Analysts seem to be looking at Intel's offerings as something that might save the PC market. I think this is misplaced.
PCs (including ultrabooks) and tablets have different use cases, and are not all that interchangeable.
A friend got an iPad, and described it as a "media consumption device." I think he was spot on. The tablet is a largely half duplex device. The assumption is that you will use it to consume media, and the interface is optimized for choosing the media you wish to consume. You don't need a keyboard or a mouse, because you aren't generally *creating* content. You're just tapping the screen to pick what you want to read/view/watch/hear next.
If a tablet fits your use cases and meets your needs, you won't buy an ultrabook, even if it is the same price, because you don't need to do the other things an ultrabook does. (And while you *can* use an ultrabook for the same purposes as a tablet, it will be far less convenient.)
Intel accepting lower margins to allow ultrabooks to be cheaper won't magically make them sell better. The question isn't what an ultrabook costs, it's how many people need what an ultrabook does. The likely answer is "Not enough to provide the sales ultrabook vendors might like."
We may well be in the position where everybody who *can* use a PC (in whatever form factor) largely *has* one, and the market is replacements and upgrades, not new sales. That's still a vary large market, but is doesn't provide the growth the analysts look for.
The tablet market is not yet saturated, and does.
AMD is not going anywhere just yet. Xbox sales alone will make their APU effort profitable. Their APUs are a nice fit for Xbox-type devices. They don't have Haswell's performance, but are cheaper and use less power. They aren't as low power as ARM, but have far better performance. AMD is wisely trying to fit into the gap between ARM and Intel's offerings.
From supply chain.
Surface 2 will be announced in June and ship 2H. MS projects highest volumn Sufrace will be the one with snapdragon 800.
Describe as good enough computing for the masses at a very agressive price.
“Haswell is only marginally better than [the existing] Ivy Bridge [CPUs] for most compute tasks,” Brookwood said.
I think that was widely expected...Haswell is focusing on "battery life".
The net result is “there’s still a big gap between the performance of the fastest Intel Atoms and the slowest Haswells,” said Brookwood. “Likewise, battery life is way better with Atom than with the best Haswell,” he said.
Thus “there’s nothing in the ARM camp that can compete with Haswell on performance, and there’s nothing in the Haswell camp that can compete with ARM for battery life,” he added.
Maybe Atom that can beat ARM with respect to performance and battery life??!! - Haswell addresses a completely different market compared to Atom.
just like BMW has a 3 series and 7 series -
Haswell is part of the Core x86 architecture the follow on to Ivy Bridge.
Atom is a separate x86 lineage aimed at lower power and performance targets with its own road map targeted at mobile and microservers as you note.
Surface Pro running on Haswell is a 4 hours device. At 999 euros , itvs far too low. 14 nm won t bring much in term of power saving. Intel is ok on high perf devices...arm better on low power. AMD is the ultimate loser
Blog Doing Math in FPGAs Tom Burke 24 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 ...