Skip to main content

MIT researchers show off supercomputer that treats NAND storage like RAM

When you’re dealing with big data, supercomputers, and massive servers, a one percent increase in energy and price efficiency can mean massive savings. You can imagine, then, how big of a deal it is that MIT researchers have found a way to replace traditional RAM with flash memory, because “it’s about a tenth as expensive, and it consumes a tenth as much power.”

The RAM is where data is stored while a computer is actively working with it. The advantage here is that traditional DRAM can be read and written at incredible speeds — far greater than that of flash memory, which is a similar order of magnitude faster than hard disk drives. The disadvantage is that storage is limited, and DRAM can’t store data unless an electrical charge is supplied to it, which isn’t great for power consumption.

Recommended Videos

Until now, the standard operating procedure was to use the DRAM as much as possible in order to lower latency. Since using a lot of RAM is cost prohibitive, both for initial cost and energy use, not all of the necessary information could be stored there. The rest of the information is stored on disk-based drives, which are cheaper and boast huge storage size, but slow access times. Very sporadically, typically less than ten percent of the time, the computer has to access that information. The result is a hitch in its giddy-up, but a big one, and the biggest contributor to slower processing times.

The MIT researchers found that in a number of common use cases for supercomputers, they were able to use NAND flash, like you might find in your PC’s solid state drive, to act as the RAM for the entire process. Even though the read and write speeds are slower overall, the computer never had to access a disk-based drive, eliminating the longest latency period.The researchers found that as long as a process would access the disk drive more than five percent of the time, the NAND architecture is a comparable choice for speed, but a much better one for energy management.

NAND-based flash memory has been a huge boon to data access across both mobile and performance applications. It far outpaces the traditional disk drives in terms of speed and power consumption, but its not without its own issues. With barriers like price and availability not limiting access, they’re becoming the norm for systems from smartphones to servers. Limited speed compared to conventional RAM means they haven’t yet found their way into desktop or mobile system memory yet, but this new development from MIT could start to change that.

The researchers tested the system across three different algorithms – searching for similar images in a huge database, running Google’s PageRank, and Memcached, a common database system used on large websites. They found that through specially designed chips and software, they were able to stay competitive with speeds traditional DRAM based systems were running at, while cutting power usage and price across the board.

For now, this is basically a proof of concept, but the system wouldn’t be too hard to implement in real world systems. This is the bleeding edge of super-computing and memory architecture, and the sort of implementation that may see consumer application down the road, especially as PCIe SSDs become more common.

Brad Bourque
Former Digital Trends Contributor
Brad Bourque is a native Portlander, devout nerd, and craft beer enthusiast. He studied creative writing at Willamette…
SanDisk’s latest drive sets new benchmark for consumer NVMe SSDs
The SanDisk WD Black SN8100 PCIe Gen 5 SSD with and without heatsink variants

SanDisk has officially introduced the WD Black SN8100, its latest high-end PCIe Gen 5 NVMe SSD targeting PC enthusiasts, gamers, and professional users. With sequential read speeds of up to 14,900 MB/s and write speeds of 14,000 MB/s, the drive sets a new bar for consumer SSD performance, surpassing some of the best NVMe SSDs currently on the market, including the Crucial T705. 

The SN8100 uses a standard M.2 2280 form factor and is available in capacities of 1TB, 2TB, 4TB, and 8TB. It’s worth noting that the 1TB model offers lower write speeds, up to 11,000 MB/s, compared to the higher-capacity versions, which reach up to 14,000 MB/s. 

Read more
Pairing the RTX 5090 with a CPU from 2006? Nvidia said ‘hold my beer’
RTX 5090.

Nvidia's best graphics cards are often paired with expensive CPUs, but what if you want to try a completely mismatched, retro configuration? Well, that used to be impossible due to driver issues. But, for whatever reason, Nvidia has just removed the instruction that prevented you from doing so, opening the door to some fun, albeit nonsensical, CPU and GPU combinations.

The instruction in question is called POPCNT (Population Count), and this is a CPU instruction that also prevents Windows 11 from being installed on older hardware. Its job is counting how many bits are present in a binary number. However, as spotted by TheBobPony on X (Twitter), POPCNT will not be a problem for Nvidia's latest graphics cards anymore.

Read more
AMD’s upcoming CPU could offer bonkers gaming performance
A fake and real AMD Ryzen 7 9800X3D side by side.

AMD's Zen 5 architecture has been a popular choice for gamers due to its outstanding performance and 3D V-Cache capacity, and now a leak suggests Zen 7 could double down on that through a new "3D Core." According to YouTuber Moore's Law is Dead, "[AMD] is moving toward a lot of official variants."

AMD reportedly plans to launch a single overall architecture, divided into different product categories, including the expected lineup: Classic Cores, Dense Cores, Efficiency Cores, and Low-Power Cores. The 3D Core is the latest addition, and it is said to "require full cache chiplets" that "seem to be leading to profound performance increases."

Read more