Gp100 isn't coming to GeForce guys, they're not going to waste all that die area on fp64, not to mention that the size of SMs are cut in half, so register file size is doubled across the chip. HBM won't be worth the cost.

half agree... the GP100 is for pro right now (considerating that pro don't have money issue) But in the past NVidia have done titan black without any cut and despite prohibitive price for normal consumer they sold this card like holly bread... but with new Intel and AMD computing card on the starting block i guess it is not for now 🙂

I think there is a chance that nvidia will counter Vega with a cut down titan version, if they can't produce enough units for the full version. Or they release it in Feb-March.

The cutdown version willl not be the Titan, but the 1080TI.. they can surely release a full version as Titan . The question still remain if it is GP102 or GP100 and whatt really differentiate GP100 and GP102 .. I will bet, all FP16 hardware disabled (dont want too go across the PX version ) + FP64 units disabled nearly completely ( for save power ) no Nvlink obiously ( limited in support anyway ).. but with a 4096bit bus , 16GB HBM2 vs 12GB on the Ti Version ( one stack disabled as the cheaper version of the PCIe tesla )

half agree... the GP100 is for pro right now (considerating that pro don't have money issue) But in the past NVidia have done titan black without any cut and despite

Not 100% right, GP100 today is only availlable to some Supercomputer center who had pass contract for them allready years ago.. who will take anyway 6 months to 1years for finish their upgrade ( the most known so far is the Switzerland national supercomputing center who will replace his 4500 GPU with 4500 tesla GP100 ( there should be some other).. Its not really publicly available, but only for serve the contract of some specific instances in computing. ( because they need to start the upgrade now for be in time ready for their own contract. ) Even if you have money, you cant forget to get one. in Q4 2016 ( mostly december ) you could buy a Tesla PCIe version. Upgrade, test the hardware, software, electric supply, cooling, emergency systems etc for supercomputers take time, even when all is going well, it can take 1 year.. Samples who are delivered right now, are all parts of specific contracts.. Nvidia can even send them non fully functional chips and replace them with good one at the moment they can realize their test ( I dont say this is the case, but it have happend allready by the past in this type of industry ) Its completely different of an commercial " availability".