The flash industry's race to smaller processs geometries has hit another milestone. This time, Micron says it's started sampling the industry's smallest 128Gb (16GB) MLC flash chip. The chip is fabbed using 16-nm process technology, which is finer than the 20-nm process currently used to build the firm's 128Gb dies. In fact, Micron claims the 16-nm process is the "most advanced processing node for any sampling semiconductor device."
Micron plans to bring its 16-nm, 128Gb chips into "full production" next quarter. It will feature them in a new family of solid-state drives due next year, as well. You can also expect to see these chips in other products—USB thumb drives, flash cards, tablets, smartphones, and server SSDs for data centers are all mentioned in the Micron announcement.
Moving production of 128Gb chips to the finer process will mean more chips can be crammed into each silicon wafer. In theory, that should translate into lower production costs per chip, and it may lead to lower SSD prices at retail.
That said, 128Gb chips aren't ideal for affordable, low-capacity drives. Today's SSD controllers tend to be eight-channel affairs that can tap into four NAND dies per channel, so they need 32 dies to achieve peak performance. However, only 16 128Gb chips are needed to build a 240-256GB drive, and just eight are required for a 120-128GB offering. Crucial's 240GB M500 SSD, which uses 16 of Micron's 20-nm, 128Gb chips, is markedly slower than its 480GB sibling. I'd expect future SSDs based on Micron's 16-nm, 128Gb dies to see a similar performance dropoff at lower capacity points.
|Rumor: Radeon R9 285 to arrive on September 2||31|
|Deal of the week: Devil's Canyon for $194.99||7|
|This gameplay clip made me pre-order The Vanishing of Ethan Carter||27|
|Leaked slides may expose next-gen NUCs||5|
|Thursday Evening Shortbread||22|
|Specs for upcoming FX-8300 chips leak out||64|
|Report: Windows Threshold preview planned for Sept 30||31|
|Only a few hours remain to win ~$1k of hardware via haiku||25|