What is the DualSense polling rate on PC, and is overclocking to 1000 Hz worth it?

A DualSense polls at 250 Hz over USB on a PC and, in gamepadla's measurements, between 542 and 799 Hz over Bluetooth. The widely repeated "125 Hz over Bluetooth" does not match any DualSense test we could find. Overclocking the cable to 1000 Hz is worth doing if you play wired: on gamepadla's Prometheus 82 rig, which physically presses the buttons and moves the sticks, average button latency fell from 8.27 ms to 3.93 ms and stick latency from 7.41 ms to 2.85 ms.

Why two answers circulate

The figure most search results repeat is from pollingratetestc.com: 250 Hz over USB and about 125 Hz over Bluetooth. The USB half is right. The Bluetooth half cites no measurement, and the measured numbers disagree with it. Gamepadla recorded 542.42 Hz over Bluetooth in March 2025 and 798.72 Hz in June 2026, both on Windows.

About 125 Hz is a real Bluetooth figure, just not the DualSense's. It is what the Xbox Series controller, the Flydigi Vader 5 Pro and the 8BitDo Ultimate 2 measured over Bluetooth in the table below. The DualSense is the outlier: its stock cable is the slow link and its Bluetooth link is the fast one. A gamepadla commenter put it as "so funny that wireless is somehow better than wired", and the site's admin replied: "the standard polling rate there is only 250. If you speed it up, it will be faster over the wire."

Six pads, measured

Gamepadla measures latency with its hardware rig and polling rate with a separate software tool. The mean polling delay column is ours: half the report interval, the same rule we use for mouse polling rate.

Sources: DualSense Edge, Xbox Series controller, Flydigi Vader 5 Pro, 8BitDo Ultimate 2 Wireless. The Edge Bluetooth button test ran at 689 Hz and the stick test at 743 Hz.

Is 1000 Hz worth it? The arithmetic undersells it

By the half-interval rule, going from 250 Hz to 1000 Hz removes 1.5 ms of mean polling delay. Gamepadla measured a much larger drop on the same controller, same firmware and same tester:

The measured saving is roughly three times the polling arithmetic. Consistency improved too: button jitter fell from 1.11 ms to 0.36 ms, and the worst button press from 10.32 ms to 4.68 ms. At 1000 Hz a standard DualSense lands within half a millisecond of the Edge and under a millisecond of the Vader 5 Pro on buttons, and half a millisecond behind the Vader on sticks.

Four milliseconds is a large saving by input standards. In the click-to-photon chain it is bigger than the gap between a 144 Hz and a 240 Hz display's frame time.

If you will not install a driver, Bluetooth is the faster stock option on average (5.96 ms against 8.27 ms), with a catch: its jitter was 2.43 ms and its slowest press took 17.11 ms, against 1.11 ms and 10.32 ms on the stock cable.

How the overclock is done, and what it costs

Pollingratetestc tells readers to raise the wired rate in DS4Windows. DS4Windows' own documentation describes its poll rate option as "BT Poll Rate", a Bluetooth setting in the profile editor's Other tab, and says nothing about the USB rate. The documented wired route is SweetLow's hidusbf filter driver, which Battle Beaver distributes with a step-by-step guide. Per that guide:

• It works on DualSense and DualShock 4 only. "This program will not work on Xbox Series X, Xbox One, or Xbox Elite controllers at this time." The rate choices are 1000 to 8000 Hz for a DualSense.
• It needs Memory Integrity off. Windows 11 users must switch off Memory Integrity under Core isolation in Windows Security. Secure Boot no longer has to be disabled, because the current drivers are EV signed by Battle Beaver and attestation signed by Microsoft.
• Check it took. In the tool, the Filter column should read "Yes" and the Rate column should match your choice, and both should survive unplugging the controller. Then confirm the delivered rate with a polling test, as you would for a mouse.

Switching off Memory Integrity lowers a Windows security protection for a few milliseconds of input. That is a fair trade for a competitive wired setup and a poor one for a general-purpose PC. The alternative is a pad that already reaches 1000 Hz without help, such as the Edge or the Vader 5 Pro.

Where this stops applying

Everything here is Windows on PC. hidusbf is a Windows driver, so it does nothing on a PS5. Gamepadla's figures come from individual units, mostly on firmware 0630, and its own disclaimer says other units of a model "might perform slightly better or worse". Its three stock-cable DualSense tests averaged 7.81, 8.27 and 11.27 ms on buttons, so treat any one figure as a point in a range. Polling also says nothing about stick quality: the standard DualSense uses potentiometer sticks, which wear in ways Hall effect sticks avoid, and a worn stick needs a bigger deadzone no polling rate can win back.