Skip to content

EPFL Team's Discovery Could Revolutionize FPGA Programming

A groundbreaking discovery could make FPGAs, crucial in telecoms, automotive, and aerospace, more efficient. The team's work won the Best Paper Award at the 33rd IEEE International Symposium on Field-Programmable Custom Computing Machines.

In this image there is a board, on that board there is some text, on the left side there are...
In this image there is a board, on that board there is some text, on the left side there are vehicles on a road, on the right side there is a footpath and buildings.

EPFL Team's Discovery Could Revolutionize FPGA Programming

Researchers from EPFL, AMD, and the University of Novi Sad have made a significant discovery that could reshape how millions of Field-Programmable Gate Arrays (FPGAs) are programmed. The team, led by Professor Davide Rossi, has identified an inefficiency in the PathFinder algorithm, a standard used since the late 1990s.

FPGAs are widely used in various industries such as telecoms, automotive, aerospace, and particle physics due to their reconfigurability. However, the PathFinder algorithm, which connects thousands of tiny circuit components in FPGAs without creating overlaps, has been found to often build routing trees larger than necessary. This increases the risk of overlaps and can lead to slowdowns and unroutable designs as circuits grow larger.

The research team discovered that the order in which PathFinder creates and adds new branches can cause these failures. The team's initial solution involved trying different orders and picking the one that results in the smallest tree. Their findings, published in a paper titled 'Efficient Routing for Large FPGA Designs', received the Best Paper Award at the 33rd IEEE International Symposium on Field-Programmable Custom Computing Machines.

This discovery could influence the design of future generations of these reconfigurable chips, making them more efficient and capable of handling larger and more complex signal paths. The team's work is a significant step forward in improving the programming of FPGAs, which are crucial components in many modern technologies.

Read also:

Latest