Intel® VTune™ Amplifier enables you to collect data on a remote application from the host system (remote usage mode) via command line interface (amplxe-cl) and view the analysis result locally in the GUI. Remote data collection using the amplxe-cl command running on the host is very similar to the native collection on the target except that the -target-system option is added to the command line.
Prerequisites:
VTune Amplifier is installed on the local host.
VTune Amplifier collectors and drivers are installed on the remote target.
Depending on your remote system, you may choose to install the VTune Amplifier remote target package (available for the VTune Amplifier for Systems only) or full command line interface (
amplxe-cl).A password-less SSH access to the target is set.
Recommended: an analysis target located on a shared drive visible to both local and remote machines.
Note
If you plan to collect data in the native mode using the full-scale command line interface of the VTune Amplifier installed on your target Linux system, see the topic About Running Analysis from Command Line. You may use the Command Line option in the VTune Amplifier graphical interface to automatically generate a command line for an analysis configuration selected in the GUI. Make sure to edit the generated command line for remote collection as described in the Generating Command Line from GUI topic.
Consider the following examples of the data collection launched from the host system in the remote mode:
To run event-based system-wide sampling collection:
The command line below collects system-wide Advanced Hotspots information without call stacks. This command automatically pulls in modules required for viewing results from the device and caches them in the temp directory on the host. This happens only on the first collection, all subsequent collections reuse modules from the cache.
host>./amplxe-cl --target-system=ssh:user@target–collect advanced-hotspots --duration 10
For system-wide collection, a lot of modules running in the system during collection are copied from the target to the host, which may take a while. However, this happens only once since amplxe-cl caches target system modules on the host for faster access on the next collection. If you do not want the command to take the modules from the device, you can specify a local directory where modules will be searched first, for example:
host>./amplxe-cl --target-system=ssh:user@target–collect advanced-hotspots --duration 10 --search-dir <local directory with modules>
In the case above, <local directory with modules> can be either a directory where modules are located, or it can be a pointer to the root file system of the target device. For example, when the collector searches for the /usr/lib64/libstdc++.so.6.0.16 file from the target device, it first tries <local directory with modules>/usr/lib64/libstdc++.so.6.0.16, then it tries <local directory with modules>/libstdc++.so.6.0.16, and only after that it attempts to copy the file from the target device.
To run event-based sampling collection for an application:
In this command line, you attach to a running application by its PID:
host>./amplxe-cl --target-system=ssh:user@target–collect advanced-hotspots --target-pid 333
To run event-based stack sampling collection for the application:
host>./amplxe-cl --target-system=ssh:user@target–collect advanced-hotspots -knob collection-detail=stack-sampling -- <path to the application on the target>
To run a user-mode sampling and tracing collection:
host>./amplxe-cl --target-system=ssh:user@host -collect hotspots -- /target_path/app
To run power system-wide collection (available for VTune Amplifier for Systems only):
host>./amplxe-cl --target-system=ssh:user@target -collect sleep --duration 10
or
host>./amplxe-cl --target-system=ssh:user@target -collect frequency --duration 10
To collect user annotations for tasks (ITT API tasks):
Pass the -knob enable-user-tasks=true option in addition to other options. Note that ITT API can be collected only when your application is launched by the collector.
host>./amplxe-cl --target-system=ssh:user@target–collect advanced-hotspots -knob collection-detail=stack-sampling -knob enable-user-tasks=true -- <path to the application on the target>
To run the General Exploration analysis:
host>./amplxe-cl --target-system=ssh:user@target -collect general-exploration -- <path to the application on the target>
To control collection from the command line:
You can pause, resume and detach collection from the host as follows:
host>./amplxe-cl -r result@@@ -C pause
host>./amplxe-cl -r result@@@ -C resume
host>./amplxe-cl -r result@@@ -C detach
For advanced event-based sampling collection on the Intel processors:
You can configure the VTune Amplifier collector to take other events, such as Cache Misses, Branch Mispredicts. To specify the events, type:
host>./amplxe-cl --target=ssh:user@target --duration 10 -collect-with runsa -knob event-config='<event name1>':modifier1=val
You can take any event supported by the Performance Monitoring Unit (PMU). Additionally, you can enable multiple event collection at a time.
For a complete list of available options for the collector, see amplxe-cl --help.
To identify potential latency or responsiveness issues, use the following collection configuration:
host>./amplxe-cl --target=ssh:user@target --duration 10 -collect-with runsa -knob event-config=“CPU_CLK_UNHALTED.REF:sa=20000”
This command line takes samples at ~2x the rate of a context switch, which gives you an approximately 20% performance hit.