USB Monitors? DisplayLink's Technology, Examined

Performance Scaling Depends On Your Workload

Because the host system does all the compression and encoding, all DisplayLink devices share the same CPU workload. Thus, the only two variables affecting CPU utilization are resolution and on-screen activity.

Simply moving your mouse around causes CPU utilization to hop up to 15% on a capable mobile Core i5 processor. And that a light workload, as only the screen data around the cursor is getting compressed and encoded. 

The absolute worst-case scenario is a contrived synthetic measurement we put together, progressing through Windows' Picture Viewer at the rate of two images per second (though the results of that aren't far off from a more real-world workload of H.264-based video playback). In both cases, the system processes entire frames, causing our Core i5 to hit 50% usage at 1920x1080.

Moving a window on the DisplayLink-enabled screen causes CPU usage to jump to 20-30% for short bursts, depending on the resolution you're using. Once the image stops changing, CPU utilization is allowed to fall back to 2%. That's good news, since we're seeing some scary-high utilization numbers.

Naturally, if you want to avoid those uglier results, avoid video playback on a USB-attached screen. Swap over to the notebook's integrated display and CPU use drops to near-nothing thanks to Nvidia's Quadro 3100M GPU with hardware-accelerated H.264 decoding.

Obviously, gaming on a screen attached via USB is a terrible idea. We wouldn't even recommend watching a movie on it. But as a mechanism for improving productivity via a second or third display, DisplayLink's solution is more compelling.