Say hello to Kal-El, Tegra 2's quad-core successor

Nvidia has scored a solid number of design wins for its Tegra 2 processor, which is going to find its way into many of the iPad competitors due out this year. The chipmaker isn't resting on its laurels, though. At the Mobile World Congress in Barcelona, Nvidia says it gave the first public demonstration of Kal-El, Tegra 2's successor.

Kal-El will feature four ARM cores as well as a "12 core" GeForce GPU—that presumably means a graphics component with 12 ALUs, or stream processors. Samples have already gone out to Nvidia's customers, who are purportedly planning to kick off production of Kal-El-based gear in August.

The Mobile World Congress demo involved web browsing, games, and streaming of 1440p video on a 2560x1600 panel. As Nvidia points out, the high-res video streaming wasn't just a show of strength. The firm has made provisions for mobile devices with 10.1" displays that have a 300-DPI pixel density. If my math is right, a 300-DPI, 10.1" display with a 16:10 aspect ratio would have a resolution of about 2569x1606, so that's not far off the mark.

Nvidia didn't capture the MWC demo on video, but its blog post includes a clip of a Kal-El development system running a browser benchmark on YouTube:

There's also a second video that shows Kal-El running the Coremark benchmark on Android. Apparently, Kal-El managed to outperform Intel's Core 2 Duo T7200 processor in that test. The benchmark is clearly heavily multi-threaded, so Kal-El's extra cores must give it an advantage—and, you know, this is a test hand-picked and run by Nvidia. Still, we're talking about an ARM-based solution here, not an x86 CPU like the Core 2.

What's next? Nvidia's MWC blog post shows a roadmap of upcoming Tegra products code-named Wayne, Logan, and Stark, which are scheduled for 2012, 2013, and 2014, respectively. Stark will purportedly be a whopping 75 times quicker than Tegra 2. Crysis 3 on next-gen slates, anyone?

Tip: You can use the A/Z keys to walk threads.
View options

This discussion is now closed.