SAN FRANCISCO—As expected, Nvidia Corp. Wednesday (Nov. 9) formally launched its Tegra 3 processor, codenamed "Kal-El," a quad-core processor that has been designed into the Asus EE Pad Transformer Prime media tablet.
According to Nvidia (Santa Clara, Calif.), Tegra 3 provides up to three times the graphics performance and up to 61 percent lower power consumption compared to the the previous-generation Tegra 2. This translates into 12 hours of battery life for HD video playback, according to the company.
Tegra 3 was initially expected to be available by the end of this summer of in early September, but the company said last month it would be officially launched Wednesday. Earlier, some reports suggested that the processor faced longer delays.
Nvidia said Tegra 3 implements a new, patent-pending technology known as variable symmetric multiprocessing (vSMP), which includes a fifth "companion" CPU core specifically designed for work requiring little power. The four main cores are specifically designed for work requiring high performance, and generally consume less power than dual-core processors, Nvidia said.
During tasks that require less power consumption—like listening to music, playing back video or updating background data—the Tegra 3 processor completely shuts down its four performance-tuned cores and, instead, uses its companion core. For high-performance tasks—like web browsing, multitasking and gaming—the Tegra 3 processor disables the companion, Nvidia said.
"Nvidia's fifth core is ingenious," said Nathan Brookwood, Research Fellow at Insight 64, in a statement issued by Nvidia. "Tegra 3's vSMP technology extends the battery life of next-generation mobile devices by using less power when they're handling undemanding tasks and then ratcheting up performance when it's really needed."
The Tegra 3 quad-core CPUs are complemented with a new 12-core NVIDIA GeForce GPU, which delivers more realism with dynamic lighting, physical effects and high resolution environments, plus support for 3-D stereo, Nvidia said.
Hey Al.Key, what makes you think it is "implemented quite well"? Do you work for Nvidia to know this? Or do you have access to one to know this? Dont say something before you can vouch for it. Marvell announced it bizillion years before Nvidia. This is what TI does on their OMAP4 with Cortex-M3. I dont know what Nathan is smoking to say that this is "ingenious". He needs to explain or else he sounds "dis-ingenous"
No, it is not big-little. It is much better. The big little had the low power core as A7 while the high power cores were A.15.
What NVIDIA did is different and remarkable. It is the same type of core (so you don't need to write your mp3 player for both types of cores). But IT IS A DIFFERENT TECHNOLOGY PROCESS in the same SoC. A high power and a low power.
Good job NVIDIA.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.