AIoT fragmented applications and computing power create new opportunities, GigaDevice diversifies storage layout

Publisher:EE小广播Latest update time:2021-08-13 Source: EEWORLDKeywords:AIoT Reading articles on mobile phones Scan QR code
Read articles on your mobile phone anytime, anywhere

AIoT fragmented applications and computing power bring new opportunities, revealing the logic behind GigaDevice's diversified storage layout


In the human brain, the hippocampus is responsible for important memory-related functions. Similarly, in electronic systems, the role of the "hippocampus" is played by memory. As an indispensable component of semiconductor components, memory is born with data and is the most widely used and market-leading basic integrated circuit product. Data shows that global memory revenue increased by US$13.5 billion in 2020, accounting for 44% of the growth in revenue of the entire semiconductor market in 2020.


image.png

Figure 1: Global semiconductor revenue growth structure distribution in 2020 (Foresight Industry Research Institute)


In recent years, technologies such as AI and IoT have driven the rapid development of consumer electronics and large-scale data centers, and the market demand for storage has become increasingly fierce. Especially in the field of IoT, the power consumption and size of products have always been the direction in which the industry is constantly seeking breakthroughs. For memory, key features such as low power consumption, small size, and high operating speed have gradually become necessary conditions for entering IoT applications.


AIoT's diverse new demands and edge computing power drive market growth and technological evolution of various memory types


In the era of the Internet of Everything, data flows continuously; at the same time, the improvement of computing power has promoted the change of computing structure - from the end-to-cloud in the past to the cloud-to-end now. In this process, not only the IoT devices are required to have network connection capabilities, but also some basic computing capabilities, that is, the computing power of the main chip, which also means that the algorithms on the end side are larger and more complex, and the memory capacity for storing algorithm codes also increases accordingly.


"NOR Flash has excellent characteristics such as random storage, strong reliability, fast reading speed, and executable code. It is expected to replace NAND Flash and DRAM in IoT devices. The processor/main chip can directly call and run system code from NOR Flash, which can meet the requirements of storage and running memory at the same time. This makes the market demand for NOR Flash increase year by year, and the annual growth rate is expected to exceed 10%." said Chen Hui, senior product marketing director of GigaDevice's storage business unit.


Obviously, the explosive growth of the Internet of Things and related AIoT products (TWS, wearables) has become the main driving force for the growth of NOR Flash memory demand. Relevant data show that the global NOR Flash market size in 2020 is around US$3.06 billion, and will further grow to US$3.72 billion by 2022.


According to Chen Hui, as the main product in the storage field, GigaDevice SPI NOR Flash already has 26 product lines, 16 product capacities, 7 temperature specifications, 4 voltage ranges and 25 packaging methods. Data shows that since launching China's first SPI NOR Flash in 2008, GigaDevice has become the world's third and China's first SPI NOR Flash supplier, and the cumulative shipments of Flash products have exceeded 16 billion!


image.png

Figure 2: GigaDevice launches comprehensive NOR Flash products and solutions


Behind the dazzling achievements is GigaDevice's determination to constantly seek breakthroughs. Chen Hui said: "GigaDevice's strategy has always been closely linked to market dynamics, relying on product definitions and features to launch products with different performances to meet the diverse needs of customers, gradually covering industrial, consumer, and even automotive fields." At present, GigaDevice's 55nm advanced process technology SPI NOR Flash has entered the full-line mass production stage, with higher performance, richer capacity, and lower power consumption than the previous generation of products. For the automotive market, automotive-grade 3.0v and 1.8v SPI NOR Flash products have also been launched, with capacities ranging from 2Mb to 2Gb to meet the needs of different in-vehicle applications.


