Shader Performance Test

A Shader Performance Test is a hardware GPU benchmark that evaluates graphics processing efficiency by executing computationally intensive GLSL fragment shaders with procedural 3D ray-marching, multi-octave noise, and specular lighting in real time.

Yes. WebGL2 compiles our GLSL ES 3.00 shader code directly into native machine instructions on your graphics hardware. All ray-marching math, normal gradients, and lighting calculations execute directly on your physical GPU cores.

A fragment shader is a GPU program responsible for computing the final color, lighting, reflections, and transparency of each screen pixel covered by a rendered primitive.

Higher complexity introduces additional procedural noise octaves, multi-pass normal gradient calculations, and secondary reflection rays. Multiplying the math operations required per pixel eventually saturates GPU arithmetic pipelines.

Iterations define the loop boundary inside the ray-marching fragment shader. Higher iterations (such as 128 to 512) allow rays to traverse deeper geometric fields and refine surface convergence, multiplying floating-point ALU cycles.

WebGL2 is a modern web standard based on OpenGL ES 3.0 that brings advanced GPU shading capabilities, 32-bit floating-point textures, uniform buffers, and GLSL ES 3.00 directly to modern web browsers without plugins.

Different browsers use different graphics translation layers (e.g. DirectX ANGLE on Windows Chrome/Edge vs native Metal on macOS Safari). Browser compositors and GPU driver hooks also introduce small scheduling variations.

Yes. For integrated graphics (like Intel Iris Xe, AMD Radeon 780M, or mobile SoCs), start with Low or Medium complexity with 32 to 64 iterations for optimal stability.

No. Standard browser security sandboxes prevent web applications from querying hardware motherboard sensors or thermal diodes. We never fabricate fake temperature or clock speed figures.

As the 3D procedural camera orbits, the volume of screen space covered by complex fractal surfaces shifts. Operating system background processes and browser garbage collection cycles can also cause minor frame time variances.

Yes. Because our benchmark uses deterministic procedural mathematical algorithms and continuous time-based formulas, running identical configurations (e.g. High complexity, 128 iterations, 100% scale, 30s) produces a repeatable workload for comparison.

Render scale adjusts the internal canvas pixel resolution. Selecting 150% super-samples pixels, forcing your GPU to compute 225% more fragment shader evaluations per frame, stressing fillrate and memory bandwidth.

1% Low FPS is the average frame rate of the slowest 1% of frames recorded during the benchmark run. It highlights brief stutter spikes and frame pacing instability that aggregate average FPS numbers might obscure.