# Playwright Parallelization

### Playwright Sharding

Running Playwright tests on CI eventually will require splitting the spec files between several CI machines for faster execution. Playwright has the built-in ability to split the tests into shards - each shard will run only a subset of tests. 

Read more about native [playwright-sharding](https://docs.currents.dev/~/changes/TYKBeqcc7WyqjmgOzWzB/guides/parallelization-guide/pw-parallelization/playwright-sharding "mention").

<figure><img src="https://3745692499-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FqmFDEiUa9mr11LUlxDnt%2Fuploads%2FQ9X9Ey8OPtDDi5qXKb6U%2Fpw-shard-fast-bg.png?alt=media&#x26;token=6e7d0445-ce17-4212-91cc-efa77f708053" alt=""><figcaption><p>Sharding vs Orchestration</p></figcaption></figure>

### Playwright Orchestration by Currents

As your testing suite grows, you can discover that the native sharding is not always satisfactory and can be optimized:

* the native split of spec files between machines can be non-optimal, leading to underutilized resources and slow CI runs
* updating the shards amount requires updating CI configuration
* spec files' names affect the distribution of tests between shards
* collecting the reports across multiple machines is cumbersome

Read more about [Playwright Orchestration](https://docs.currents.dev/~/changes/TYKBeqcc7WyqjmgOzWzB/guides/parallelization-guide/pw-parallelization/playwright-orchestration) to discover how Currents improves the native orchestration.