Thanks for the quick response. I'll be happy to test this when it's released. I confirmed this bug on another 440BX by a different manufacturer as well.
What did the debug information show that helped identify the issue?
It has shown that the problem was in detection of the SIO chip. If HWiNFO detected the Winbond chip but left it's config space unlocked, a new start and check for another SIO at the same I/O location would cause the issue. I believe that putting the SIO back into locked state after detection will solve this problem. That's what the next build of HWiNFO will do.