Skip to content

Start Volume Shader Test

The GPU draws a ray marched volume. You watch live FPS.

Older phones can hitch. Start on Standard if you are unsure.

This browser cannot draw a canvas.

Drag to turn. Scroll to zoom. Right click to pan.

Your result

Your recent runs on this device

Best on : FPS

Scores are anonymous: GPU name, browser, OS and FPS only. · Privacy

Top GPUs on Standard

How your score compares: real runs, browser (WebGL)

Full leaderboard
#NameFPSRuns
1 NVIDIA GeForce RTX 5090 NVIDIA · Desktop / laptop GPU 182
28
2 NVIDIA GeForce RTX 5080 NVIDIA · Desktop / laptop GPU 135
264
3 NVIDIA GeForce RTX 4090 NVIDIA · Desktop / laptop GPU 126
39
4 NVIDIA GeForce RTX 4070 Ti SUPER NVIDIA · GPU 120
41
5 AMD Radeon RX 7900 XT AMD · Desktop / laptop GPU 118
27
6 AMD Radeon RX 7900 XTX AMD · GPU 114
28
7 NVIDIA GeForce RTX 5070 Ti NVIDIA · Desktop / laptop GPU 109
411
8 AMD Radeon RX 9070 XT AMD · Desktop / laptop GPU 102
452
9 NVIDIA GeForce RTX 4080 SUPER NVIDIA · Desktop / laptop GPU 102
42
10 AMD Radeon RX 9070 GRE AMD · Desktop / laptop GPU 101
29

What this test is doing

Volume Shader loads the GPU by ray marching a Mandelbulb in WebGL. You tap Start when you are ready. Live FPS is the main result. Each pixel walks a ray through a 3D field until it finds a surface. That loop is expensive because every pixel runs it, not only the silhouette of a mesh. For the name, the kernel, and Volume Shader BM, read What is Volume Shader.

How to read the numbers

FPS sits on the canvas while the test runs. Compare devices only when complexity, browser, and heat match. Bands and hitch notes are in what FPS means. A fair compare method is the GPU benchmark.

  • FPS: 60 and up is comfortable. 30 to 60 is usable. 15 to 30 is strained. Under 15 means this level is too heavy.
  • Complexity: Simple, Standard, Advanced, or Extreme under the canvas. Standard matches the classic Volume Shader load.
  • Fullscreen: Use fullscreen to record a clip. Escape or the same control exits.
  • Share: The header share button opens the system share sheet on phones that support it.
  • Your score: after a 2 second warmup the test measures 10 seconds of frames. The result card shows your average FPS, how many runs you beat, and how your GPU ranks.

Controls

  • Start / Stop: Start on the overlay, then the stop icon. Keyboard: P stops the test.
  • Complexity: Buttons under the canvas rebuild the shader. Changing level restarts the load.
  • Desktop: Left drag to rotate, scroll to zoom, right drag to pan. Space resets the view.
  • Phone: One finger rotates, pinch zooms, two fingers pan.
  • Advanced: Optional kernel editor for people who already know GLSL. Leave it alone for a fair comparison. Notes for that panel are in the BM test developers guide.

When to use Extreme

Standard is the comparison point. Extreme is a stress load. If you want heat and throttling, stay on Extreme for one to two minutes after you prep the device. Details are on GPU stress test.

When to pick another test

This canvas is ray marching, not raster 3D. Games almost never run a full screen Mandelbulb. If you want timed meshes and lights, open the 3D rendering test. If you want fragment iterations without a volume, open the shader test. If you want texture sampling pressure, open the VRAM test.

Name, GitHub, and BM

You do not need a GitHub project to run Volume Shader. Public demos popularized this Mandelbulb look years ago. History and naming, including Volume Shader BM, are on What is Volume Shader.