 GB RAM, a 3 GB GPU, DirectX 11, 4 GB storage, SteamVR, a keyboard, and a mouse.

vkmark is a free, extensible Vulkan benchmarking suite for Linux, built around targeted scenes that measure different aspects of Vulkan performance. Each scene can take configuration options. The `-b, --benchmark` command-line option lets users select scenes and their order; without it, vkmark runs a default set. The project lists X11, Wayland, and KMS backends. It can detect the window system automatically, or a user can choose one with `--winsys`. After installation, vkmark runs from the command line, and it can also be run directly from the project build directory without installation. Building requires Meson, a C++17 compiler, Vulkan development files, libglm, and libassimp, as well as dependencies for the selected backend. The documented build uses Meson and Ninja, with installation through `ninja -C build install`. The README includes an apt-install example for dependencies on recent Debian or Ubuntu systems. The repository includes the GNU Lesser General Public License, version 2.1.