This summer, Intel is expected to introduce a new version of its x86 CPU architecture, one that has spanned more than 30 years and multiple attempts to put it out to pasture. This new generation, codenamed “Haswell,” will introduce new functions and play a big role in certain platforms, particularly for a new class of hybrid ultrabooks.
Many CIOs and IT leaders don’t have time to closely follow developments in the CPU arena. However, the changes in this industry ultimately drive advances in the capabilities of the computing devices you’re choosing for your workforce. The reason we’re singling out Intel’s Haswell is because of its potential impact on a new generation of ultrabooks.
Is a new CPU that’s still at least half a year from hitting the market reason enough to hold off on your enterprise device purchases? Probably not, but it’s worth watching where this one goes.
First, it’s worth learning a bit about the Intel engineering policy known as “Tick/Tock.” “Tick” is when they take the existing processor architecture and reduce its manufacturing design. “Tock” is a whole new architecture built on the current processor technology. The company alternates between the two policies every year.
The current product on the market is the Ivy Bridge line, which is a “Tick” for Intel because it took the Sandy Bridge architecture, originally introduced on a 32nm design, and shrunk it to 22nm. Haswell will also be 22nm, but it’s a whole new processor design.
Intel separates architecture from shrink because doing both at the same time is too complicated and risky.
What’s the big deal about Haswell?
So what’s the big deal about Haswell? Well, it will be aimed squarely at the ultrabook market. Ultrabooks have improved slowly over the first generation, with battery life creeping up to be nearly comparable to regular notebooks.
Ivy Bridge notebook and ultrabook processors primarily draw 35 watts, although some are as low as 15 watts. They also have a much slower clock speed, about 1.6Ghz, to show for the reduced wattage.
Haswell will draw less power and get more out of its clock. The conventional wisdom among hardcore CPU tech sites is that Haswell will demonstrate a 10 percent to 15 percent improvement in instructions per clock cycle over its predecessor.
Haswell processors will start at 35 watts but drop as low as 10 watts for ultrabooks and tablets. CPU speed is about the same, but that wasn’t Intel’s focus. With Haswell, Intel is aiming for improved power management and graphics performance.
For example, the voltage regulator normally found on the motherboard is being moved into Haswell, so voltage management is done inside the CPU, improving efficiency and putting one less component on the motherboard.
Haswell introduces new Advanced Vector Extensions (AVX2) that will double the graphics performance over the Ivy Bridge generation. The new AVX2 extensions will reportedly give ultra-high definition resolution, 3840×2048, double that of current high definition video. Not that your notebook needs it. However, a doubling of graphics performance will help accelerate Windows 8, which relies on the GPU more than any previous operating systems.
Intel has not said officially when it will ship Haswell. The current rumor is the processors will make their grand debut at the Computex show in Taiwan in June 2013. That would be the most logical time to do it, as Computex is when all eyes are on the PC industry.
It will be worth watching folks at Tom’s Hardware, AnandTech, and other benchmarking and testing sites as they tear into Haswell later this year to examine the promised improvements like battery life and graphics performance.
About the Author
Andy Patrizio is a writer for EnterpriseEfficiency.com, a UBM Tech community.