While testing an AMD Barcelona chip, Tom Yager found that he could incur variances in benchmark tests ranging from 10 to 60 percent through combined manipulation of the server’s BIOS settings, BIOS version, compiler flags, and OS release — then the whole thing ran aground after he’d fiddled with settings and flags so much that the testbed stabilized. “I was told that I was seeing an effect that’s widely known among CPU engineers, but seldom communicated to IT,” Yager writes in Your servers are wasting your CPUs. Most times, trying to outthink AMD’s engineers is a fruitless exercise, but system and OS makers do it all the time, Yager adds. “When they get it wrong — and this is far easier to do than getting it right — it costs you. You end up with systems that aren’t performing to their potential, are letting power efficiency features go unexploited, or both.” Technology Industry