[Beowulf] Liinpack benchmark

[feldy at acm.org]

On Mon, 14 Nov 2005, Robert G. Brown wrote:

> I'm looking to see if one can WRITE to the TSC post boot -- if one can
> then it might be a fix for the kernel team -- install an ntp-like
> handshaking/sync process at a controlled point in the boot so that all
> CPUs emerge from a linux boot with their counters in sync.  Assuming the


I believe this linux already does this (meaning sync the TSC), at least using 
FC4 on an x86_64 machines (intel). Here's a snippet of the
dmesg from a dual processor system with hyperthreading turned off)

CPU 1: Syncing TSC to CPU 0.
Brought up 2 CPUs
time.c: Using PIT/TSC based timekeeping.
testing NMI watchdog ... OK.
checking if image is initramfs...
CPU 1: synchronized TSC with CPU 0 (last diff 52 cycles, maxerr 1207 cycles)

and with hyperthreading turned on
CPU 1: synchronized TSC with CPU 0 (last diff -23 cycles, maxerr 1290 cycles)
CPU 2: synchronized TSC with CPU 0 (last diff -4 cycles, maxerr 915 cycles)
CPU 3: synchronized TSC with CPU 0 (last diff -19 cycles, maxerr 1290 cycles)