Feeding The Smartphone Frenzy? Its A Matter Of Time And Space
Sponsored Feature When considering a new smartphone, consumers will generally evaluate the screen size, on-board storage, battery, and possibly, the processor. They're far less likely to consider the amount of on-board DRAM, never mind its specific characteristics.
Which seems a glaring omission. Not just because every smart device in fact uses DRAM, but because its capacity and speed has a direct bearing on the performance of the applications users find most compelling, and which drive mobile phone innovation more broadly.
Afterall, we're all generating incredible amounts of data. IDC predicts total annual data creation will reach 221 zettabytes (ZB) in 2025, compared to just 1ZB as recently as 2010, and 81ZB in 2021. Smartphones are a major contributor to this information explosion, whether through the creation of photos, videos and audio on individual devices which then finds their way to the cloud and beyond, or through the capture of the more ephemeral stuff which is processed by back-end platforms.
Moreover, cutting edge applications require ever larger amounts of internal memory with massive bandwidth and low latency to run smoothly and efficiently. These can include natural language models with upwards of 1 trillion parameters, or imaging systems that can seamlessly remove unwanted individuals from images, or produce perfectly balanced night vision images.
Whatever the use case, the memory inside of your smartphone underpins the overall experience. Chris Moore, vice president of marketing at Micron's mobile business unit, explains that many of the glitches users may experience while using their smartphones are simply due to insufficient DRAM. Users with older smartphones with smaller memory capacities on older interfaces might experience latency when opening apps, for example. Opening data intensive and more demanding apps can maximize the available amount of DRAM capacity, which in turn will force other apps to close instead of staying open and active in the background.
Some of this performance inefficiency may also be due to the processors in older devices, he adds, "A big part of the overall performance challenge may be due to the DRAM itself. Once your app is opened, can it stay open? Or do you get that load screen again and have to go through that re-launch waiting experience."
"Some end users may not notice the delay of the application responsiveness they are using on the smartphone," Moore says. "However, once you notice it, you become keenly aware of the lack of responsiveness."
So, there is a significant imperative to deliver more DRAM for smartphones. But the clever trick is to do that without compromising the form factor, or battery life, by simultaneously making the device larger or more power hungry.
Which is why Micron's launch of LPDDR5X memory built on its 1-beta DRAM node promises to put memory at the center of the conversation. LPDDR5X for mobile delivers a 16Gb per die capacity alongside a data rate of 8.5Gbps, a 33 percent increase on LPDDR5.
"From a density point of view, the 1-beta node gets about a 35 percent improvement over 1-alpha," says Moore. "That's where you're getting the geometry savings here at the node level." While the initial products will be 16Gb per die, Moore says "There's nothing to keep us from going above that in the future."
Power player
A key feature is Micron's High-K Metal Gate (HKMG) technology, which contributes to power and performance efficiencies. The architecture also features new JEDEC enhanced dynamic voltage and frequency scaling extensions core (eDVFSC) techniques for more efficient power management.
Together with the change in geometry, these contribute to over 20 percent power savings, says Moore, "just going from 1-alpha to 1-beta." He adds that this will vary by application. When running at higher speeds, "the power savings are closer to 30 percent."
"If you're capturing video content for Tiktok, which requires those higher bandwidth speeds, you are going to notice an increase in your smartphones battery life." he says. The same applies with night mode shots which require multiple frames of the same scene stitched together to create the picture perfect low light image.
But Moore adds it's important to not consider DRAM in isolation. The technology has been developed in close cooperation with the manufacturers who will be integrating it into their own components, whether it be a system on chip (SoC) or finished smartphones.
"We as a team are really focused on the mobile phone system, the entire system, not one individual component," Moore says. This has allowed Micron to gain a deeper understanding of "how that component works with the SoC and as a whole can impact battery life."
