A. Eassa pulverizing INtel on the Ihub board, but won't show his face here
What a snake.
Tuesday, 10/29/13 05:43:16 PM
If Intel's Cherry Trail can't decisively beat a 28nm planar part targeted at a similar power envelope in a smaller die area, then Intel's process lead doesn't mean #$%$ in the real world. I'm not ready to accept that given how that lead has manifested itself nicely in the PC/server space once the designs are competent.
Bottom line, though, is that while Intel's designs for this space at both an individual IP block + SoC level are getting much better and closing the gap even normalized for process, they've still got two major problem spots (GPU and baseband).
Bay Trail should have been built with a "Rogue" GPU if it wanted to have decisive leadership. The reason they didn't do that, of course, was that building drivers for both Windows and Android would have been quite the task especially given the work that already goes into GenX drivers for Windows.
That's why Merrifield is built with a Rogue, and that's why its (most likely quad core Silvermont) successor is also built with Rogue. It's also why these two processors will be aimed at both Android tablets and phones, not just phones.
I think Merrifield will gain some traction as an excellent alternative to Snapdragon 400 and Tegra 4i in both mid-tier "large" phones but, more importantly, high end "small phones".
Intel's target seems very much to be iPhone type devices and I doubt very much that Intel had Phablets on the mind when designing Merrifield.
I think that's a mis-characterization of Essa's position. Essa had taken the view that Intel's growing success in tablets and mobile would come at ARMH's expense and hence had advocated short ARMH positions. What Essa didn't understand is that in a rapidly growing & expanding market Intel and ARMH can grow simultaneously - at least for a while. Sure some of Intel's success will come at the expense of ARMH's growth potential and OEM profitability but ARMH will still appear to be growing. I believe Essa is simply acknowledging that the second piece of his bullish Intel growth view isn't necessarily true.