HBM is the expensive, it requires and organic substrate (or silicon? I forget) which is expensive, the increased thermal density is more difficult to cool which is expensive.
Even with how difficult, expensive, and hot GDDR6X is that was still preferable for Nvidia's consumer cards.
From what I remember, it can be power intensive which required more power to the motherboard to supply additional power to HBM. HBM2 will consume more than 100w due to increased I/O. It made sense why they are using it in GPU since it can handle additional power from it with 4/6/8 PIN Connectors from PSU directly.
If it is being used as a RAM on the motherboard, I imagine they would need to redesign the motherboard layout to accommodate the power and bandwidth for RAM slot to support HBM. Or a possibly HBM stick get it own pin connectors for PSU to avoid motherboard redesign.
IIRC HBM is actually more power efficient but the problem is power density since HBM is stacked. Could also be trace length.
>Therefore, HBM provides at least an order of magnitude higher bandwidth than DDRx DRAM [15] at a lower power consumption (nearly 7pJ/bit as opposed to 25pJ/bit for a DDRx DRAM) with a smaller form factor [65].
I figured they are making improvement in HBM but wasn't expecting an massive improvement with power saving. That is great to know.
I forgot that HBM is stacked. I wonder if it a form factor issue? HBM is clearly "thick boy" than conventional DRAM. I imagine they need to make some room for the HBM modules including double-sided stick.
