Introduction

Consumer demand for smart electronics functionality such as smartphones, consumer electronics, and in-car entertainment systems is endless.

As such, the choice of storage device has been, and will continue to be, very important for designers.

These systems have become very complex over the past decade, and designers are demanding higher capacity, faster non-volatile memory to store the code that will deliver the next generation of innovative functions and features. On the other hand, however, the challenge is to achieve high capacity without increasing costs. Consumers want more features and functionality, but they don't want to pay more for it.

In short, the challenge for non-volatile memory manufacturers is to be able to deliver higher densities and faster speeds without significantly increasing costs for both themselves and their users, providing consumers with a better user experience. I'm here.

Flash memory market and applications for embedded systems

All electronic systems now use some form of flash memory, and the emergence of innovative applications has dramatically increased the demand for higher performance products. In particular, the markets and applications listed below are rapidly evolving and demanding “modern advanced memory subsystems” capable of faster read/program/erase operations and instant boot.

・ In-vehicle: Advanced Driver Assistance Systems (ADAS), instrument cluster, push buttons
・ artificial intelligence: Training model storage, database storage, autonomous driving
・ OTA (Over-the-Air): Consumer, Automotive, Gaming

In the 'automotive field', throughput is very important when choosing flash memory.

The instrument cluster in the digital display is designed to ensure that the driver is always comprehensively informed.

These displays must power on instantly and display 2D/3D images quickly and accurately.

Nowadays, it is common to use serial NOR flash (Octal SPI or Quad SPI) or serial NOR flash with eMMC for instrument clusters. However, these functions and features are becoming more and more sophisticated, and the increase in code size requires higher capacity and increases cost.

Also, even if the code size increases, the startup time should be the same or even faster than before.

This is very important in the automotive market for features like pushbuttons that need to display multiple screens instantly when the user presses a button.

“OTA applications” for code updates are used in electric vehicles and many consumer devices due to their flexibility and ease of use. All of these devices are mobile and need to receive important updates easily and efficiently. Therefore, fast erasing/programming is required, and speed is a key requirement.

Other applications that require high-speed operation include AI and facial recognition.

An "AI application" generally needs to run different training models to make decisions and tell the end device what to do. For that, training models need to be stored inside the system, and in many cases these are very large models that need to be processed.

Especially for autonomous driving, decisions must be made in a very short time (milliseconds), and these models must be loaded into RAM from very high-throughput flash memory, as well as pedestrian and vehicle avoidance. and so on, an accurate judgment is required. Data throughput is absolutely critical here, and it has to be instantaneous.

In the next article, I will explain NOR flash memory, which is approaching its limit, and NAND flash memory, which is still evolving, so please take a look.

Product introduction

Click here for product introduction.

Click here for related links

Inquiry

If you have any questions regarding this content, please contact us using the form below.

Back to Manufacturer Information Top

If you want to return to Winbond manufacturer information top page, please click below.