* Intel on the ‘*T technologies’ that will improve processor performance
Intel recently provided an update on what it calls “*T technologies,” such as Hyper-Threading, Vanderpool, LaGrande and Centrino Mobile technology – all of which will help shift the company away from the exclusive pursuit of gigahertz.
The *T’s (pronounced “star T’s,” not “splat T’s”) are part of Intel’s “platformization strategy” to ensure its products provide a real solution to users’ problems. They address how Intel is going to get increased parallelism and performance out of the computing elements in the platform.
The company says the *T jargon originated in its technical groups.
“We’re a bunch of tech geeks here at Intel,” says Ron Curry, director of marketing for the Corporate Technology Group in Santa Clara. “The asterisk in computer-speak is a wildcard character, so they call it ‘*T’s’ – and it just seemed to be the right thing to do to keep the family of technologies *T’s.”
Curry said that as Intel has improved silicon processor technology over the last 10 years the company has continued to increase clock speed. Intel is now looking to provide processor improvements that will increase the clock speeds over time but also address power envelope needs and get more dramatic increases in performance. To do that Intel is moving from Hyper-Threading toward dual- and multi-core technology, in which two processors reside on the same die.
Several of the *T technologies, such as Hyper-Threading and Centrino Mobile technology, are already built into Intel processors.
The *T’s Intel is working on now are LaGrande, Vanderpool, active memory technology and EM64T, which it introduced last summer. LaGrande is security technology that lets a PC protect itself against viruses or denial-of-service attacks. Vanderpool is hardware-enabled virtualization technology.
In this newsletter we’ll talk about Hyper-Threading. Later this week we’ll talk about LaGrande and Vanderpool.
Intel first introduced Hyper-Threading in 2002, in the Pentium processor line. Hyper-Threading is a part of a path to get increased performance out of processors regardless of cost. It adds more parallelism to the processor.
The Pentium was capable of executing two instruction streams and it had two pipelines. If one instruction stream got busy, it could start executing the other instruction stream. The Pentium Pro incorporated an innovation called out-of-order execution, which allowed it to be more efficient and more effectively execute instruction streams, Curry says.
“In the Pentium 4 generation we introduced Hyper-Threading,” Curry says. “Hyper-Threading allows the processor to appear as if it is two processors, so the operating system thinks it is running on a dual-processor machine.”
Hyper-Threading is, Curry claims, a stepping-stone to the upcoming dual-core and multi-core architectures that Intel will introduce next year. It helps Intel get software in place that will execute on those dual-core and multi-core machines. Intel has shipped 50 million Hyper-Threading-enabled processors.
