Samsung Starts Mass Production of Second-Gen 16GB LPDDR5 RAM for Future Premium Smartphones
Samsung has announced that it will kick off mass production of the world's first 16GB LPDDR5 RAM package for future smartphones. Last year, the Korean giant stated that it started mass production of 12GB LPDDR5 RAM. For 2020, Samsung has taken that production dial to the next phase and claims that the new RAM packages will enable users to experience enhanced 5G and AI features ranging from graphic-rich gaming and smart photography.
According to the company, the data transfer rate for the 16GB LPDDR5 [package] is 5500Mb/s (megabits per second), making it significantly faster than the previous-generation LPDRR4X RAM package, which peaks out at 4266Mb/s. That's not the only benefit of using these chips, because compared to an 8GB LPDDR4X package, the new mobile DRAM can deliver more than 20 percent power savings while offering twice the memory capacity.
16 GB DRAM packages could also be used in single board computers and other compact systems. For example, the BCM2711 SoC used in the Raspberry Pi 4 Model B can theoretically address up to 16 GB of memory.
Samsung press release. Also at AnandTech.
Previously: Samsung Announces 8 GB DRAM Package for Mobile Devices
Samsung Announces LPDDR5 DRAM Prototype Before Specification is Finalized
Samsung Begins Mass Producing 12 GB DRAM Packages for Smartphones
Samsung Mass Producing LPDDR5 DRAM (12 Gb x 8 for 12 GB Packages)
Get Ready for Smartphones with 16 GB of RAM
Related Stories
Samsung has announced an 8 GB LPDDR4 DRAM package intended for smartphones and tablets, using four 16 Gb (2 GB) chips manufactured on a 10nm-class process (probably 18nm):
Samsung this week announced its first LPDDR4 memory chips made using its 10nm-class DRAM fabrication technology. The new DRAM ICs feature the industry's highest density of 16 Gb, are rated to run at 4266 MT/s data rate, and open the door to more mobile devices with 8 GB of DRAM.
Earlier this year Samsung started to produce DDR4 memory using its 10nm-class DRAM manufacturing process (which is believed to be 18 nm) and recently the firm began to use it to make LPDDR4 memory devices, just as it planned. The thinner fabrication technology allowed Samsung to increase capacity of a single LPDDR4 DRAM IC to 16 Gb (up from 12 Gb at 20nm introduced in August, 2015) while retaining a 4266 MT/s transfer rate.
The first product to use the 16 Gb ICs is Samsung's 8 GB LPDDR4-4266 mobile DRAM package for smartphones, tablets, and other applications that can use LPDDR4. The device stacks four memory ICs and provides up to 34 GB/s of bandwidth when connected to an SoC using a 64-bit memory bus. The 8 GB DRAM package comes in a standard 15 mm x 15 mm x 1 mm form-factor, which is compatible with typical mobile devices, but Samsung can also make the package thinner than 1 mm to enable PoP stacking with a mobile application processor or a UFS NAND storage device.
The press release confirms the high data rate:
The new 8GB LPDDR4 operates at up to 4,266 megabits per second (Mbps), which is twice as fast as DDR4 DRAM for PCs working typically at 2,133 Mbps per pin. Assuming a 64 bit (x64) wide memory bus, this can be viewed as transmitting over 34GBs of data per second.
Tune in next year when I post about Samsung putting 12 GB of RAM in smartphones.
Previously:
Samsung Announces 12Gb LPDDR4 DRAM, Could Enable Smartphones With 6 GB of RAM
Samsung Announces "10nm-Class" 8 Gb DRAM Chips
Samsung Announces First LPDDR5 DRAM Chip, Targets 6.4Gbps Data Rates & 30% Reduced Power
[Samsung] is announcing that they have completed fabrication, functional testing, and validation of a prototype 8Gbit LPDDR5 module. The company is targeting data rates up to 6.4Gbps-per-pin with the new memory, and while Samsung isn't ready to start mass production quite yet, the company's press release notes that they're already eyeing it.
