The older I get, the dumber I get. I picked on Robert for asserting
that there are physical disconnects between the Red and Black
partitions. I shouldn't have done that.
in <http://www.lanl.gov/orgs/ccn/salishan2003/pdf/camp.pdf>,
Slide 68 clearly shows the four places where the disconnect can be
located. Only one disconnect is used at one time. A partition can be
assigned 0%, ~25%, ~75%, or 100% of the CPUs, with the other partition
taking what's left.
Slide 67 is less clear, but more specific: The breakpoints are at
2688, 4992, and 2688 CPUs (plus 0).
In other words, the location is not generally programmable like the
size of a FAT16 hard drive partition, which I had incorrectly assumed
was the case.
|
Similar Threads
