For some time there are lots of rumors circulating around the web that next-Generation of AMD Graphics cards will take advantage of the modern memory HBM (High Bandwidth Memory). One of the AMD’s employees somewhat confirmed this information. It will totally change the game as HBM 3D Stacked memory is up to 9x faster than GDDR5.
AMD Radeon R9 390X to feature HBM Memory
Linglan Zhang (PMTS at AMD) posted to his profile on LinkedIn projects in which he participated as a Project Manager – among them the most amazing line of our interest is
Developed the world’s first 300W 2.5D discrete GPU SOC using stacked die High Bandwidth Memory and silicon interposer – Linglan Zhang
New memory has a much higher bandwidth and lower power consumption with respect to the currently used GDDR5. If you refer to 300W TDP, we suspect that it was a top model – eg. The AMD Radeon R9 390X. Is weaker models will also be equipped with a memory HBM? That, unfortunately we don’t know about this case.
We know, however, that the engineers are working on it too – ASIC physical design engineer, team leader is impossible to find information that worked on the design of the model. By the way, it is worth noting postscript, he was the biggest “king of the hill” to the entire product line. Well, one thing is certain – that the next generation Radeon will make a huge leap in terms of development in order to beat the others in the race.
In case you are wondering what’s the difference between 2.5D and 3D stacking, here’s a good quote from ChipDesign:
In 2.5D stacking, two dies are flipped over and placed on top of an interposer. All of the wiring is on the interposer, making the approach less costly than 3D stacking but requiring more area. Heat dissipation is not much of a concern, since cooling mechanisms can be placed on top of the two dies. This approach is also lower cost and more flexible than 3D stacking because incorrect connections can be reworked.