[...] In terms of performance, Samsung is targeting up to 6.4Gbps/pin with the new memory. Which for a typical 32-bit bus chip works out to 25.6GB/sec of memory bandwidth. This is a 50% increase in bandwidth over the current LPDDR4(X) standard, which tops out at 4.266Gbps under the same conditions. So for a high-end phone where 64-bit memory buses are common, we'd be looking at over 50GB/sec of memory bandwidth, and over 100GB/sec for a standard 128-bit bus PC.
Samsung Begins Mass Production of 12 GB LPDDR4X for Smartphones
Samsung announced late on Wednesday that it had started volume production of 12 GB LPDDR4X-4266 memory for high-end smartphones. The chip is currently the highest-density DRAM for mobile applications. The first smartphone to use Samsung's 12 GB LPDDR4X DRAM package will be the company's own Galaxy S10+ handset formally announced last month.
Samsung's 12 GB LPDDR4X package integrates six 16 Gb memory devices featuring a 4266 MT/s data transfer rate at 1.1 Volts and produced using the company's second-generation '10nm-class' process technology (also known as 1y-nm). The 12 GB memory module is 1.1 mm tall, which is a bit higher than standard quad-die LPDDR4X packages (which are thinner than 1 mm), but Samsung has managed to incorporate the device into its latest premium smartphone.
Were the previously announced 12 GB DRAM smartphones using two packages instead of this one thick package?
Related: Samsung Announces 12Gb LPDDR4 DRAM, Could Enable Smartphones With 6 GB of RAM
Samsung Announces 8 GB DRAM Package for Mobile Devices
SK Hynix Announces 8 GB LPDDR4x DRAM Package for Mobile Devices
Oppo Likely to Release the First Smartphone With 10 GB of RAM
Xiaomi Announces Smartphones with 10 GB of RAM
Lenovo Announces a Smartphone With Up to 12 GB of RAM
Samsung Commences Mass Production of First Ever 12Gb LPDDR5 DRAM for Premium Handsets
After the LPDDR4X standard, Samsung is ready to take mobile computing to the next plateau, as the Korean giant has announced that mass production of the industry's first 12-gigabit (Gb) LPDDR5 mobile DRAM, a component that has been optimized for 5G and AI features for future smartphones.
Additionally, Samsung plans on mass producing 12-gigabyte (GB) LPDDR5 packages later this month, which each package combining eight of the 12Gb chips. This reveals that future premium devices will demand the best when it comes to faster, more efficient memory, and Samsung wants to be ahead of the curve in both supply and demand.
Data rate will be 5,500 MT/s, compared to 4,266 MT/s for LPDDR4X, with up to 30% less power consumption than LPDDR4X. Future LPDDR5 chips could hit 6,400 MT/s.
Samsung plans to start producing 16Gb LPDDR5 chips next year. Smartphones with 16 GB of DRAM are sure to follow.
Samsung press release. Also at AnandTech.
Previously: Samsung Announces LPDDR5 DRAM Prototype Before Specification is Finalized
Samsung Begins Mass Producing 12 GB DRAM Packages for Smartphones
Micron shared details of its 3rd generation of "10 nm-class" DRAM fabrication:
Micron's 3rd Generation 10 nm-class (1Z nm) manufacturing process for DRAM will allow the company to increase the bit density, enhance the performance, and the lower power consumption of its DRAM chips as compared to its 2nd Generation 10 nm-class (1Y nm) technology. In particular, the company says that its 16 Gb DDR4 device consumes 40% less power than two 8 Gb DDR4 DRAMs (presumably at the same clocks). Meanwhile, Micron's 16 Gb LPDDR4X ICs will bring an up to 10% power saving. Because of the higher bit density that the new 1Z nm technology provides, it will be cheaper for Micron to produce high-capacity (e.g., 16 Gb) memory chips for lower-cost, high-capacity memory sub-systems.
