Test GPU workload on battery in a virtual desktop

A graphics test in a virtual desktop does not always put a graphics workload on your laptop. If the browser is running inside a remote desktop, the remote computer may do the rendering, while your laptop mainly displays the streamed image. To test your laptop's GPU on battery, run the test in a browser on the laptop itself.

The Graphics Test measures browser graphics performance through shading and fill rate at 1080p, then gives a score out of 100. Compare runs made in the same browser and graphics environment, first on AC power and then on battery. The score is not a direct measure of battery drain or of a GPU's full performance in every app.

Choose which graphics system to test

Test your laptop's GPU while it is running on battery

Open the Graphics Test in a browser running directly on the laptop, outside the remote desktop window. That tests the graphics path available to the local browser, which may use integrated or dedicated graphics depending on the laptop and its settings. Disconnect power only after the laptop is set up and the test page is ready.

Test the remote desktop's graphics path

Open the test inside the remote desktop if you want to measure what the remote session's browser can render. The result can depend on the remote computer, its virtual graphics setup, the remote desktop software, and the connection. It does not tell you how much work the laptop's GPU is doing to display the session.

Keep those two tests separate. A local browser test and a browser running in a remote session may show different graphics devices and use different hardware, even if both windows appear on the same laptop screen.

Compare GPU performance on battery and AC power

1. Open the Graphics Test in the browser context you want to measure. Use a local browser for the laptop GPU, or a browser inside the remote desktop for the remote graphics path.
2. Keep the same browser, display arrangement, screen resolution, and remote session state for each run. Close extra tabs or apps that are doing heavy graphics work, and avoid changing display settings between tests.
3. With the laptop plugged in, run the test and note the shading result, fill-rate result, and overall score. The test reports a score out of 100, but that score is most useful for comparing the same device under similar conditions.
4. Let the laptop return to a similar temperature if it warmed up noticeably. Then unplug it, wait briefly for the power state to settle, and run the test again in the same browser and session.
5. Record the battery level and power mode for both runs. On Windows 11, check Settings › System › Power & battery. On macOS, check System Settings › Battery, including whether Low Power Mode is enabled.
6. For a separate view of battery behavior, open the Battery Meter on the laptop itself, if its browser reports a battery. Note its drain rate before and during a local graphics test, but do not treat that reading as a direct GPU power measurement.

One run can be affected by background activity, heat, or a change in the remote session. If a difference matters, repeat the comparison under the same conditions rather than relying on one score.

Understand what a change in score means

The score falls on battery

The local graphics path is performing differently when unplugged. Some computers reduce performance on battery to conserve power or manage heat. Check the power mode and battery settings, then compare again with the same settings and a similar starting temperature. A lower score alone does not prove that the GPU is faulty.

The score is similar on battery and AC power

The browser test did not show a large change under those conditions. This does not guarantee identical performance in a game or editing app, which may use different graphics features and workloads. It also does not show whether the battery is draining faster during that other task.

The score is low or varies between runs

Browser graphics performance can be affected by other active apps, heat, power settings, browser graphics support, and virtual desktop configuration. Close graphics-heavy tasks, use the same session and screen setup, and run a fresh comparison. A virtual desktop may also expose a virtual graphics device rather than the physical GPU model.

What the test cannot tell you

A browser graphics score is not a direct readout of GPU utilization, temperature, wattage, or battery health. It measures how the browser completes the test's graphics work. It cannot confirm that a remote desktop host is using a particular physical GPU, and it cannot separate the host's rendering time from network delay in the streamed picture.

If you are checking laptop battery use, run the graphics workload locally and compare the Battery Meter's reported drain rate over a similar interval with the test stopped and running. The browser reports an estimate, and other activity on the computer can affect it. If the browser does not report a battery, the Battery Meter may not be available.

Keep the laptop on a hard, clear surface during a graphics test. If it becomes unusually hot or the fan runs loudly, stop the test and let it cool before repeating it.