He cites the example of the development of LPDDR5. "It came out with lower power at the component level, but what we found was actually running the higher speeds of the LPDDR5, as opposed to LPDDR4X, allows the SOC to go into standby mode much faster because you're transferring data that much faster." This resulted in "really big power savings that are noticeable'' on battery life.
Denser, faster memory potentially boosts all applications. While there is much talk about the 5G killer app, Moore questions whether this is the right way to look at progress in smartphones. "That's not the way 3G and 4G worked. It is really more incremental."
However, 1-beta should help some nascent applications and form factors really take off, such as foldable phones. Gamers might already love these sorts of devices, says Moore, and "one thing you're going to see out of LPDRAM, just like you would in a PC or a gaming console, is that faster and more DRAM is going to give you a richer, more efficient end user experience."
One of the nice things about foldable phones is you can have multiple apps open simultaneously, which is where the amount of DRAM really becomes critical. And increasingly, these are likely to be enterprise apps. While this is a usage model that is just in its infancy today, it's a space where the amount of DRAM and the speed of that DRAM is going to really matter in the future, predicts Moore.
A foldable world
Similarly, AI and contextual awareness applications – whether on device or at the backend – will become more of a priority over time. 5G connectivity will be a major enabler here, but these models also assume more data being moved around and to and from the device.
"You need super-fast storage and massive capacity to be able to process and store the influx of rich AI data that is derived from these new applications" explains Moore.
Ultimately, the hardware needs to get out of the way of the software. "So what we focus on is making sure that we are enabling these technologies that enhance the end user experience," says Moore.
"These advanced use cases demand higher capacity smartphones, 12GB to 16GB, and in the future a potential for a demand for 24GB. Contextual awareness requires real time processing and space. We consider these next-generation use cases when developing our technology, specifically, prioritizing increasing the capacity and speed potential to allow for the bandwidth that these advanced applications demand to operate at their full potential."
Parts using Micron's 1-beta node are already being sampled and qualified by OEMs. By the second quarter of this year, this node will ship in very high volumes. And while the initial focus will be on high end and flagship smartphones, the technology will inevitably cascade to the mid and lower tier devices over time.
A high-density, lower power memory architecture also has an obvious appeal for Internet of Things (IoT) devices, thin and light laptops, and other applications. And after a tough couple of years for the car industry as manufacturers scrabbled to source enough silicon, this is a vertical that could be on the cusp of a rapid expansion which will see 1-beta "advance in the automotive side in a very impactful way" predicts Moore.
More broadly, it will find its way into traditional DDR and graphics applications. When so much energy is consumed in the datacenter, anything that reduces power draw and footprint will be of interest. "In terms of doing good for the world, the green side, these newer nodes definitely help on that as well."
A few decades ago it was customary to be amazed at the prospect of a smartphone that could deliver the power of a computer in the palm of your hand. Now perhaps we can look forward to talking about having the benefits of a smartphone in the enterprise.
Sponsored by Micron.
From Chip War To Cloud War: The Next Frontier In Global Tech Competition
The global chip war, characterized by intense competition among nations and corporations for supremacy in semiconductor ... Read more
The High Stakes Of Tech Regulation: Security Risks And Market Dynamics
The influence of tech giants in the global economy continues to grow, raising crucial questions about how to balance sec... Read more
The Tyranny Of Instagram Interiors: Why It's Time To Break Free From Algorithm-Driven Aesthetics
Instagram has become a dominant force in shaping interior design trends, offering a seemingly endless stream of inspirat... Read more
The Data Crunch In AI: Strategies For Sustainability
Exploring solutions to the imminent exhaustion of internet data for AI training.As the artificial intelligence (AI) indu... Read more
Google Abandons Four-Year Effort To Remove Cookies From Chrome Browser
After four years of dedicated effort, Google has decided to abandon its plan to remove third-party cookies from its Chro... Read more
LinkedIn Embraces AI And Gamification To Drive User Engagement And Revenue
In an effort to tackle slowing revenue growth and enhance user engagement, LinkedIn is turning to artificial intelligenc... Read more