[...] As for mobile memory, Micron's 16 Gb LPDDR4X chips are rated for transfer rates up to 4266 MT/s. Furthermore, along with offering LPDDR4X DRAM packages with up to 16 GB (8x16Gb) of LPDDR4X for high-end smartphones, Micron will offer UFS-based multichip packages (uMCP4) that integrate NAND for storage and DRAM. The company's uMCP4 family of products aimed at mainstream handsets will include offerings ranging from 64GB+3GB to 256GB+8GB (NAND+DRAM).
Finally, a reasonable amount of RAM for smartphones. But I think we may need at least 24 GB, if not 32 GB.
Related: Xiaomi Announces Smartphones with 10 GB of RAM
Samsung Mass Producing LPDDR5 DRAM (12 Gb x 8 for 12 GB Packages)
SK Hynix Commences Mass Production of 18GB LPDDR5 RAM Chips for Smartphones With 6,400Mbps Speeds
Android phone makers will continue to push the limits of hardware specifications, and from the looks of it, SK Hynix will lend out more than just a helping hand. The memory manufacturer today announced that it has started mass production of 18GB LPDDR5 RAM chips for flagship smartphones, meaning that premium handsets touting more memory than notebooks will become a commonplace.
SK Hynix claims that its 18GB LPDDR5 RAM for smartphones can operate up to 6,400Mbps, making it around 20 percent faster than the previous-generation LPDDR5 RAM, which could run up to 5,500Mbps. The manufacturer also mentions that it has supplied ASUS with these DRAM chips for the upcoming ROG Phone 5 flagship. Keep in mind that during a specifications leak, the ROG Phone 5 was spotted with the aforementioned RAM count.
Why does a smartphone need 18 GB of memory instead of the previous 16 GB? From the press release:
"This product will improve the processing speed and image quality by expanding the data temporary storage space, as the capacity increases compared to the previous 16GB product," an official from the company said.
So we will see smartphones with 18 GB of RAM, or perhaps smartphones or laptops with 16/32 GB of error correction code (ECC) LPDDR5 memory.
Previously: Samsung Begins Mass Producing 12 GB DRAM Packages for Smartphones
Samsung Mass Producing LPDDR5 DRAM (12 Gb x 8 for 12 GB Packages)
Get Ready for Smartphones with 16 GB of RAM
Samsung Announces Mass Production of 16 GB LPDDR5 DRAM Packages
Exclusive: Valve is making a Switch-like portable gaming PC
Video game and hardware studio Valve has been secretly building a Switch-like portable PC designed to run a large number of games on the Steam PC platform via Linux—and it could launch, supply chain willing, by year's end.
Multiple sources familiar with the matter have confirmed that the hardware has been in development for some time, and this week, Valve itself pointed to the device by slipping new hardware-related code into the latest version of Steam, the company's popular PC gaming storefront and ecosystem.
[...] In recent years, the "Switch-like PC" category has exploded. In early 2020, Alienware revealed its first Switch-like gaming PC, but the "concept" device has not yet turned into a commercial product. If you want to buy a similar device today, you're largely looking at products from Chinese OEMs like GPD, One-Netbook, and Aya, who have slapped ultramobile PC processors and parts into a Switch-like chassis.
Rumors point to an AMD "Van Gogh" APU (Zen 2 quad-core with RDNA 2 graphics and support for LPDDR5 RAM), 7/8-inch screen, at a $400 price point for a Q4 2021 release.
Also at Wccftech.
Samsung Announces LPDDR5X DRAM for Smartphones; 1.3x Faster Than LPDDR5 With Speeds up to 8.5Gbps
Samsung today officially announced LPDDR5X DRAM chips for smartphones and other applications. Compared to the LPDDR5 standard, the new chips bring increased speeds, and it will be no surprise that we will see them in action in several 2022 flagship handsets.
[...] In contrast to LPDDR5's 6.4Gbps maximum bandwidth, LPDDR5X can achieve 1.3-times the performance with processing speeds that go up to 8.5Gbps. Samsung has used its 14nm technology to mass produce the next-generation DRAM chips, and it will be advantageous for portable devices too because the new standard is 20 percent more energy-efficient than LPDDR5.
