Guidelines for Configuring Virtual Partitions on Cellular Platforms

6
During a SPECjbb run, the benchmark is run repeatedly with an increasing number of
warehouses starting with a specified number and ending with twice that number. The
average throughput of all the runs is used as the performance measure. The unit of measure,
BOPS, stands for business operations per second.
For this focused study, the benchmark was configured with 16 warehouses. Given a set of
resources a virtual partition configuration having the best resource locality should result in
highest number of BOPS.
Hardware and Software Environment
An HP Integrity rx7640 Server with two cells was used for this experiment. Each cell was
populated with 32 GB memory and four dual-core Intel Itanium 2 CPUs (8 processing cores).
The rx7640 is a single crossbar system. To enable for workload performance comparisons
with a single cell nPartition configuration this system was configured to be capable of
hosting either:
• Two virtual partitions (HP-UX 11i v3 with vPars A.05.01) in a two-cell nPartition or
• Two single-cell nPartitions each capable of hosting one virtual partition.
The I/O associated with each workload was attached to each cell: the I/O for SAP SD 2-tier
was connected to cell 0, and the I/O for SPECjbb was connected to cell 1.
Design of Experiments
The main objective behind this study was to empirically measure the performance impacts on
the above-mentioned sample workloads due to:
a) The vPars software stack.
b) The different resource layouts in a virtual partition.
c) The influence of workloads running simultaneously in co-hosted virtual partitions.