Thu Sep 26 14:00:00 UTC 2024: ## Researchers Optimize Golomb-Rice Coding for Efficient Data Compression in VLSI Testing

**Vellore, India (September 26, 2024) -** A team of researchers from multiple institutions in India, Egypt, Saudi Arabia, and Jordan have developed new hardware optimizations for Golomb-Rice (GR) coding, a widely used data compression technique. Their findings, published in PLOS ONE, demonstrate significant improvements in bit usage, area, delay, and power consumption, making GR coding even more effective for applications like VLSI testing, biomedical image compression, and data transmission in AI/IoT systems.

The study focused on optimizing the “unary part” of the GR code, which represents the frequency of a symbol in a data set. The researchers proposed three new schemes, HSGRC, LPCGRC, and EBRGRC, that reduce the number of bits used for unary encoding, leading to a more efficient compression.

Their results show a substantial reduction in bit usage, ranging from 6% to 19% for linearly distributed data sets. This, in turn, leads to improvements in other crucial metrics, such as:

* **Delay:** Reduction of 3% to 8% in worst-case delay.
* **Area:** Reduction of 22% to 36% in area usage.
* **Power:** Reduction of up to 7% in switching power.

The researchers have also provided a detailed comparison of their proposed methods with the original GR encoding scheme and other existing state-of-the-art compression techniques. Their optimized design demonstrates a significant advantage in terms of area, power, and speed, making it a promising solution for various data compression applications.

**The study highlights the following key findings:**

* GR coding offers a high compression ratio, particularly for data sets with a geometric distribution.
* Optimizing the unary part of the GR code can significantly enhance compression efficiency.
* The researchers’ proposed schemes demonstrate substantial improvements in various design metrics, making GR coding more suitable for resource-constrained applications.

**The researchers believe their work can contribute to:**

* More efficient VLSI testing, reducing testing time and power consumption.
* Improved biomedical image compression and transmission, facilitating faster and more efficient analysis.
* Enhanced data transmission in AI/IoT systems, enabling more efficient and reliable data sharing.

The research team will continue to explore further optimizations for GR coding, aiming to develop even more efficient and versatile data compression solutions for various applications.

Read More