In addition to NOR Flash, SLC NAND Flash and DRAM are also the storage product directions that GigaDevice focuses on, and continues to increase investment. In October 2020, GigaDevice officially launched the GD5F4GM5 series, a 4Gb SPI NAND Flash product with a nationally produced 24nm process node. This series has achieved pure localization and independence in all links from design and development, production and manufacturing to packaging and testing, and has been successfully mass-produced. In addition, in terms of DRAM memory chip layout, in June 2021, GigaDevice announced the mass production of its first self-branded 4Gb DDR4 product, the GDQ2BFAA series. This product has achieved national localization from design, tape-out, to packaging and testing, and verification, marking that GigaDevice has successfully expanded its business tentacles to the mainstream DRAM storage market.


image.png

Figure 3: GigaDevice's 4Gb SPI NAND Flash product - GD5F4GM5 series


Chasing size, power consumption, safety, and performance? Superb product definition is an art


The booming Internet of Things has injected a shot in the arm for the NOR Flash market, but the diversity and complexity of IoT application scenarios have also brought new challenges to NOR Flash. For example, although small package size and fewer signal leads are the main advantages of SPI NOR Flash, it is imperative to further reduce the product package size under the requirement of increasingly miniaturized IoT devices; most IoT applications may even take more than 10 years in some cases due to geographical location or cost reasons, so power consumption is also one of the most important indicators of IoT terminals or modules. In addition, aspects such as capacity and security are also challenging topics that manufacturers have to face.


Wearable devices, which have become increasingly popular in recent years, need to collect user data and information to provide intelligent suggestions, such as location information and heart rate information, and transmit external data to users, such as music, video and calls, etc. These processes cannot be separated from the support of NOR Flash. In the case of "every inch of land is valuable" internal space, its electronic components need to meet the requirements of small size and low power consumption.


For example, as TWS headphones enter the 2.0 era, with the emergence of multiple functions including OTA upgrades, bone conduction, voice recognition, noise reduction, touch, etc., in order to store more firmware and code programs, it is often necessary to expand an external 32Mbit or even larger SPI NOR Flash, while also requiring small size and low power consumption.


As a response, GigaDevice announced earlier this year that it would introduce WLCSP packaging into its own storage devices. Compared with other packages, GigaDevice's storage devices with the same capacity can be as small as less than 1mm in length and width, and as thin as 0.25mm, which is about 1/10 of the USON8 (3mm x 2mm) package volume. For TWS headset products, GigaDevice can also stack Flash and the main chip according to customer needs, without having to open a separate space for Flash, which greatly helps reduce the area of ​​the entire system PCB.


image.png

Figure 4: WLCSP package comparison with other packages


image.png

Figure 5: WLCSP package vs. USON8 package, the volume is about one-tenth of the latter


In addition, GigaDevice has also launched the industry's lowest read power consumption LE series SPI NOR Flash, with an operating voltage of 1.65~2.0V and a product capacity of 2Mb~512Mb, which can meet different application requirements. In terms of power consumption, the sleep current of the LE series is as low as 0.1uA, and the read power consumption of the four-channel 133Hz is as low as 5mA, which is about 40% lower than the industry average.


image.png

Figure 6: GigaDevice provides a wide range of capacity options for the same package size


Security is another major consideration in memory design and application. In IoT applications, Flash is more vulnerable to attack than SoC because it stores many critical system codes, so improving security is particularly important. "To improve system security, we cannot just start with Flash security specifications or encryption technology. We need to consider Flash and main chips, such as MCU and SoC, as a whole. Although the higher the security level, the better, we consider that the security function of Flash should be matched with the terminal product positioning and the main chip positioning, so that we can make a balanced product that meets market demand." Chen Hui said, "Of course, GigaDevice Flash is combining its own MCU products or working closely with more SoC manufacturers to improve the security performance of IoT devices and systems."


