Samsung on Monday announced what it claims is the industry's first 30nm class DRAM to successfully complete customer evaluations in 2Gb (gigabit) densities.
"Our accelerated development of next generation 30nm-class DRAM should keep us in the most competitive position in the memory market," said Soo-In Cho, president, Memory Division, Samsung Electronics. "Our 30nm-class process technology will provide the most advanced low-power DDR3 available today and therein the most efficient DRAM solutions anywhere for the introduction of consumer electronics and server systems."
According to Samsung, shrinking down to a 30nm manufacturing process allows the company to raise production by 60 percent over 40nm-class DDR3. And as far as consumers are concerned, the company's Green DRAM lowers power consumption by up to 30 percent over 50nm-class DRAM. To give a real world example, Samsung says a 4GB, 30nm module will consume only 3W per hour in a new generation notebook.
Any large technology company relies on their server infrastructure to serve their customers. The sort of power that runs Google or Facebook doesn’t come cheap. It's not so much the cost of the hardware, it’s the massive cost of powering that infrastructure that eats into the bank account. Two start-ups aim to change the server game with some new, low-power alternatives to conventional servers.
SeaMicro, from Santa Clara, is putting together servers based on the low power Atom chip seen most often in Netbooks. Those in the know have indicated that SeaMicro will be able to pack 80 Atom chips in a very small chassis. These Atom servers would offer massive reductions in energy costs, but still provide adequate processing power to serve up data. After all, how much power does it really take to push out some Google results?
In Austin, Texas, there’s an even more ambitious server project afoot. Smooth-Stone is working to integrate the ARM chips you’ve seen in smartphones, like the iPhone, into a new server architecture. Smooth-Stone CEO, Barry Evans, accumulated a great body of knowledge working for Intel’s mobile products group. This seems to jive nicely with the company’s apparent goals. Details on this one are scarce, but if the performance is sufficient, the energy savings could be staggering. Could it be that the era of companies running rack after rack of Xeon-based web servers is coming to a close?
Nvidia has licensed Transmeta’s power conserving technology for a sum of $25 million. The technologies that Transmeta has leased out to Nvidia include its flagship power management technologies, Longrun and Longrun 2. Transmeta has quickly mastered its current business model of licensing IP to bigger companies and its coffers are loaded with cash.
It shouldn’t surprise anyone that Nvidia has licensed Transmeta’s power management technology as most chip manufacturers are concentrating on increasing power efficiency.