On Wed, Aug 19, 2009 at 07:19:02PM -0700, john stultz wrote:

On Wed, Aug 19, 2009 at 2:48 PM, Thomas Gleixner[off-list ref] wrote:

quoted

We are pleased to announce the next update to our new preempt-rt
series.

Hey Thomas, Peter, Ingo,
    I'm still seeing the odd SMP scheduler behavior (where all the
processes line up on a single cpu) with -rt5. This can be easily
reproduced with a "make -j8 bzImage".  I'm I the only one seeing this?

I am seeing it with 2.6.31-rcX-rtY kernels on both a 8 CPU AMD/Intel box
Specjbb was returning pathetically low numbers which turns out to be
because all threads are running on just one of the cpus

	-Dinakar

 Any suggestions on further debugging this?

thanks
-john