Does the use of infinity cache with slower cheaper memory is really AMD intentionally nerfing RDNA 2 GPUs so they would not be attractive to miners (that would seem an odd business decision or maybe not could help your more pro product line in similar workloads) or simply the consequence of design decision that had nothing to do with it ? If memory slower bandwidth is really the reason they are not attractive I could imagine it is just cheaper than higher bandwidth solution, or it is something else ?Yep. Supply far outstripping demand. Just remember that AMD nerfed it's RDNA 2 GPUs so those would not be attractive to miners.
Last edited: