How to get an accurate benchmark result

Small things change a benchmark result: the charger, heat, other apps, battery saver, even switching tabs. A checklist for a fair run on a phone, tablet, laptop or desktop, and how to compare two devices honestly.

A benchmark measures a device as it is at that moment, and “as it is” includes far more than the chip: how warm it is, whether it is charging, what else is running. None of that is cheating, and none of it is a fault in the test — but it is the difference between a result you can compare and one you cannot. Here is how to give a device a fair run.

Before you start

• Unplug the charger. Charging changes how fast a battery device is allowed to run. Many laptops run faster on the charger and hold back on battery to save power, while a phone can run slower when charging, because charging warms it up. Testing on battery is the fairest common ground, and the site shows a note when it can see a charger plugged in.
• Close other apps and tabs. Anything working in the background competes for the same processor, and on a phone, apps left open can keep working.
• Let a hot device cool down. If it has just been gaming, charging or sitting in the sun, give it a few minutes. A warm device slows itself down to protect the chip, and the test will measure that slowdown rather than the device at its best.
• Turn off battery saver or Low Power Mode. These modes cap the processor on purpose. A test run in battery saver measures battery saver.
• Check the power mode on a computer. On Windows, a “best power efficiency” power mode slows the processor too. For a fair comparison, use the same mode every time you test.
• Pick the right device type. The tier compares a phone with phones and a laptop with laptops. If the button says “Test my tablet” on a phone, tap Phone first.

While it runs

• Do not touch the screen. Taps and scrolling make the device do extra work in the middle of the measurement.
• Stay on the tab. Browsers slow down pages that are not in front — a background tab is given far fewer frames — so switching away during the drawing test makes the device look much slower than it is.
• Take a phone out of a thick case if it warms up quickly. Cases hold heat in, and heat is what makes a device slow itself down.

Run it twice

One run is one sample. The results screen shows how far the three passes inside a run differed, which says how steady that run was; a second run a minute later says how repeatable the device is. If the two scores agree, trust them. If they are far apart, something changed in between — usually heat, a background task or the charger — and the guide on repeatability explains what to look at.

Bear in mind that the test causes some of the heat it measures. About a minute of full load warms any device, so a second run straight after the first can come out a little lower. If the drop is large, the device struggles to hold its speed under load, and that is worth knowing in itself.

Comparing two devices

Comparisons are where conditions matter most. Test both devices the same way: both on battery, both cool, both with nothing else running. A comparison between a cool phone and a warm one measures the temperature, not the phones.

The score uses one scale for every device, so a phone can be compared with a laptop directly. The easiest way is the “Dare a friend to beat it” button under your result: it makes a link that carries your measurements, and when the other device runs the test from that link, both results appear side by side under “Compare with another device”.

Testing a demo unit in a shop

A shop’s demo unit is a worst case. It has usually been switched on all day, it is often plugged in and warm, and other people may have left apps running. Ask before you test it, close what you can, and treat a poor result as a reason to ask for a second run rather than as a verdict. The shop guide goes through it step by step.

What even a fair run cannot tell you

A careful run is still a measurement of the processor and of drawing speed in a browser. It does not measure storage speed, the camera, battery health or the graphics chip, and it cannot tell you whether a device is genuine. A device that scores well below its type might be warm, busy, on battery saver or carrying an aged battery — and the right response to a surprising result is a second, careful run.

Related

• Why is my new phone slower than my old one?My new phone feels slower than my old phone. Why?
• What makes a device throttle?What does it mean when a phone or laptop throttles?
• Is my phone throttling?How do I know if my phone is throttling?
• What is 1% low FPS?What does 1% low FPS mean?
• Why does the same device score differently?Why did my score change between runs?
• How to test a phone in a shopCan I test how fast a phone is before buying it?