Storage, computing and sensing are integrated, and the synergy effect of GigaDevice's product ecosystem is beginning to emerge


In fact, the potential of the NOR Flash market has been fully activated, driven not only by the Internet of Things, but also by the demands of automotive electronics, 5G, smart phones, and TDDI integrating touch functions into driver ICs. "At present, GigaDevice has reached cooperation with the world's top communication equipment manufacturers, and its larger capacity 512Mb-2Gb products for the automotive market have also passed the AEC-Q100 certification." Chen Hui said that with the increase in user demand, the update of system functions, and the increase in code complexity, GigaDevice's Flash products will continue to evolve in the direction of large capacity, high performance, high reliability, low power consumption, and small packaging in the future.

[1] [2]
Keywords:AIoT Reference address:AIoT fragmented applications and computing power create new opportunities, GigaDevice diversifies storage layout

Previous article:BlackBerry launches BlackBerry Jarvis2 to address new global embedded network security challenges
Next article:Intel AI consultants travel across the stars to protect astronauts' health

Recommended ReadingLatest update time:2024-11-16 13:38

With AI in government services, things can become simple and efficient
In the past, going to government agencies to do business was a headache. Not only did one have to prepare complicated documents, but there were also many troubles such as "running back and forth to do business, the business location was far away, and the business points were scattered", which increased the difficulty
[Embedded]
With AI in government services, things can become simple and efficient
State Grid Corporation of China's ubiquitous power Internet of Things construction in November
As of the end of November, the construction of the ubiquitous power Internet of Things of State Grid Corporation was generally carried out as planned. The outline of the key construction tasks of the ubiquitous power Internet of Things in 2020 was issued to guide various units to arrange next year's work and special
[New Energy]
Renesas Electronics’ PMIC and GreenPA combination powers Kria SOM vision AI platform
Renesas Electronics (Dialog) has announced that it has expanded its collaboration with Xilinx, a leader in adaptive computing, and has been selected to provide power management companion chips for Xilinx's new Kria Adaptive System-on-Module (SOM), which targets visual AI applications in smart cities and factories.
[Power Management]
Amazon Cloud Technology’s artificial intelligence and machine learning technology helps scientists draw complete brain maps
Despite decades of research by neuroscientists, advances in treatments for brain diseases such as Alzheimer's and Parkinson's have been slow because the brain is the most complex and least understood organ in the human body. Data sets from brain research, while numerous, are discrete and diverse, and often cannot be d
[Medical Electronics]
Mapbox launches conversational AI service MapGPT to enable automakers to control voice assistants
According to foreign media reports, Mapbox, a map and positioning platform provider that powers a new generation of location-aware applications, has announced the launch of a new conversational artificial intelligence service, MapGPT, that will allow automakers to create customized voice assistants that can be used on
[Automotive Electronics]
Mapbox launches conversational AI service MapGPT to enable automakers to control voice assistants
OPPO Find X8 launches the first AI one-click screen query function
OPPO Find X8 is the first to launch the new AI one-click screen inquiry function, creating a mobile AI that everyone can use through leading multimodal interaction. Summary generated by OPPO AI On October 18, OPPO announced the new AI feature of the Find X8 series - "AI One-Click Screen I
[Mobile phone portable]
OPPO Find X8 launches the first AI one-click screen query function
Xiaomi Group's first quarter 2020 results released: A strong comeback despite adversity
Xiaomi Group released its first quarter 2020 financial report: total revenue reached RMB 49.7 billion, a year-on-year increase of 13.6%; adjusted net profit was RMB 2.3 billion, a year-on-year increase of 10.6%. The degree of globalization continues to deepen. Overseas market revenue was RMB 24.8 billion, accounting f
[Embedded]
Xiaomi Group's first quarter 2020 results released: A strong comeback despite adversity
JFrog acquires Qwak AI to simplify the entire process of AI model development to production
JFrog will continue to expand its solution to bring advanced MLOps capabilities to enterprises, enabling them to build, deploy, manage, and monitor AI workflows including generative AI, large language models (LLMs), and general machine learning models on a unified platform. July 1, 2024 – JFrog,
[Network Communication]
Latest Embedded Articles
Change More Related Popular Components

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号