Ever onward

Arjan Van de Ven hipped me to a BIOS bug that can be mitigated by passing idle=poll to the kernel at boot. There is a performance hit, but the RedHat 2.4.21ELsmp kernel now boots on the 2xOpteron 248 box. MySQL is also happier. The crux of the issue, of course, was the inaccessibility of the RTC. The MySQL bench tests show the 2.2Ghz 248 system well ahead of a dual 3.06Ghz HT XEON and a dual 1.6Ghz Itanium2 syst

Arjan Van de Ven hipped me to a BIOS bug that can be mitigated by passing idle=poll to the kernel at boot. There is a performance hit, but the RedHat 2.4.21ELsmp kernel now boots on the 2xOpteron 248 box. MySQL is also happier. The crux of the issue, of course, was the inaccessibility of the RTC. The MySQL bench tests show the 2.2Ghz 248 system well ahead of a dual 3.06Ghz HT XEON and a dual 1.6Ghz Itanium2 system. Well ahead. Now if I could just get 2.6.0-test10 to boot...

Recommended
Join the discussion
Be the first to comment on this article. Our Commenting Policies