News

Researchers Develop Bug-blocking Chip Monitor

Researchers at the University of Michigan have developed technology that can keep microprocessor bugs from seizing up a PC

By Robert McMillan, IDG News Service (San Francisco Bureau)

September 30, 2008 — CSO —

Researchers at the University of Michigan have developed technology that can fence off microprocessor bugs and keep them from seizing up a PC.

For the past two years they've been working on what they call a "semantic guardian." It's a tiny monitor that lives on the microprocessor, checking it to see if the chip is being asked to do something that its designers hadn't predicted in their quality assurance testing.

Companies such as Intel and Advanced Micro Devices test their products rigorously, but in the real world chips are often asked to do new things that could potentially cause a crash. The semantic guardian can identify these untested states and then slow down the processor by kicking it into a safe mode, where many of the chip's performance-enhancing bells and whistles are disabled.

This slows the computer down for a tiny fraction of a second, but it also makes it much less likely to crash, said Valeria Bertacco, an assistant professor with the university's computer science department.

Her team has built the semantic guardian into a software-based chip simulator and is now trying to get it up and running on a programmable microchip called a field-programmable gate array. In their current design, the monitor takes up about 3 percent of the chip's real estate, but they expect that it would be much smaller if ever developed commercially. "If any commercial company decides to do this it would be much less than 1 percent," Bertacco said.

Chip makers often end up discovering hundreds of bugs, called errata by the industry, in their products after they ship. These bugs can be addressed in BIOS updates, but if a flaw is serious enough, it can derail a company's product plans. Last year, AMD's quad-core Opteron processor, code-named Barcelona, was delayed by nearly half a year when a serious flaw was discovered after the chip's official launch.

Insight 64 analyst Nathan Brookwood is unconvinced that a semantic guardian would have helped AMD with its Barcelona problem. According to him, there are at least two big problems with this approach: First, it would be hard to keep track of all the tested states on a commercial processor. "There are a very large number of legitimate states, so I really question whether this is anything that could ever be made to be a practical solution," he said.

A second problem is that it would take a lot of design work to create a slimmed-down processor safe mode that really worked.