The press release says that 16Gb LPDDR5X chips will enable 64 GB memory packages, "accommodating increasing demand for higher-capacity mobile DRAM worldwide." In other words, Samsung is planning to put 32 dies in a single package, and eventually stick 64 gigabytes of memory in smartphones (or tablets, or laptops). Recently, Samsung has been making 16 GB packages with only 12 or 8 dies:
The 16Gb LPDDR5 can build a 16GB package with only eight chips, whereas its 1y-based predecessor requires 12 chips (eight 12Gb chips and four 8Gb chips) to provide the same capacity.
Also at AnandTech.
Previously: SK Hynix Announces 8 GB LPDDR4x DRAM Package for Mobile Devices
Samsung Announces LPDDR5 DRAM Prototype Before Specification is Finalized
Samsung Announces Mass Production of 16 GB LPDDR5 DRAM Packages
SK Hynix Begins Production of 18 GB LPDDR5 Memory... for Smartphones
(Score: 4, Interesting) by DannyB on Tuesday February 25 2020, @08:54PM (8 children)
Imagine cheap $35 dollar small computer boards that have four cpu cores, gigabytes of memory, and solid state fast storage.
In January 1975, an Altair 8800, with 1 K of memory was expensive, bulky, heavy, slow, power hungry and required hand loading a boot loader on the front panel.
Imagine the power of packing that large Altair 8800 box with these hypothetical cheap computer boards I speak of.
Now think of our future. In time, you can expect small inexpensive computers that have hundreds of cpu cores, astonishing amounts of memory, capable of running software of untold complexity and abstraction. And C programmers will still whine that their low level language has the best performance, and pretend that it somehow matters. Developer time will still cost money. Most problems to solve will be even more complex than today's.
If a lazy person with no education can cross the border and take your job, we need to upgrade your job skills.
(Score: 1, Insightful) by Anonymous Coward on Tuesday February 25 2020, @09:16PM (3 children)
... and all that power will be used to try to sell you stuff you don't need.
(Score: 4, Funny) by takyon on Tuesday February 25 2020, @09:17PM (2 children)
Like a foldable smartphone with 32 GB of RAM?
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Wednesday February 26 2020, @01:48AM
will 32GB even be enough to containerize a docker instance on a Hackintosh-Air ARM64 edition?
(Score: 2) by DannyB on Wednesday February 26 2020, @02:58PM
What I want is for a phone to fold eight times to become so small it disappears into the quantum foam so that it doesn't cause a large bulge in my pocket.
This would be an improvement on George Jetson's car folding into a briefcase that is too heavy to lift.
If a lazy person with no education can cross the border and take your job, we need to upgrade your job skills.
(Score: 5, Interesting) by takyon on Tuesday February 25 2020, @09:16PM (3 children)
The future is going to be monolithic 3D chips with either x86, ARM, or RISC-V cores and memory placed just tens/hundreds of nanometers away from the cores. The performance increase will be so vast that single board computers or docked smartphones could replace most desktops. Ideally, we will see a universal memory technology with much higher density and lower cost than today's DRAM, also capable of replacing NAND.
I don't know how many more layers of abstraction can be added. But if there is a demand to optimize software for every iota of performance, such as for supercomputing, then someone will get paid to do it.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: -1, Offtopic) by Anonymous Coward on Wednesday February 26 2020, @07:40AM
Gu tossed Xiu Xiang's project onto the lawn. He reached into the drive compartment and flipped out the loose hull section. A ragged and maybe disdainful cheer rose from the kids behind him. "Hey, dork! There has to be a latch. Why didn't you scam the lock?"
Gu didn't seem to hear. He leaned forward to look into the interior. Juan edged closer. The compartment was in shadow, but he could see well enough. Not counting damage, it looked just like the manual said. There were some processor nodes and fiber leading to the dozens of other nodes and sensors and effectors. There was the steering servo. Along the bottom, just missed by Gu's cutting, was the DC bus to the left-front wheel. The rest was empty space. The capacitor and power cells were in the back.
