AMD continues to be a thorn in market leader Intel’s side.
The latest figures from Mercury Research shows AMD’s Opteron processors now account for 22.1 percent of the market for x86 server processors chips shipped worldwide in the first quarter, up from 16.4 percent last quarter and a whopping 254 percent increase over AMD’s share a year ago.
Precise figures for Intel won’t be released until next week, but Mercury analyst Dean McCarron said Intel “makes up most of the rest” of the market at about seventy-five percent.
Intel did not address AMD’s server gains directly, but did roll out a new business brand and features aimed at corporate computing this week. However, the new vPro brand technologies are aimed at the desktop and won’t be available till later this summer along with a new “Conroe” Core micro architecture dual-core processor.
At Monday’s vPro event and other forums Intel has noted the power saving efficiency of its forthcoming chips. Intel’s CEO Paul Otellini said on Monday that Intel could save a company deploying 50,000 Conroe-based PCs over $1.5 million in energy costs over the lifetime of those systems (typically four years) versus earlier Intel-based models.
But AMD also has a strong performance-per-watt story for systems already in the market.
“I give Intel credit in selling that message that they’re going to have great power efficiency,” Steve Demski, AMD Opteron product manager, told internetnews.com. “But we have the edge now and we’re not sitting still.”
Demski pointed to new Opterons coming later this summer that use DDR-2 memory, which Intel already uses. But AMD will continue to integrate it’s memory controller to the processor, considered a more efficient design than Intel which uses an external memory controller with its own power requirements.
In related news, the HyperTransport Consortium released version 3.0 of the HyperTransport specification. Developed by AMD, HyperTransport
AMD, a founding member of the consortium, said future AMD64 processors will be designed to take advantage of the new spec, which is designed for faster processor-to-processor and processor-to-I/O throughput in multiprocessor servers and multi-system environments.