AMD: Carrizo APU targets mobile platform only, Kaveri to power Desktop side till 2016
So something pretty interesting showed up at the Italian publication Bitsandchips.it. The pub has sourced their article from a study published in PC Watch and the contents are very interesting. However, since Google Translate is the only tool at my disposal and AMD PR will be off for the holidays, I cannot confirm whether this is authentic or not although this is in all probability legit since its based of ISSCC 2015. Basically, the piece states once again that Carrizo APU has been designed specifically for mobility platforms and wont be appearing on the desktop APU socket like Kaveri till 2016.
Carrizo APU wont come to desktops till 2016 - all but confirmed
Some technical details about Carrizo are present as well, particularly the fact that it employs a Resonant Clock Mesh, something which apparently gives you the same performance as a regular clock mesh but with decreased power consumption. Pretty neat, and a good feature to have a in a mobile part. AMD has also apparently stated (and I am quoting the pub on this, not the company) that Carrizo is more like Beema than Kaveri at the moment, with the higher powered implementation coming later (2016). Carrizo measures 244.62 mm2 and packs more than 3.1 billion transistors. Its new Excavator core is 23% smaller and uses 40% less power than AMD’s previous x86 core.
On the GPU side, the APU packs Third Generation GCN (Volcanic Islands) core which is pretty damn great. The integrated graphics would still be based on the 28nm process and its hard to say whether the Compute Unit count will increase in the Carrizo APU generation of products. AMD will also add AVX2, BMI2, MOVBE and RDRAND support to the instruction set which brings the extension feature set close to Intel’s Haswell. Since the targeted platforms for these APUs are notebooks, all-in-ones and convertibles, they will be shipped in BGA, FP4, package and will ship in variants ranging in TDPs of 12/15/35W.
Also, Carrizo is built on the 28SHP process from Global Foundries and is optimized to increase transistor density while maintaining low power consumption. This decision will allow AMD to use both Carrizo (Excavator) and Carrizo-L (Puma +) using the same socket. It is also a major hint that AMD does not consider the 20nm node mature enough for double patterning and therefore not valid for any high performance ASIC such as a GPU. Finally the pub states that the APU will not be arriving to the FM2+ socket till 2016 and Kaveri will stay with us throughout 2015.