Gu stared into the shadows. There was no fire, no explosion. Even if he had chopped into the back, the safeties would have prevented any spectacular outcome. But Juan saw more and more error flags float into view. A junk wagon would be coming real soon.
Gu's shoulders slumped, and Juan got a closer look at the component boxes. Every one had physical signage: "No user-serviceable parts within".
The old guy stood and took a step away from the car. Behind them, Chumlig and now Williams were on the scene, herding the students back into the tent. For the most part, the kids were fully stoked by all the insanity. None of them, not even the Radner brothers, ever had the courage to run amok. When they committed something major, it was usually done in software, like what the guy had shouted from the crowd.
Xiu Xiang gathered up her weird, Gu-improved, project. She was shaking her head and mumbling to herself. She unplugged the gadget and took a step toward Robert Gu. "I object to your appropriation of my toy!" she said. There was an odd expression on her face. "Though you did improve it with that extra bend." Gu didn't respond. She hesitated. "And I never would have run it with line power!"
Gu waved at the guts of the dead car. "It's Russian dolls all the way down, isn't it, Orozco?"
Juan didn't bother to look up "Russian dolls". "It's just throwaway stuff, Professor Gu. Why would anyone want to fool with it?"
(Score: 2) by DannyB on Wednesday February 26 2020, @03:19PM (1 child)
Memory closely integrated with processors at the chip level makes sense. You would upgrade memory and processing power together.
Another thing I think will eventually happen, but that will be controversial.
Hardware assisted GC
Note that all modern languages in the last 2 freaking decades have garbage collection. Remember "lisp machines" from the 1980's? Like Symbollics? Their systems didn't execute Lisp especially fast, but what they did was provide hardware level assistance for GC which made GC amazingly fast.
I look at the amazing things JVM (Java Virtual Machine) has done with GC. If only the JVM's GC could benefit all other languages (Python, JavaScript, Go, Lisps, etc). Of course, those languages could use JVM as a runtime. And GraalVM _might_ make something like that happen where lots of different languages run in the same runtime and can transparently call each others functions and classes and have a common set of underlying data types. Red Hat's Shenandoah and Oracle's open source ZGC are amazing garbage collector technology. Terabytes of memory with 1 ms GC pause times. Now imagine if you had hardware assistance for GC. (btw, why is Red Hat investing so much into Java development? I thought they were a Linux company? Could Red Hat, which is a publicly tiraded company, have some economic reason Java is making them lots of money?)
Rationale: GC is an economic reality. Ignore the whining of he C programmers in the peanut gallery for a moment. They'll jump up and down and accuse other professionals of not knowing how to manage memory. Ignore it. Why do we use high level languages (like C) instead of assembly language? Answer: human productivity! Our code would be so much more efficient if we wrote EVERYTHING including this SN board directly in assembly language!!! So why don't we??? Because, as C programmers are simply unwilling to admit, the economic reality is that programmers are vastly more productive in higher and ever higher level languages. Sure there is an efficiency cost to this. But we're optimizing for dollars not for bytes and cpu cycles. Hardware is cheap, developer time is expensive.
Slight aside: ARM processors already have some hardware provision for executing JVM bytecodes (gasp! omg!).
I'm surprised that modern Intel or AMD designs haven't introduced some hardware assistance for GC.
Symbollics hardware, IIRC, had extra bits in each memory word (36 bit words I think) to "tag" the type of information in every word. Then a way to efficiently find all words that happened to be a "pointer". A way to tag all words that were "reachable" or "marked" from the root set, etc.
Maybe this can happen if memory and processing elements become highly integrated and interconnected. Hardware design will follow the money just as programming languages and technology stacks do.
Others will believe that system design will stand still to conform to a romantic idealism that was the major economic reality once upon a time.
If a lazy person with no education can cross the border and take your job, we need to upgrade your job skills.
(Score: 2) by DannyB on Wednesday February 26 2020, @03:33PM
I just reposted this as a Journal entry. [soylentnews.org]
If a lazy person with no education can cross the border and take your job, we need to upgrade your job skills.