x11perf (xorg-x11perf) — a specialized X11 (libx11) console application to run one or more tests measuring the performance of the X Window System (X11). Developed as part of the project X.Org (X.Org Foundation together with freedesktop.org community).
X Window System (X11 or just X) — window system, standard tools and protocols for building the graphical user interface provide the basic functionality for the graphical environment but does not define details of the user interface (this is done by the window managers). For this reason, the appearance of programs in the environment of X11 can be very different depending on the capabilities and settings of a specific window Manager.
In X11 network transparency is provided: a graphics application can run on another machine in the network and their interface will be transmitted across the network and displayed on the user’s local machine. In the context of the X Window System , the terms “client” and “server” are unfamiliar to many users, the value of: “server” means the local user display (display server)and “client” is an application that uses this display (can be run on a remote computer).
Many tests of the X Window System assume that the graphics device used to display the output for a single graphics application. Such tests typically measure the speed of drawing lines, polygons, text, etc… But the workstations mostly are not used as standalone graphics engines, so the test x11perf measures not only the performance of window management, and examines the output performance of the traditional graphics.
x11perf includes tests on the time required to create and display Windows (as when you run the app), displaying, on the screen of an already existing set of Windows, and moving Windows. The test also measures graphics performance for operations not normally used in standalone graphics displays, but are nonetheless frequently used applications X11 (scrolling, touches, etc).
Before running tests that x11perf measures the initial time and sinhroniziruete it with the end time, it guarantees the correct execution of the tests when working with graphics accelerators. The number of repetitions of each test “default” is automatically calibrated, so each of them requires about the same length to run on workstations with very different speeds. Some tests window displays, cancel the show or move it one parent window and report how many child Windows (in second) can be displayed or moved.
x11perf is focused on the analysis of specific strengths and weaknesses of X11, for a comprehensive study of almost all operations. The test does not pretend to absolute accuracy performance X11, although can be used as a reference when comparing. The resulting figures have no informative value and can be misleading, however, for specific applications of the obtained data can be used in predicting performance on the current version of the X Window System.
Utility x11perfcomp (included with test)can be used to compare results of different runs of x11perf, providing a list of files containing results from multiple tests. The tests are documented in a readable tabular format, adding missing results, when comparing the results of the first file is used to determine the set of currently running tests.
To obtain repeatable results tests x11perf should be run at least five times to see if each test takes about the same amount of time in each run. Sometimes are not all tests, in this case, the result is automatically filled in as zero (with a rate of 0.0 objects / sec). Since the first file determines which tests are run, it is best that this file contains an expanded set of tests (which are available in other files)or the comparison of results will fail.
x11perf was used as the engine when writing utilities x11perfcomp, which helps to simplify the comparison of test results between different machines, allowing you to cover the most important and not functions of the X Window System with a maximum number of running graphics operations.
License: MLT (The MIT License)