Skip to content

Commit

Permalink
[INTERNAL] Document the UI5 Tooling benchmarking process
Browse files Browse the repository at this point in the history
PRs where this process was already used:
* SAP/ui5-builder#488
* SAP/ui5-builder#390

JIRA: CPOUI5FOUNDATION-261
  • Loading branch information
RandomByte committed Dec 2, 2020
1 parent ab398bc commit 1db338f
Showing 1 changed file with 149 additions and 0 deletions.
149 changes: 149 additions & 0 deletions docs/pages/Troubleshooting.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,3 +8,152 @@ You need to delete the HSTS mapping in [chrome://net-internals/#hsts](chrome://n

## Issues Not Listed Here
Please follow our [Contribution Guidelines](https://github.com/SAP/ui5-tooling/blob/master/CONTRIBUTING.md#report-an-issue) on how to report an issue.

## Benchmarking UI5 Tooling

For benchmarking UI5 Tooling we typically make use of the open source tool [hyperfine](https://github.com/sharkdp/hyperfine).

In general we only benchmark calls to the UI5 CLI. However, we might add scripted benchmarks for some components in the future.

The following is a walk-through on how to compare the performance impact of an imaginary change in UI5 Builder project.

### Setup

1. Install [hyperfine](https://github.com/sharkdp/hyperfine#installation)
1. Prepare the UI5 Tooling projects you want to measure *(optional if your development environment already reflects this)*:
1. Start in an empty directory
```sh
mkdir ui5-tooling-benchmark && cd ui5-tooling-benchmark/
```
1. Clone [UI5 CLI](https://github.com/SAP/ui5-cli)
```sh
git clone git@github.com:SAP/ui5-cli.git
```
1. Clone [UI5 Builder](https://github.com/SAP/ui5-builder) (or your fork!)
```sh
git clone git@github.com:SAP/ui5-builder.git
```
Make sure you checkout the `master` branch since we will perform the baseline test first
1. Install npm dependencies in both directories
```sh
(cd ui5-cli && npm install)
(cd ui5-builder && npm install)
```
1. Create global npm links for both projects
```sh
(cd ui5-cli && npm link)
(cd ui5-builder && npm link)
```
1. Link UI5 Builder into UI5 CLI
```sh
(cd ui5-cli && npm link @ui5/builder)
```
1. Verify your setup
```sh
ui5 --version
```
This should output the version and location of the UI5 CLI you just cloned.

For example:
```
2.6.6 (from /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js)
```

1. Prepare your test project (we chose the [openui5-sample-app](https://github.com/SAP/openui5-sample-app))
1. Clone the project
```sh
git clone git@github.com:SAP/openui5-sample-app.git
```
1. Navigate into the project
```sh
cd openui5-sample-app
```
1. Install any required npm dependencies
```sh
npm install
```
Note: We will not link UI5 CLI into the but call the global install directly
1. Verify that the previously installed UI5 CLI can be called with the following command
```sh
UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js --version
```
On Windows:
```sh
set UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js --version
```
*(Replace the path to ui5.js with the one shown in the previous `ui5 --version` output)*

### Benchmark

1. Depending on how reliable you want the measurements, consider preparing your system:
1. Connect your computer to a power supply
1. Make sure no updates or anti virus scans are taking place
1. Close all applications. This includes your IDE since it might start indexing any new files created during the build, thus impacting I/O
1. Don't interact with your system wile the benchmark is running
1. Perform the baseline measurement
1. In the project, start your first benchmark
```sh
hyperfine --warmup 1 \
'UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build' \
--export-markdown ./baseline.md
```
On Windows:
```sh
hyperfine --warmup 1 \
'set UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build' \
--export-markdown ./baseline.md
```
1. Your baseline benchmark is now stored in `baseline.md` and should look similar to this:
| Command | Mean [s] | Min [s] | Max [s] | Relative |
|:---|---:|---:|---:|---:|
| `UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build` | 1.439 ± 0.036 | 1.400 | 1.507 | 1.00 |
1. Prepare your change
1. Switch to the branch that contains your change
```sh
(cd ../ui5-builder && git checkout my-change)
```
1. If your change requires different npm dependencies, reinstall them
```sh
(cd ../ui5-builder && npm install)
```
1. The link from UI5 CLI is still in place. However, if you have changes in **multiple** UI5 Tooling modules, you might need to `npm link` those again
1. Perform the change measurement
1. In the project, start your second benchmark
```sh
hyperfine --warmup 1 \
'UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build' \
--export-markdown ./my_change.md
```
On Windows:
```sh
hyperfine --warmup 1 \
'set UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build' \
--export-markdown ./my_change.md
```
1. Your change's benchmark is now stored in `my_change.md`

### Compile Results

1. Merge both measurements into one markdown
1. Either remove the `Relative` column or calculate the relative difference yourself.
1. You should end up with a markdown like this:
```md
| Command | Mean [s] | Min [s] | Max [s] |
|:---|---:|---:|---:|
| `UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build` | 1.439 ± 0.036 | 1.400 | 1.507 |
| `UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build` | 1.584 ± 0.074 | 1.477 | 1.680 |
```
Rendering like this:

| Command | Mean [s] | Min [s] | Max [s] |
|:---|---:|---:|---:|
| `UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build` | 1.439 ± 0.036 | 1.400 | 1.507 |
| `UI5_CLI_NO_LOCAL=X node /my/home/ui5-tooling-benchmark/ui5-cli/bin/ui5.js build` | 1.584 ± 0.074 | 1.477 | 1.680 |

1. You can now share these results on GitHub or wherever you might need them!

**Happy benchmarking! 🏎**

0 comments on commit 1db338f

Please sign in to comment.