As proven on the above slide, AMD shall be bringing a re-architected Compute Unit and pipeline enhancements to RDNA 3 graphics, plus a “next-gen” Infinity Cache, along with the headlining transfer to a chiplet method fairly than a big monolithic die. AMD claims that these enhancements, together with the transfer to a 5nm course of, have the potential to spice up effectivity by better than 50% over RDNA 2.
As to the general roadmap, AMD confirmed a slide indicating that after the upcoming RDNA 3 (“Navi 3x”) merchandise at 5nm, the subsequent technology RDNA 4 (“Navi 4x”) shall be on an “superior node”, although it was not revealed whether or not this shall be an enhanced 5nm or a smaller course of tech.