This is an HDMI to DisplayPort adapter. It is for connecting an HDMI output (graphics card, laptop, etc.) to the DisplayPort input on a monitor. It will not connect a DisplayPort output to an HDMI display.
======================
VIDEO FORMAT SUPPORT
======================
This adapter seems to support up to ≈300 MHz TMDS clock which is slightly below the maximum allowed by HDMI 1.4. A 300 MHz TMDS clock gives it around 9.0 Gbit/s of bandwidth (7.2 Gbit/s data rate). This data rate is enough to support:
• 1920 × 1080 @ 120 Hz
• 2560 × 1440 @ 60 Hz
• 3840 × 2160 @ 30 Hz
assuming 8 bpc (24 bit/px) RGB or YCbCr 4:4:4 color.
I have tested the adapter at 1080p 120 Hz and at 1440p 60 Hz and can confirm that it does work for both of those formats.
The adapter was tested with two different 1080p 144 Hz monitors (a ViewSonic XG2401 and a AOC G2460PF) using both an AMD graphics card (a Radeon RX 480) as well as an NVIDIA card (a GeForce GTX 1080), and it worked up to 120 Hz on all four combinations. 144 Hz did not work; this was somewhat expected since the adapter only claims to support up to 120 Hz anyway, but I just wanted to check.
I verified 120 Hz was working properly without frame skips for all four combinations using a high speed camera.
The adapter was also tested at 2560 × 1440 @ 60 Hz using a Dell UP2516D. It functioned properly for that too, however it should be noted the adapter does not appear to support 10 bpc (30 bit/px) color; only the 8 bpc option was available.
==================
OTHER CAPABILITIES
==================
The adapter does support inline audio.
The adapter does not support AMD FreeSync.
The adapter claims to support Dual-Mode DisplayPort (which means it has support for DisplayPort to DVI/HDMI passive adapters), so I decided to test that out just to see if it was true, and in fact it does work with both DP to DVI and DP to HDMI passive adapters. It apparently supports the updated Dual-Mode spec, since I tested it with a Type 2 DP to HDMI passive adapter and it did work at the full 1080p 120 Hz even through both adapters. I don't know why you would want to convert your HDMI to DP and then back to HDMI again rather than just using an HDMI cable end to end, but I guess if that's what you have laying around...
=======
POWER
=======
The Amazon picture shows "USB 3.0 for power"; since USB 3.0 has a higher maximum power delivery (900 mA at 5 V compared to USB 2.0's 500 mA) I assumed the adapter would require the extra power. However, believe it or not the adapter only appears to use around 40–50 mA during operation, at least according to my USB power meter.
=========
CAUTIONS
=========
The adapter can be finicky at times, and when hot-plugging the adapter between different graphics cards and monitors it would sometimes need to be unplugged and reconnected several times in order for a connection to be established. Please note that it may take several seconds for the image to appear on the monitor.
========
LATENCY
========
For curiosity I also tested to see if the adapter adds any latency; the short answer is no, the adapter does not add any amount latency worth worrying about.
I measured this using a custom device I built using a disassembled USB mouse and an oscilloscope; I measure the time delay between the left mouse switch being depressed and a light appearing on the screen (in this case, the muzzle flash from my gun in CS:GO) as detected by a light-to-voltage converter.
Since this measures the entire latency chain from the mouse click to the system, to the game engine, to the GPU driver and hardware, to the display unit, to the screen, the actual latency number is not important here, but rather the difference between the latency when using a straight DisplayPort cable and when adding the adapter inbetween. Any latency added by the adapter should be detectable as an increase compared to not using the adapter.
Since the screen does not update continuously, but only every 8.333 ms (since it is operating at 120 Hz) in fixed intervals, a random amount of latency between 0 and 8.333 ms will be added after every mouse input, depending on how the timing of the click lines up with the when the screen is refreshing. So to mitigate this somewhat I have picked the lowest latency measurement from a sample of 50 measurements for each configuration. I measured 43.8 ms minimum (with 55.8 ms average) with a straight DisplayPort connection, and 45.2 ms minimum (58.7 ms average) with the adapter. Considering the 8.333 ms margin of error, my conclusion is that the latency of the adapter is negligible.