What is Memory

Why can’t we Scale memory chips?

Memory chips have a harder time when they get smaller because they have to keep their tiny boxes (cells) from getting mixed up and confused.
Share this STORY


Imagine you have a bunch of tiny boxes, and you want to use these boxes to store information. These boxes represent the cells in memory chips.

Memory chips, like the ones used in your computer or phone, are really good at storing lots of information because they have many of these tiny boxes, or cells, to hold data like pictures, videos, and apps.

Now, let’s say you want to make these boxes even smaller so you can fit more of them in the same space.

This is like trying to fit more books on a bookshelf by making the shelves smaller. When you make the boxes smaller, each box can hold less information, but you can fit more boxes overall.

This is what we mean by “scaling down” the size of the memory cells.

However, when you make these boxes really small, they become a bit finicky. They might start listening to their neighbors and get confused about what information they’re supposed to hold.

It’s like if you and your friend are whispering secrets to each other, and someone else nearby can hear and mix up the secrets.

These tiny boxes can start to “whisper” to each other because they’re so close together, and this can make it harder to keep the right information in each box.

Now, let’s talk about logic chips. These are like the brain of your computer or phone. They’re in charge of doing calculations, making decisions, and telling the memory chips what to do.

Logic chips use simple switches, like light switches, to perform calculations and make things happen. These switches are a bit like tiny traffic signals inside the chip.

When you scale down logic chips, you’re making these switches smaller, which means they can work faster and use less energy.

It’s like having smaller and more efficient traffic signals that control the flow of information inside the chip.

Because logic chips are mainly about making decisions and calculations, they don’t need to worry as much about getting confused by their neighbors like the memory cells do.

So, in simple terms, memory chips have a harder time when they get smaller because they have to keep their tiny boxes (cells) from getting mixed up and confused.

Logic chips, on the other hand, use simple switches to make decisions and don’t need to deal with this confusion problem as much. That’s why logic chips have been able to scale down more easily compared to memory chips.

Also Read: Explained: What the hell is memory?

Why Memories don’t scale as much as logic?

Memory chips and logic chips have different characteristics and challenges when it comes to scaling down their sizes and increasing their performance.

While both types of chips have benefited from advancements in semiconductor technology, there are specific reasons why memory chips have not scaled as easily as logic chips.

Density vs. Complexity: Memory chips, like DRAM (Dynamic Random-Access Memory) and NAND Flash, are designed for high-density data storage.

They require a large number of cells to store data, and each cell needs to be reliable and consistent in its behavior. Logic chips, on the other hand, are designed for processing information and performing computations.

While both types of chips benefit from scaling, the requirements for maintaining consistent behavior and reliability are more demanding in memory cells, making the scaling process more challenging.

Cell Integrity: As memory cells are scaled down, they become smaller and more susceptible to various forms of interference and defects.

For example, as cells get smaller, they can become more sensitive to electrical noise, cross-talk between neighboring cells, and other issues that can affect their ability to store and retrieve data accurately.

Ensuring the integrity and reliability of memory cells becomes increasingly complex as their size is reduced.

Read and Write Operations: Memory cells need to be read from and written to frequently, which introduces wear and tear on the cells.

As memory cells are scaled down, the physical effects of these operations become more pronounced, potentially reducing the overall lifespan of the memory chip.

Manufacturers need to develop new techniques and materials to mitigate these effects.

Technological Complexity: Memory cells, especially NAND Flash cells, are more complex in terms of how data is stored and retrieved compared to logic gates used in processors.

This complexity can lead to challenges when it comes to scaling down the cells while maintaining the necessary levels of reliability and performance.

Economic Factors: The cost of developing new manufacturing processes and materials for scaling memory cells can be significant.

Manufacturers need to weigh the benefits of scaling against the costs and challenges associated with producing viable, high-performance memory chips.

Applications and Use Cases: Memory and logic chips serve different purposes in electronic devices. Logic chips are critical for computations and processing, which drives demand for faster and more efficient chips.

As a result, there is a strong incentive to invest in scaling logic chips. Memory chips, while important for storage, face different usage patterns and demand characteristics, which may influence the priorities for scaling.


It’s important to note that while memory chips may not have scaled as dramatically as logic chips, there have still been significant advancements in memory technology.

Manufacturers have introduced new types of memory like 3D NAND Flash, which stacks memory cells vertically to increase storage density, and new non-volatile memory technologies like PCM (Phase Change Memory) and MRAM (Magneto-Resistive RAM) that offer different trade-offs between speed, density, and endurance compared to traditional memory technologies like DRAM and NAND Flash.

Share this STORY