The UFS standard aims to bring us lightning fast microSD cards that perform on-par with SATA SSDs. Samsung introduced theirs earlier this month, and now Micron has announced their solution:
As you can see, UFS is not just for SD cards. These are going to be able to replace embedded memory in mobile devices, displacing the horror that is eMMC with something way faster. These devices are smaller than a penny, with a die size of just over 60 mm squared and boast a 32GB capacity.
One version of the UFS 2.1 devices also contains Micron's first packaged offering of LPDDR4X. This low power RAM offers an additional 20% power savings over existing LPDDR4.
Also up is an overdue branding of Micron's XPoint (spoken 'cross-point') products:
QuantX will be the official branding of Micron products using XPoint technology. This move is similar to the one Intel made at IDF 2015, where they dubbed their solutions with the Optane moniker.
More to follow from FMS 2016. A few little birdies told me there will be some good stuff presented this morning (PST), so keep an eye out, folks!
Press blast for Micron's UFS goodness appears after the break.
Micron Introduces Mobile 3D NAND Solution for Next-Generation Smartphones
New Micron Mobile 32GB 3D NAND memory technology delivers better quality, performance and reliability for next-generation mobile multimedia and streaming experiences
Santa Clara, CALIF., August 9, 2016 — Micron Technology, Inc., (Nasdaq: MU) today introduced the company’s first 3D NAND memory technology optimized for mobile devices and its first products based on the Universal Flash Storage (UFS) 2.1 standard. Micron’s initial mobile 3D NAND-based 32GB solution is targeted specifically for the high and mid-end smartphone segments which make up approximately 50 percent of worldwide smartphone volume [1]. As mobile devices bypass personal computers as consumers’ primary computing device, user behaviors heavily impact the device’s mobile memory and storage requirements. Micron’s mobile 3D NAND addresses these concerns, enabling an unparalleled user experience that includes seamless high definition video streaming, higher bandwidth gameplay, faster boot up times, camera performance and file loading.
“Micron continues to advance NAND technology with our introduction of 3D NAND and UFS products for the mobile segment,” said Mike Rayfield, vice president of Micron’s mobile business unit. “The improved performance, higher capacity and enhanced reliability of 3D NAND will help our customers meet the ever-growing demand for mobile storage and will enable much more exciting end user experiences.”
To meet the elevated hardware demands stemming from increased mobile video and multimedia consumption, as well as the anticipated increased storage demands that will result from the introduction of 5G wireless networks, Micron 3D NAND technology stacks layers of data storage cells vertically with extraordinary precision to create storage solutions with three times more capacity than previous generation planar NAND technologies. Because capacity is achieved by stacking cells vertically, Micron is able to pack more storage cells into a much smaller die area, resulting in the delivery of the industry’s smallest 3D NAND memory die measuring only 60.217mm2. A smaller die enables a tiny memory packaging footprint which can free up space for additional mobile battery size or enable smaller form factor devices.
“3D NAND technology will be vital to the continued development of smartphones and other mobile devices,” said Greg Wong, founder and principal analyst at Forward Insights. “With the advent of 5G and mobile’s increasing influence in our digital lives, smartphone makers are in need of the most advanced technology to store and manage the ever-increasing volume of data. Micron’s 3D NAND for mobile is well suited to address the market’s evolving data storage needs by enabling a more seamless user experience for high resolution video, gaming, and photography.”
3D NAND: Powering the Changing Mobile Landscape
Micron's first 3D NAND for mobile offers several competitive technical advantages. New features include:
- Industry’s first mobile product built on floating gate technology, a universally utilized design refined through years of high-volume planar flash manufacturing
- Micron’s first memory devices with UFS 2.1 standard, which enable best-in class Sequential Read Performance for the mobile market
- 3D NAND-based multichip packages (MCPs) also include low power LPDDR4X, providing up to 20 percent more energy efficiency than standard LPDDR4 memory
- Industry’s smallest 3D NAND memory die, measuring only 60.217mm2, allows tiny memory packaging ideal for ultra-small form factor devices; Micron’s 3D NAND die is up to 30% smaller than planar NAND die of the same capacity.
Micron’s 3D NAND solutions for mobile are now sampling with mobile customers and partners and will be widely available by the end of 2016.
[1] Source: Micron internal forecast
Additional Resources:
- Lean more about Micron mobile memory solutions
- Learn more about Micron’s 3D NAND products
- 3D NAND for Mobile Data Sheet
Follow us online! Take part in Micron's social conversations where we're talking all things storage and memory:
- Blog: www.micron.com/about/blogs
- Twitter: www.twitter.com/MicronTech
- LinkedIn: www.linkedin.com/company/micron-technology
- YouTube: www.youtube.com/microntechnology
So now for AMD, Nvidia,
So now for AMD, Nvidia, Samsung, SK Hynix, etc. to all get together with JEDEC and get a new HBM/NVM standard to allow for the Micron XPoint(QuantX) to be added to the HBM stacks and wired up through TSVs to the DRAM dies for some really fast NVM texture storage/other storage right on the HBM stacks. XPoint is many times denser than DRAM but not as fast as DRAM, but XPoint is many times faster than NAND and many times as durable! So both AMD’s and Nvidia’s GPUs could make use of a New JEDEC HBM standard that allows for an XPoint die on the HBM stacks with the DRAM die stacks communicating over TSVs. Some on HBM NVM to store textures and Kernels that the GPU can get at quickly with algorithms in the HBM’s bottom control logic die that are able to intelligently predict what textures/kernels code should be moved in the background to the HBM’s DRAM from the on HBM NVM to hide any XPoint to DRAM latency.
There are already such products just announced for the DRAM/NVM DIMM market with algorithms in the DIMM’s control logic that are able to intelligently predict and hide most NAND to DRAM latency, for that new DIMM DRAM/NVM(NAND) product just announced.
Here is the new DIMM DRAM/NVM
Here is the new DIMM DRAM/NVM product from Samsung/Netlist!
“Say hello to Samsung and Netlist’s flash-DRAM grenade: HybriDIMM”
http://www.theregister.co.uk/2016/08/08/samsung_and_netlist_hybridimm/
AMD Radeon SSG anybody?
AMD Radeon SSG anybody?
Yes that SSG is fast over its
Yes that SSG is fast over its on card PCIe connection, but the next step is to go with a new JEDEC XPoint/NVM on the HBM stacks Standard so both GPU makers can have a very fast on HBM die stacks NVM store. I’d imagine that for AMD’s APUs on an interposer having some gigabytes of XPoint NVM on the Interposer based APU’s HBM die stacks would be very desirable for both the Zen cores die, and the Polaris/Vega die to get at. Having XPoint ON HBM NVM wired up directly to the HBM’s dies with TSVs is going to have a higher effective bandwidth/wider connection to the NVM XPoint HBM die than any PCI SSG NVM. Nvidia will probably get its own version of on GPU card NVM similar to AMD’s SSG, but why not go with NVM on the HBM/stacks JEDEC standard and wire things up with thousands OF TSVs, and many times faster.
AMD hasn’t shown direct to
AMD hasn't shown direct to NVMe communications yet. Their demo system had those SSDs mounted by the host OS. Their vision is to mount the storage directly by the GPU, but they are not there yet (and it's not a simple problem to solve, either).