But the real end-user benefit in Haswell is power efficiency, with Intel saying that the chips can slash power consumption by as much as 41 percent in notebooks and ultrabooks. In real terms, power consumption has been cut from 17 watts to 10 watts, offering a massive improvement in performance per watt.
The problem now is that processors no longer sell devices, it's the devices that sell the processors, and as such buyers aren't concerned any more about GHz, cores, and L2 cache sizes. However, a processor that features lower-power consumption is something that translates into a metric that everyone can understand -- longer battery life.
That's not entirely true. The 10W TDP of
some new Haswell chips should certainly allow for a more fitting CPU for an Ultrabook platform, but that has little to do with battery life. Most Ultrabooks with 17W TDP chips actually have
worse battery life than their 35W brethren. That mainly has to do with the limited space for a larger battery. The 35W chips idle at very low voltages anyway so the savings in power are only during full load situations and for most computing that's negligible. A majority of the battery is burned up by the display and not the processor, thus any increase in efficiency for the chip is still only going to result in a very small gain in battery life and only under certain circumstances.
Not that I'm complaining, just clearing things up a bit. The article is obviously written by someone who doesn't know what he's talking about or purposely overstating the improvements and their potential effects. I think Ultrabooks are a great idea, but only if they offer something tangible over the current flock of notebooks. At the moment they really aren't any lighter, any more portable, they're more expensive, don't offer better performance nor do they offer any more time away from a socket. Being thin alone isn't going to sell Ultrabooks. Well, some, but not enough. Thankfully, Haswell seems to be at least addressing the issue of the persistent throttling we see in current Ivy and Sandy chips so it's certainly a start
Hopefully Intel can follow through with either decreasing the prices across the board or addressing some of the other issues that plague the platform. Cutting the TDP down from 17W to 10W isn't going to help the battery any, though.
- editing to not sound like a Debbie Downer.
Intel has pushed for standardizing the design of the Li-Polymer batteries which can be a significant factor in the overall cost of the laptop. Unlike Li-ion batteries they do tend to degrade a little quicker but the turnover period of notebooks means you're usually not going to see it (500 charges or 3 years. For me though, that's about 2 years). Standardizing the batteries should drive costs down.
Intel has also pushed for standardizing the mSATA cards in the laptops as well which, too, should decrease costs.