Looks like it’s that time of the year again. As we near the launch of FinFET based offerings from both IHVs – the rumor mill is in full tilt. A recent report form DigiTimes is claiming that AMD will be launching its Polaris GPUs in June. The Polaris architecture will be manufactured on Globalfoundries or Samsung's 14nm FinFET processes and is equipped with its fourth-generation GCN architecture, and support for HDMI 2.0a and DisplayPort 1.3.
AMD launching Polaris series of GPUs in June - targeting the 'minimum VR spec' price level of $349
This leak ties in well because we saw atleast one Polaris board pass RRA certification recently - indicating that everything about the GPU has been finalized and is ready for full scale production and shipping. We also saw many Arctic Islands codenamed samples surfacing on the Zauba database including Baffin XT and Weston PRO. . Baffin XT is a name that we are very familiar with by now. This is clearly a flagship GPU of sorts yet it is shown to be listed with only a 4GB memory channel. Assuming the card is only using 4GB of memory, this can be either HBM or GDDR5X (GDDR5 would have been clearly mentioned as such). In either case, we are looking at almost twice the bandwidth of the previous generation product (remember, Baffin =/= Fiji Successor). Its highly likely that one of these chips are infact the Polaris GPU
AMD’s Polaris architecture is built on the 14nm FinFET node and will offer a huge leap in performance per watt over the last generation. The raw leap gained just by jumping the node (from 28nm) is roughly 2 times, although AMD is throwing around the word “2.5 times” a lot, which shouldn’t be that hard to achieve with IPC gains. Polaris architecture will initially be divided into Polaris 10 and Polaris 11 – one of which will be big Polaris and one, small. AMD has also previously revealed that the Polaris GPUs will target the ‘minimum VR spec’ at affordable price levels below the $349.
We have also previously heard reports that atleast one of the Polaris GPUs will be around 232mm². According to the information we have about the 14nm LPP process, and based on transistor density increase, a 232mm² GPU would be roughly equivalent to a 464mm² 28nm processor – at the same TDP levels. Since we already know that AMD is going to be focusing not just on performance but power efficiency as well – this number could be be much higher. We can however safely say that this die is more than capable of meeting the ‘minimum VR spec that AMD promises.
The shift to 14nm becomes highly significant in this aspect. Not only does the die shrink allow for more transistors to be placed on the same surface area, it increases economies of scale – lowering cost. But the maturity of the process (14nm LPP has a 20nm backbone) would allow good yields and more volume to be shipped. AMD hopes to exercise this advantage by lowering the cost of “minimum spec VR” to a point below the $349 mark – making VR available to everyone and increasing the Total Available Market for VR.
If AMD is able to succesfully get out Polaris GPUs in June, it will allow it to compliment sales of the HTC Vive and Oculus Rift headset since the graphics card will allow more people with less budget to be able to successfully engage VR content.