One more test of 2.5.62-mm1 on a big memory system w/qlogic FC
controller.
JennyZ has run OSDL's dbt3 workload on an 8-CPU system CONFIG_HIGHMEM4G=y
(she's out today, so i'm Mr.Hands)
Dbt3 is a decision support workload.
"DBT3 power
load has some disk write and the beginning and end of
the run, most disk activity is read. I did not see any
problem during the run.
The file calc_power.out records the mesurement. The
higher the power, the better the performance. "
Results for 2.5.62-mm1 at (vmstat data in file vmstat.out)
http://www.osdl.org/archive/jenny/2.5.62-mm1
Run of same test on 2.4.18-rh kernel
http://www.osdl.org/archive/jenny/2.4.18-redhat
cliffw