Re: x86 clock speed patch -- Possible bug?
Stephan Meyer (firstname.lastname@example.org)
Wed, 5 Feb 1997 13:50:27 +0100 (MET)
On Sun, 2 Feb 1997, Nicholas J. Leon wrote:
> On Sun, 2 Feb 1997, Stephan Meyer wrote:
> # On Sat, 1 Feb 1997 email@example.com wrote:
> # > I have a P133. When I first started, the /proc/cpuinfo said that I had a
> # > 133MHz processor. About an hour later, I look at /proc/cpuinfo from my
> # > user account, and get this:
> # > clock_speed : 132 MHz
> # > That happened several times. Then it reverted back to:
> # > clock_speed : 133 MHz
> # This is definitely not a bug. On my machine, I get 119 instead of 120
> # about every 20th time I read /proc/cpuinfo.
> # Does your harddrive run because of some other process while doing this?
> Not necessarily. The true speed is 66(external clock) x 2(multiplier) for
> a total of 132. Or, if you prefer, 66.66Mhz x 2 (133.33).
> It's common for programs to miscalculate the actual internal speed of a
> processor. It's one of the reasons I'm surprised that it made it into the
> kernel (being so hard to do and mostly innacurate). Even ctcm (a quality
> low-level benchmark util) fluctuates between 131 and 134 for a P133.
I don't really get the point of the problem...
Is getting 131 of 134 that bad? Will someone sue the hardware
manufacturer for it?
"being so hard to do" - it is not hard to do. Have a look at the bogomips
code and consider again.