AMD might be sticking to the good old-fashioned 18 Gbps GDDR6 memory solution for its next-gen RDNA 4 "Radeon RX 8000" GPUs, alleges @Kepler_L2.
While memory manufacturers are talking about their newest solutions for consumer graphics cards such as Samsung's GDDR6W or the truly next-gen GDDR7 standard which will ship with 28 Gbps die speeds in the start, it looks like AMD is planning to stick with the existing GDDR6 memory solution to power its upcoming Radeon RX 8000 graphics cards based on the RDNA 4 GPU architecture. The move may sound a bit odd at first but it might make sense which we will get to in a bit.
The AMD RDNA 4 GPU family which is expected to power the Radeon RX 8000 line of discrete cards has seen various developments. The lineup was initially expected to feature high-end GPU SKUs till they got canceled and replaced with two SKUs in the mid-tier and low-tier segments. We know the two SKUs as the Navi 48 and Navi 44. The Navi 48 GPU is reportedly the higher-end SKU of the two while the Navi 44 chip is an entry-level SKU. These GPUs are part of the GFX12 family and they have already been listed in patches within ROCm as GFX1200 and GFX1201.
Now based on the latest rumor from @Kepler_L2, it looks like the AMD Radeon RX 8000 "RDNA 4" GPUs will be utilizing GDDR6 memory and no special or faster variant. They will stick with the just 18.00 Gbps speeds which are slightly lower than the current fastest models such as the 7900 XTX or the 7800 XT. The Radeon RX 7900 GRE, RX 7700 XT, RX 7600 XT & the RX 7600 feature 18 Gbps dies. The current lineup features the following: