I believe most of the NVMe SSD heatsinks' contact surface is just one flat piece, but not all the chips on the NVMe SSDs have the same height.
From my recent observation, there are air gaps between some chips and the thermal pad/heatsink after putting them together, the thermal pad that came with the heatsink is not enough to fill the gaps.
Here's a picture. The SSD in this picture is the same model I've observed in person. The lowest chip is the third one from the left (probably DRAM,) the one with smallest top surface. I've read somewhere the 4th one from the left is the controller, 1st and 2nd ones are NAND memory chips:
Those which are not in contact with the heatsink will have worse heat dissipation, but does it matter?
My concerns are:
A heatsink that's not in contact with the components underneath is also restricting the air flow around those components, so it's probably worse for those components than without the heatsink and there are more components without contact to the heatsink than there are with contact. In such case, is the heatsink actually doing more harm than good?
The vibes I'm getting from infos on the interweb is that the temperature frequently hovering around the throttle threshold is bad for reliability and longevity of the SSD.
Is it true that as long as the temperature stays under the threshold, through throttle or not, reliability and longevity won't be considerably worse?