diff options
author | Chris Wilson <chris@chris-wilson.co.uk> | 2009-06-10 08:49:39 +0100 |
---|---|---|
committer | Chris Wilson <chris@chris-wilson.co.uk> | 2009-06-10 08:52:50 +0100 |
commit | 81b5dc42b0e754d602506a8ccd231df9afd71593 (patch) | |
tree | 5a08dae49652b52bbede6a05b4658d59472d52a9 /perf/README | |
parent | ec92e633edd377747155b60aa225b266c38bc498 (diff) | |
download | cairo-81b5dc42b0e754d602506a8ccd231df9afd71593.tar.gz |
[perf] Expand the section on cairo-perf-trace in the README
Promote the information on how to use cairo-perf-trace and include it
immediately after the details on cairo-perf. This should make it much
clearer on how to replay the traces, and the difference between the two
benchmarks.
Diffstat (limited to 'perf/README')
-rw-r--r-- | perf/README | 37 |
1 files changed, 34 insertions, 3 deletions
diff --git a/perf/README b/perf/README index abf0687d1..fd6eb308d 100644 --- a/perf/README +++ b/perf/README @@ -9,8 +9,20 @@ more details on other options for running the suite below. Running the cairo performance suite ----------------------------------- -The lowest-level means of running the test suite is with the -cairo-perf program, (which is what "make perf" executes). Some +The performance suite is composed of two types of tests, micro- and +macro-benchmarks. The micro-benchmarks are a series of hand-written, +short, synthetic tests that measure the speed of doing a simple +operation such as painting a surface or showing glyphs. These aim to +give very good feedback on whether a performance related patch is +successful without causing any performance degradations elsewhere. The +second type of benchmark consists of replaying a cairo-trace from a +large application during typical usage. These aim to give an overall +feel as to whether cairo is faster for everyday use. + +Running the micro-benchmarks +---------------------------- +The micro-benchmarks are compiled into a single executable called +cairo-perf, which is what "make perf" executes. Some examples of running it: # Report on all tests with default number of iterations: @@ -29,6 +41,25 @@ when using cairo-perf-diff to compare separate runs (see more below). The advantage of using the raw mode is that test runs can be generated incrementally and appended to existing reports. +Running the macro-benchmarks +---------------------------- +The macro-benchmarks are run by a single program called +cairo-perf-trace, which is also executed by "make perf". +cairo-perf-trace loops over the series of traces stored beneath +cairo-traces/. cairo-perf-trace produces the same output and takes the +same arguments as cairo-perf. Some examples of running it: + + # Report on all tests with default number of iterations: + ./cairo-perf-trace + + # Report on 100 iterations of all firefox tests: + ./cairo-perf-trace -i 100 firefox + + # Generate raw results for 10 iterations into cairo.perf + ./cairo-perf-trace -r -i 10 > cairo.perf + # Append 10 more iterations of the poppler tests + ./cairo-perf-trace -r -i 10 poppler >> cairo.perf + Generating comparisons of separate runs --------------------------------------- It's often useful to generate a chart showing the comparison of two @@ -180,7 +211,7 @@ added: 64x64. -How to benchmark traces +How to record new traces ----------------------- Using cairo-trace you can record the exact sequence of graphic operations made by an application and replay them later. These traces can then be |