Update CONTRIBUTING.md

CONTRIBUTING.md

@@ -153,7 +153,6 @@ FontoXPath contains different test sets:
 |The QT3 XQueryX tests|`npm run qt3testsxqueryx`|
 |The XQUTS tests|`npm run xqutstests`|
 |The XQUTS XQueryX tests|`npm run xqutstestsxqueryx`|
-|The QT3 performance tests|`npm run performance`|
 
 They all run in Node. By running the tests with the `--inspect` flag,
 they can be debugged by the browser: `npm run test -- --inspect
@@ -185,16 +184,23 @@ If you are adding a new feature, don't forget to edit the file
 `test/runnableTestSets.csv`. This file disables tests for features we
 have not yet implemented.
 
-To check the performance of fontoxpath we pick a random subset of the
-qt3tests as running all will take too long (hours). This random subset
-is not checked in but can be generated using
-`npm run performance -- --regenerate [<number-of-tests>]`, this will create
-and populate `test/runnablePerformanceTestNames.csv`. You can manually
-edit this file to run specific tests. By storing these in a file you can
-run the same set even when switching between different commits. With the
-generated file present you can run the tests using `npm run performance`,
-this will run a benchmark for each qt3 test using
-[benchmarkjs](https://benchmarkjs.com/).
+### Running benchmarks
+
+FontoXPath has 2 options to run benchmarks.
+
+In one we run benchmarks over scenarios defined in javascript which are located in the directory
+`/performance`. These can be run using `npm run performance` which will run the benchmarks in the
+console. Or you can start a server which hosts them using `npm run performance-server` which allows
+you to test the performance in different browsers. Some of these benchmarks are set up as a
+comparison which indicates the performance overhead of FontoXPath.
+
+To check the performance of fontoxpath with the qt3tests, we pick a random subset of the qt3tests as
+running all will take too long (hours). This random subset is not checked in but can be generated
+using `npm run qt3performance -- --regenerate [<number-of-tests>]`, this will create and populate
+`test/runnablePerformanceTestNames.csv`. You can manually edit this file to run specific tests. By
+storing these in a file you can run the same set even when switching between different commits. With
+the generated file present you can run the tests using `npm run qt3performance`, this will run a
+benchmark for each qt3 test using [benchmarkjs](https://benchmarkjs.com/).
 
 ### Building