diff options
author | Arseny Kapoulkine <arseny.kapoulkine@gmail.com> | 2014-12-11 21:08:20 -0800 |
---|---|---|
committer | Arseny Kapoulkine <arseny.kapoulkine@gmail.com> | 2014-12-11 21:08:20 -0800 |
commit | cdf0836b8fd08497766015b174d51474cfe9dda3 (patch) | |
tree | 8e1860a20b27374c3d40e16451eea4f8a3a73546 | |
parent | fd761d5b1c4299302061fbab2835804e74e3bb55 (diff) |
Update benchmark.md
-rw-r--r-- | benchmark.md | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/benchmark.md b/benchmark.md index 75c5b37..df806c2 100644 --- a/benchmark.md +++ b/benchmark.md @@ -17,7 +17,7 @@ All tests were performed with Microsoft Visual Studio 2015, with all optimizatio Benchmarks are performed on nine different test files; all of them are in 1-20 Mb range and display a different range of markup/content ratios - e.g. there's a COLLADA file that has little markup, there's a book file that has almost no markup, there are some files where there is a lot of markup and almost no content, etc. The test files include a 10 Mb [XMark](http://www.xml-benchmark.org) test file. -Due to a number of test files all results are normalized by pugixml results - the ratio between time or memory for a particular parser to pugixml is displayed (so if it's 1 or more then the parser is slower or takes more memory than pugixml). Note that the horizontal axis is using logarithmic scale. +Due to a number of test files all results are normalized using pugixml results as a baseline - the ratio between time or memory for a particular parser to pugixml is displayed (so if it's 1 or more then the parser is slower or takes more memory than pugixml). Note that the horizontal axis is using logarithmic scale. Benchmark results are visualized using [Google Charts](https://developers.google.com/chart/) and require JavaScript and SVG support to render. |