From an ARM-paid reporter such as Rick, We really cant expect more than this. Rick, how much did ARM pay you to write this trashy of an article. Nothing is good at Intel and everything is great at ARM. Right?
Intel architects claim the tablet/notebook and server markets have similar max perf per joule requirements so the core design is substantially the same.
What's really different is, they say, the "uncore" block that has different versions for I/O and memory for servers and clients.
Only intel could also market 2 chips as a
"New! 1 chip BGA solution".
When in fact 2 chips are needed. Even the PCH chip is on an advanced node 32nm. In total intel is using more than 200mm^2 of advanced silicon area for Haswell. My system contact says pricing for the ULV parts is absurd at ($300-400) due to very few parts bin to the less than 15W spec.
Intel seems to be taking a shot-gun approach to chip design and missing the sweet spots in the performance / power-consumption spectrum. After getting hammered for the last 5 years with iPhone, iPad and ARM are they now perhaps punch - drunk ?
squeezed between smart phones and laptops, tablets appear the most threatened. None appear to be making money except iPads, and even the latter has fallen victim to the Mini. Between my iPhone5, iPad3 and quad core i7-3x Dell Latitude with SSD, my iPad3 is the one gathering dust. Too big compared to smart phone (and no cellular unless pay extra), too slow/limited compared to a good laptop.
“I believe that the closer the PC platform gets to the tablet’s convenience and battery life, more people will choose a new PC over a new 10-inch tablet,”
This is so true..I would have chosen a surface pro over ipad, if only the price was similar to that of an iPad. You can't package a full pc in tablet formfactor, and price it 999 and compete with 499 ipad. There should be a price parity. Unfortunately that will mean... lower margins for Intel
The schocking data is the load power (when CPU is running) is higher for Haswell than ivy bridge or sandy bridge.
11.8% increase over ivy bridge (113W versus 101W). I take this as Intels finfet just like the foundry has high gate Capactitance (back to levels not seen since 130 or 90 nodes) and is causing the high power.
Until intel gets the high gate capacitance fixed, my guess it will not be a meaningful player in smartphone market
Blog Doing Math in FPGAs Tom Burke 15 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...