VolumeShader_BM Leaderboard

This leaderboard compares accepted results from the fixed VolumeShader_BM benchmark workflow. It is not a ranking for every shader experiment made in the editable lab. The benchmark page locks the workload into known presets, records FPS and frame-time behavior, and routes each eligible submission into a specific comparison bucket. That bucket structure is why the directory is useful: it lets a reader choose the same device class, workload, and rendering API before interpreting any number.

Desktop and mobile devices are separated because they usually operate under different thermal envelopes, power limits, browser restrictions, and GPU architectures. A phone can throttle quickly under a long shader-heavy run, while a desktop GPU may sustain a higher load with more cooling headroom. The VolumeShader_BM leaderboard keeps these classes separate so the FPS table is easier to interpret and a mobile result is not judged against hardware that was allowed to draw much more power.

The Low, Balanced, cz github reference, High, and Extreme presets represent different workload levels. A Low result is meant to cover more devices, while High and Extreme put more pressure on stronger GPUs. The VolumeShader_BM leaderboard treats each preset as its own comparison lane because a lower preset can show higher FPS simply because it asks the browser to render a lighter scene.

WebGL and WebGPU are different rendering backends. They can use different browser code paths, driver behavior, scheduling, and feature support. The same GPU can therefore score differently between WebGL and WebGPU, so the VolumeShader_BM leaderboard keeps those results in separate buckets. If you are checking browser API behavior, compare the WebGL and WebGPU pages for the same device class and preset instead of mixing every API into one table.

A fair comparison uses the same device class, preset, and API. Mixing Mobile Low WebGPU with Desktop High WebGL hides the real question because both the hardware category and the rendering workload changed. Open the VolumeShader_BM leaderboard bucket that matches your run, then compare only entries inside that bucket. The segmented navigation is designed around that workflow: each link is a real URL to a complete, indexable bucket.

VolumeShader_BM leaderboard pages show accepted results after the aggregation step has processed submitted data. Each detail page summarizes accepted entries, top FPS, mean FPS, leading GPU strings, and recent rows. If a bucket has no accepted entries yet, the page still exists as a valid bucket and shows a waiting state until comparable submissions arrive.

Start from the volumeshader_bm benchmark page, choose the preset and rendering API that match the VolumeShader_BM leaderboard bucket you want to enter, keep the session running long enough for the metrics to stabilize, then submit the eligible result. Closing other GPU-heavy apps and avoiding power-saving mode usually makes the run easier to compare.

Higher FPS is useful, but it is not a complete statement about overall GPU quality. It describes performance in this shader-heavy browser scene under the submitted conditions. Frame time, stability, browser context, backend, driver behavior, thermals, and system power mode can all change the result, so use the VolumeShader_BM leaderboard as focused benchmark evidence rather than a universal hardware verdict.