The problem arises from the fact that the 1st Generation HBM memory is limited by design to just 4GB worth of capacity. While 2nd Generation memory can support up to 8GB of capacity, its production time line is too far for AMD to be able to use it in its Radeon 300 series designs. And AMD is well aware of the fact that 4GB of memory is no longer enough for a flagship product, therefore the only solution is dual link interposing. With a Dual Link Interposing design, SK Hynix will be able to stack 4x (Dual 1GB HBM modules) via an Interposer (2.5D stacking). The design will have very high performance at low clock speeds and will also be very power efficient as compared to GDDR5.Source: Bitsandchips via WCCF Tech
2x 4-HI HBM1 (which should technically be called 8-Hi-Hi according to nomenclature rules) features a 1024-bit interface, two prefetch operations per IO (dual command) and can push 128GB per second per pin. The tRC is 48nm, tCCD is 2ns (1tCK), and VDD is 1.2V. The 4-Hi HBM2 (generation 2) features a 1024 bit interface, two prefetch operations per IO (dual command), 64 Byte access granularity (=I/O x prefetch) and can push 256 GB per second per pin. The tRC is 48nm, tCCD is 2ns (1tCK), and VDD is 1.2V.
AMD Radeon R9 390X to get 8GB HBM after all?
Posted on Monday, April 27 2015 @ 13:13 CEST by Thomas De Maesschalck