YuvalGGYuvalGG

Tools

Graphics test

See which GPU string the browser reports, then run a short WebGL2 shading and fill-rate benchmark. Scores are relative and uncalibrated — useful for a quick check, not a lab ranking.

Graphics test

Detects the GPU string the browser exposes, then runs a ~15 s WebGL2 shading + fill benchmark at internal 1920×1080. Scores are relative / uncalibrated — not a lab 3DMark result.

Detected GPU

Waiting for WebGL2…

No catalog match for this renderer string. Browse GPU compare

Progress

Idle — press Start test.

Preview of the internal 1920×1080 canvas (scaled to fit). Work runs on the GPU via WebGL2.

Overall

Relative / uncalibrated

Shading power

—/ 100 relative

— (browser-estimated, uncalibrated)

Fill rate

—/ 100 relative

— (browser-estimated, uncalibrated)

Browser WebGL2 paths differ from native DirectX/Vulkan/Metal games. Drivers, power limits, thermal throttling, background apps, and discrete vs integrated routing all change results. Use this as a quick relative check, then compare catalog specs in GPU compare.

How GPU detection works

Modern browsers can expose a vendor and renderer string through WebGL. This tool prefers the WEBGL_debug_renderer_info extension and reads UNMASKED_VENDOR_WEBGL plus UNMASKED_RENDERER_WEBGL. When the extension is blocked or missing, it falls back to gl.VENDOR and gl.RENDERER, which are often more generic. Privacy settings and some enterprise policies can hide the real adapter name entirely.

If the renderer text fuzzy-matches a card in our GPU compare catalog — short names like RTX 5080, RX 7900 XTX, or Arc B580 — you get a direct compare link. Matching prefers longer Super / Ti names so a “4070 Ti Super” does not collapse to a plain “4070”.

How the score stays uncalibrated

The benchmark runs two phases at an internal 1920×1080 framebuffer (the on-page canvas can display smaller):

  • Shading power — heavy fragment-shader arithmetic. We count estimated floating-point work and report approximate GFLOPS labeled as browser-estimated and uncalibrated.
  • Fill rate — many translucent fullscreen layers per frame. We report approximate Gpixels/s, also uncalibrated.

Each raw metric is placed on a log scale between loose reference anchors defined only as mapping points in code comments — not as claims about specific retail cards. Sub-scores combine into an overall 0–100 number marked relative / uncalibrated. Bands such as Weak, Basic, Good, Fast, and Very fast add plain-English context for games without inventing frame-rate numbers. This is not 3DMark and should not be cited as a lab result.

Tips for a fairer run

  • Plug in a laptop and use the performance power profile when you care about peak numbers.
  • Close other GPU-heavy apps, overlays, and browser tabs that might steal cycles.
  • On dual-GPU machines, confirm the browser is routed to the discrete adapter (OS graphics settings or GPU control panel).
  • Enable hardware acceleration in the browser; a software renderer warning means the score is CPU-bound.
  • Run twice after a short warm-up; the first pass can look different while clocks settle.

Limits of browser WebGL

  • WebGL2 is a constrained, sandboxed path — not the same as a native game engine.
  • Driver bugs, ANGLE translation, and browser caps can under-utilize high-end cards.
  • Thermal and power limits still apply; a short ~15 s run can differ from a long gaming session.
  • Context loss (driver reset, tab discard) aborts the run — refresh and retry.
  • Detection strings can be spoofed, redacted, or wrong after remote desktop / VM setups.

For buying and build context, see graphics cards, RTX 5070 vs 5070 Ti, and RTX 5090. Pair this test with the bottleneck estimator, PSU wattage calculator, and frame-time calculator.

Common questions

How does GPU detection work?

The page creates a WebGL2 context and reads WEBGL_debug_renderer_info when the browser allows it (UNMASKED_VENDOR_WEBGL and UNMASKED_RENDERER_WEBGL). If that extension is missing, it falls back to gl.VENDOR and gl.RENDERER. Some browsers intentionally hide or genericize these strings for privacy.

Is the 0–100 score a real 3DMark-style result?

No. Every score on this page is relative and uncalibrated. Shading and fill metrics are mapped through a log curve between loose browser reference anchors defined in code comments. They are not claimed to be measurements of specific retail cards, and they are not comparable to lab suites.

Why do I see a software-render warning?

If the renderer string matches known CPU software paths (for example SwiftShader, llvmpipe, softpipe, or Basic Render), the browser is not using your discrete GPU. Turn on hardware acceleration, check dual-GPU laptop routing, then run the test again. A software path produces a weak score that reflects the CPU, not the card.

What do shading power and fill rate measure here?

Shading power stresses fragment-shader arithmetic and reports an approximate GFLOPS estimate from work completed over time. Fill rate draws many translucent fullscreen layers and reports approximate Gpixels/s. Both are browser-estimated and uncalibrated.

Why might my gaming PC score lower than expected?

WebGL2 is not the same path as native DirectX, Vulkan, or Metal games. Drivers, power limits, thermal throttling, background GPU apps, VSync-like pacing, and integrated-vs-discrete routing all change results. Treat the number as a quick relative check in this tab only.

Can this page match my card in GPU compare?

If the unmasked renderer string fuzzy-matches a short name in our catalog (for example RTX 5080, RX 7900 XTX, Arc B580), you get a link to GPU compare with that card preselected. Super and Ti variants prefer the longer name. If nothing matches, the generic GPU compare link still works.

What if WebGL2 is missing?

The tool needs WebGL2. Update your browser, enable hardware acceleration, and retry. Without WebGL2 the page shows a friendly error instead of running the benchmark.

  • Sensitivity converter — Convert sensitivity between Overwatch, Valorant, CS2, and other titles using documented yaw values.
  • cm/360 calculator — Turn mouse DPI and in-game sensitivity into centimeters per 360° turn.
  • eDPI calculator — Multiply mouse DPI by in-game sensitivity for a same-game comparison number.
  • FOV converter — Convert horizontal and vertical FOV across 4:3, 16:9, 16:10, and 21:9. Includes Overwatch, CS2, and Valorant.
  • Frame-time and Hz calculator — Convert FPS to milliseconds, compare that wait to the monitor refresh, and read a 1% low as a frame time.
  • Mouse polling-rate tester — Measure how often the browser sees pointer events. Average and max Hz.
  • Mouse raw input test — See every mouse report as a dot, spot sensor skips, and check jitter, swipe speed, and acceleration.
  • Mouse test — Check mouse buttons, the scroll wheel, double-click chatter, and a pointermove polling-rate reading in the browser.
  • Keyboard test — Press keys on a US QWERTY board, then time your typing on real sentences or random letter strings.
  • Gamepad test — See every controller button, analog trigger and stick live, check stick drift with a suggested deadzone, and test vibration.
  • GPU compare — Specs side by side with sources.
  • Bottleneck / FPS estimator — CPU- vs GPU-bound from numbers you measured.
  • PSU wattage calculator — Board power presets plus headroom you choose.
  • Frame-time calculator — ms per frame from an FPS target.
  • Graphics cards hub — Comparisons and guides.
  • RTX 5070 vs 5070 Ti — Spec comparison article.

Comments

No comments yet. Be the first to share what you think.