Measurement Update Performance Issues

Some measurement scenarios require the use of large amounts of RAM memory or a high volume of data transfer. In such cases, you might experience measurement update performance issues if your PC runs out of memory or processing power, or with data transfer over a low-bandwidth interface.

This topic discusses the symptoms of measurement update performance issues and how to manage them for best results.

Symptoms

Measurement updates are too slow.
Long delays in application response to changing measurement parameters.
VSA application consumes too much memory.

Possible Problems and Solutions

The length and number of channels of data acquisition required to satisfy the running measurement(s) have a significant impact on measurement performance (speed).

Parameters like Result Length or Search Length, coupled with Span and Symbol Rate or Bandwidth of the signal under test determine the amount of data that needs to be acquired and transferred from instrumentation, or read into memory from a recording file. Reducing the required bandwidth or time duration necessary to complete a measurement may significantly improve measurement performance.
If data is acquired from hardware, the transfer time of the data from instrumentation to the VSA application can sometimes be improved by switching to a higher speed interface, or by changing the location of the processing (Using Remote Desktop to control a VSA session rather than transferring acquired data over a VPN Virtual Private Network: A private telecommunications network created using the resources of the PSTN and customized dialing, switching and routing functions., for instance).

The number of measurements

Shared measurements impose additional requirements on the acquisition length. All constraints imposed by the multiple measurement definitions must be satisfied by a single acquisition vector. Variations between measurements such as span and time length allow unique cross-measurement insights, but also requires more data than independent measurements utilizing multiple acquisitions.
See the Multi-Measurements Overview topic for a complete discussion of multiple measurements using the VSA software.

The number of traces with mapped data, and visibility of traces with mapped data

If data is mapped to a trace, every point within that mapped data is processed and presented on the trace. For measurements with hundreds of thousands of points, this can represent significant memory and processing time. If unneeded traces use a mapping to "No Data", this overhead may be reduced, and measurement update performance can be improved.
For remote programmatic users, data is available directly from the measurement and requires no trace data mapping. See the .NET API 1) Access Preamble Indicator, or 2) Application Programming Interface reference for more information on using this approach to improve performance in automated measurements.

The computer running the VSA

Measurements which require large amounts of data analysis or trace processing require more resources from the host computer.
Increasing the computer RAM, Video subsystem performance, or number of cores will generally improve the performance of VSA measurement update rates.