Read and Execute: Prevent System Bottlenecks with RWW

Read and Execute: Prevent System Bottlenecks with RWW

Moving intelligence to the edge can lead to systems with better real-time performance, better power efficiency and enhanced security. But more intelligence requires more code, more data, and the ability to execute commands without delay or latency, even during system firmware updates.

Whether performing high-speed data transfer or executing code directly out of memory, a key system design requirement for edge devices is the selection of the right non-volatile memory.

This becomes apparent when performing a write operation. Write operations are required to install new software, change configuration settings and security keys, update a video data file or log frequently changing data points. When a standard flash memory device is busy with a write operation under any of these conditions, the host controller is unable to read, or in other words retrieve, data or instructions. If the system controller is prevented from retrieving and executing the next instruction, this can throttle the overall system performance.

How can your system effortlessly perform uninterrupted read operations and continue normal operation even during a system code or data file update, or while performing a data-logging operation?

A novel feature for flash memory is Read While Write (RWW), which allows reads and writes to execute concurrently. Without integrated RWW, multiple devices and CPU resources are required to achieve the same functionality, leading to increased system complexity, cost and power. RWW makes it easier to meet strict real-time functionality, and it reduces latency and improves throughput.

View our TechShop video for a visual demonstration of the performance advantages of RWW.

In the demonstration, you can see two systems using execute in place (XiP) to play back a video file while simultaneously receiving a request to write data to the memory, as in the case of a system or application update.

XiP is a proven-effective configuration to handle a range of emerging applications. Because the flash memory in an XiP architecture is external to the chip, the external memory can be as large as required to efficiently process even large code sets.

In the video, one system is using the RWW feature available on Adesto’s EcoXiP NOR flash memory so it can play back the video using XiP all the time – even while it receives requests to write to the memory. The other system is not using RWW.

You can see the difference in the quality of the playback. Without RWW, the system cannot perform XiP 100% of the time while it receives the same requests to write to the memory.

Advantages of RWW include:

  • Allows over-the-air (OTA) updates without interrupting normal operation
  • Simplifies software design and system architecture
  • Reduces system cost and power compared to dual chip / dual die solutions

To learn more about Read While Write on Adesto’s EcoXiP, download the product brief.

To learn more about EcoXiP, to order samples or get an EVK board, visit

For a complete guide to choosing a memory for AI devices, download the whitepaper “AI on the Edge: Don’t forget the memories”.