|
HTML rendering created 2025-09-06 by Michael Kerrisk, author of The Linux Programming Interface. For details of in-depth Linux/UNIX system programming training courses that I teach, look here. Hosting by jambit GmbH. |
|
|
NAME | SYNOPSIS | DESCRIPTION | INSTALLATION | FILES | PCP ENVIRONMENT | DEBUGGING OPTIONS | SEE ALSO | COLOPHON |
|
|
|
PMDANVIDIA(1) General Commands Manual PMDANVIDIA(1)
pmdanvidia - nvidia gpu metrics domain agent (PMDA)
$PCP_PMDAS_DIR/nvidia/pmdanvidia [-D debug] [-d domain] [-l
logfile] [-t interval]
pmdanvidia is a Performance Metrics Domain Agent (PMDA) which
extracts performance metrics describing the metrics available on
NVIDIA GPU cards via the NVML library.
The nvidia PMDA exports metrics that measure gpu activity, memory
utilization, fan speed, etc on NVIDIA Tesla and Quadro cards.
Metrics are unlikely to be available for consumer class cards.
A brief description of the pmdanvidia command line options
follows:
-d It is absolutely crucial that the performance metrics domain
number specified here is unique and consistent. That is,
domain should be different for every PMDA on the one host,
and the same domain number should be used for the same PMDA
on all hosts.
-l Location of the log file. By default, a log file named
nvidia.log is written in the current directory of pmcd(1)
when pmdanvidia is started, i.e. $PCP_LOG_DIR/pmcd. If the
log file cannot be created or is not writable, output is
written to the standard error instead.
-t Enables and sets a sampling interval for automatic refreshing
of metric values. The functionality is disabled by default,
however this option allows a time interval to be specified on
which all values are sampled - this has the effect of
constantly updating the accumulating metrics, with the goal
of assisting client tools such as pcp-atop(1) and pmlogger(1)
to observe sub-sample time changes in GPU and process state.
Typically these tools have longer sampling intervals, and can
thus 'miss' activity happening during their sampling
interval.
The nvidia PMDA is not installed and available by default. If you
want to undo the installation, do the following as root:
# cd $PCP_PMDAS_DIR/nvidia
# ./Remove
If you want to establish access to the names, help text and values
for the nvidia performance metrics once more, after removal, do
the following as root:
# cd $PCP_PMDAS_DIR/nvidia
# ./Install
pmdanvidia is launched by pmcd(1) and should never be executed
directly. The Install and Remove scripts notify pmcd(1) when the
agent is installed or removed.
$PCP_PMCDCONF_PATH
command line options used to launch pmdanvidia
$PCP_PMDAS_DIR/nvidia/help
default help text file for the nvidia metrics
$PCP_PMDAS_DIR/nvidia/Install
installation script for the pmdanvidia agent
$PCP_PMDAS_DIR/nvidia/Remove
undo installation script for the pmdanvidia agent
$PCP_LOG_DIR/pmcd/nvidia.log
default log file for error messages and other information
from pmdanvidia
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(5).
The -D or --debug option enables the output of additional
diagnostics on stderr to help triage problems, although the
information is sometimes cryptic and primarily intended to provide
guidance for developers rather end-users. debug is a comma
separated list of debugging options; use pmdbg(1) with the -l
option to obtain a list of the available debugging options and
their meaning.
Debugging options specific to pmdanvidia are as follows:
┌────────┬───────────────────────────────────────────────────────┐
│ Option │ Description │
├────────┼───────────────────────────────────────────────────────┤
│ appl0 │ interaction with pmcd(1) │
├────────┼───────────────────────────────────────────────────────┤
│ appl1 │ loading the libnvidia-ml DSO │
├────────┼───────────────────────────────────────────────────────┤
│ appl2 │ trace select(2) calls │
└────────┴───────────────────────────────────────────────────────┘
PCPIntro(1), pcp-atop(1), pmcd(1), pmlogger(1), pcp.conf(5) and
pcp.env(5).
This page is part of the PCP (Performance Co-Pilot) project.
Information about the project can be found at
⟨http://www.pcp.io/⟩. If you have a bug report for this manual
page, send it to [email protected]. This page was obtained from the
project's upstream Git repository
⟨https://github.com/performancecopilot/pcp.git⟩ on 2025-08-11.
(At that time, the date of the most recent commit that was found
in the repository was 2025-08-11.) If you discover any rendering
problems in this HTML version of the page, or you believe there is
a better or more up-to-date source for the page, or you have
corrections or improvements to the information in this COLOPHON
(which is not part of the original manual page), send a mail to
[email protected]
Performance Co-Pilot PCP PMDANVIDIA(1)
Pages that refer to this page: pcp-atop(1)
|
HTML rendering created 2025-09-06 by Michael Kerrisk, author of The Linux Programming Interface. For details of in-depth Linux/UNIX system programming training courses that I teach, look here. Hosting by jambit GmbH. |
|