MPVIS(1) General Commands Manual MPVIS(1)

mpvis - visualize cpu utilization/performance

mpvis [-biV] [-r rowlength] [-R rowlength] [pmview options] [cpuid ...]

mpvis displays a three dimensional bar chart of CPU utilization. The display is updated with new values retrieved from the target host or archive every interval seconds (default is 2 seconds).

The height of the bars is proportional to the CPU utilization in each of the modes idle (with no I/O pending), wait (idle, but waiting for I/O - only visible in IRIX), intr (interrupt processing - only visible in IRIX), nice (nice state - only visible in Linux), sys (in the kernel) and user.

The number of CPUs per row in the scene is governed by the default maximum row length, and the options -b, -r and -R. If none of these flags are specified, mpvis uses the default maximum row length to calculate the actual number of rows and columns in the view, according to:

nrows = (ncpus + maxrowlen - 1) / maxrowlen
ncols = (ncpus + nrows - 1) / nrows.

The -r option uses the above algorithm, but allows the user to override the maximum row length.

The CPUs are sorted in ascending order. They are displayed from left to right, and front to back (in the case where there is more than one row of CPUs).

The user can specify a list of CPUs as cpuid arguments using the naming scheme reported by

$ pminfo -f kernel.percpu.cpu.user

i.e. cpuN or cpuA:B depending on the operating system version and platform. Alternatively, if the cpuid argument contains one of the characters ^ or . or [ then cpuid will be treated as a regular expression in the style of egrep(1) and the matching set of CPU names will be used instead of cpuid.

If one or more cpuid arguments is specified, then only the CPUs in this list are displayed in the view. The list of CPUs is sorted into ascending CPU number and they are displayed from left to right, and front to back (in the case where there is more than one row of CPUs).

mpvis generates a pmview(1) configuration file, and passes most command line options to pmview(1). Therefore, the command line options -A, -a, -C, -h, -n, -O, -p, -S, -t, -T, -Z and -z, and the user interface are described in the pmview(1) man page.

The mpvis specific options are:

Use a ratio for row:column of 1:8. This option is best for large numbers of CPUs. If the row length is not a factor of the number of CPUs, then the last row will be shorter than all other rows. This option ignores the maximum row length.
Label each CPU column with the CPU number.
Split the view into a number of rows, where each row has a maximum length of n CPUs. This is a soft limit. The actual row length is calculated by an algorithm (see above) which prduces a balanced view with no more than n CPUs per row.
Split the view into a number of rows, where each row has a length of n CPUs. This is a hard limit. If the row length is not a factor of the number of CPUs, then the last row will be shorter than all other rows.
Verbose mode - output the generated pmview(1) configuration file.

The behavior of pmchart(1) when launched from mpvis is dependent on the number of CPUs in each row. If there are more than 6 CPUs in a row then each chart will contain a separate metric, otherwise each chart will contain a separate CPU instance.

$PCP_VAR_DIR/config/pmlogger/config.mpvis
A pmlogger(1) configuration file for mpvis metrics.
$PCP_SHARE_DIR/lib/pmview-args
Shell procedures for parsing pmview(1) command line options.

Environment variables with the prefix PCP_ are used to parameterize the file and directory names used by PCP. On each installation, the file /etc/pcp.conf contains the local values for these variables. The $PCP_CONF variable may be used to specify an alternative configuration file, as described in pcp.conf(4).

dkvis(1), nfsvis(1), pmcd(1), pmchart(1), pmlogger(1) and pmview(1).

The CPU view for pmchart(1).

Performance Co-Pilot