Sandy Bridge: Intel's Next-Generation Microarchitecture Revealed

At this week's Intel Developer Forum in San Francisco, Intel wasted no time revealing the most anticipated technology of the three-day conference: its second-generation Core microarchitecture, code-named Sandy Bridge. It's due for both desktop and laptop release at the first part of next year, but has been generating buzz for months as tantalizing little details have slowly leaked out. When it came to Sandy Bridge, we wanted to know two things: What exactly was it, and was it worth all the hype? The first question has been definitively answered—and we think we've got a good start on the second.

At its most basic, Sandy Bridge is the latest "tock" in Intel's tick-tock development strategy: the latest iteration of the Westmere-based, 32nm die-shrink of the Nehalem microarchitecture that Intel released at the beginning of the year. They won't fit the LGA1156 socket those chips use, however—they'll need a new one, LGA1155, which will be released on motherboards using the new Series 6 chipset at the same time Sandy Bridge CPUs are released.

But beyond all that, Sandy Bridge is really about unity. Yes, it finally brings the memory controller, PCI Express (PCIe) controller, and video functions within the processor die, which by itself is an enormous change that has major performance repercussions. But it also strikes up a much clearer parity with the modern computing world: finding new ways, for example, to take advantage of large numbers of processing cores (necessary, as Intel released the first consumer-oriented six-core processor earlier this year), and use less power to do everything.

Most users won't notice these changes, but many of them are substantial ones that could be setting the ground for even more exciting developments to come. Read on for our look at what's going on in Sandy Bridge, and what benefits it may hold for computer users and builders starting in 2011.

It can't be easy for a long-running game series to stay fresh after 19 years and several iterations of one general idea, yet Civilization has managed remarkably well. Through its first four versions, the quintessential PC strategy game has evolved from a habit-forming time-waster into a colorful and involved fact of life so addicting, we're astonished the FDA hasn't banned it yet. The newest chapter remains the computer gaming equivalent of crack cocaine. But even with a few more weeks of play under our belts than we had when we presented our in depth first at the game back in august Civilization V still strikes as more a reboot than a revelation.

Overclocking the Nvidia GeForce GTS 450

In our recent overclocking tests of the Nvidia GeForce GTX 460 and GTX 465, we found that pushing the cards to their performance limits essentially provided the equivalent of a free upgrade. After all, the overclocked GTX 460 outperformed a stock-clocked GTX 465, and upping the GTX 465's clock speed boosted performance up to 20 percent. So when Nvidia touted its bargain-priced GeForce GTS 450 video card ($129 list) as being highly overclockable, we were anxious to put it to the test and see if it too might reach into the performance range of its more expensive big brother.

Video card performance is about a lot more than just clock speed. Though the GTS 450 is a decent performer at resolutions of 1,680 by 1,050 or lower, its hardware is pretty stripped down compared to the GTX 460 and higher cards.

As you can see, the lower-end cards actually have faster clock speeds. But with fewer CUDA cores and SMs, and a narrower memory bandwidth, each of the individual components in the GPU has to do more work to render an image, so higher clock speeds become necessary just to maintain a baseline of performance.

In this story, we take the stock GeForce GTS 450 card and, using the standard software available from the card's designer and some trial and error, determine just how fast it can run without introducing instability to the system. What we find is that, once you cut back the number of cylinders under the hood far enough, running them at a higher RPM seems to have diminishing returns.

Though we did get a noticeable speedup from the GTS 450 when we pushed its clock rates higher, the results didn't edge it into the performance territory of pricier cards. Let's take a closer look.

The Wireless Web

The wireless web refers to the use of the World Wide Web through equipments like cellular phones, pagers,PDAs, and other portable communications devices. The wireless web service offers anytime/anywhere connection.

Bluetooth

A short range wireless technology. Operate at approx. 1Mbps with range from 10 to 100 meters. Bluetooth is an open wireless protocol for data exchange over short distances.

Wireless LANs

Wireless local area network use a high-frequency radio technology similar to digital cellular and a low-frequency radio technology. Wireless LANS use spread spectrum technology to enable communication between multiple devices in a limited area. Example of open-standard wireless radio-wave technology is IEEE 802.11b.

Cellular and PCS Systems

Use several radio communications technologies. The systems are divided to different geographic area. Each area has low-power transmitter or radio relay antenna device to relay calls from one area to the next area.