I tested the load in the Crystal Disk program, a Samsung 980 pro hard drive. At the beginning of the first run, the temperatures of the memory chips were a pleasant surprise, by the end they were almost equal to the results without a heatsink. On the second run, they were only 3 degrees lower than without the heatsink. Under load in the test without a heatsink 67, with a heatsink 64. The memory controller without a heatsink was 90, now it's 75, it's already better here. Definitely better to look for another solution for normal cooling :) Tested in a PC, vented be quiet 500dx case