},
                    )
                }.sortedBy { -it.totalTime }
                .filter { onlyNativeSubprojectsForIntelMacs(testCoverage, it.subProject.name) }

        return parallelize(subProjectTestClassTimes, testCoverage) { numberOfBatches ->
            when (testCoverage.os) {
                Os.LINUX -> ParallelizationMethod.TestDistribution

## Deployment Concepts { #deployment-concepts }

Here you saw how to use multiple **workers** to **parallelize** the execution of the application, take advantage of **multiple cores** in the CPU, and be able to serve **more requests**.

* Create a new Deployment in kube-system for every version. ([#23512](https://github.com/kubernetes/kubernetes/pull/23512), [@Q-Lee](https://github.com/Q-Lee))
* Use SCP to dump logs and parallelize a bit. ([#22835](https://github.com/kubernetes/kubernetes/pull/22835), [@spxtr](https://github.com/spxtr))

* AWS kube-down: Issue warning if VPC not found ([#27518](https://github.com/kubernetes/kubernetes/pull/27518), [@justinsb](https://github.com/justinsb))
* gce/kube-down: Parallelize IGM deletion, batch more ([#27302](https://github.com/kubernetes/kubernetes/pull/27302), [@zmerlynn](https://github.com/zmerlynn))

* Replace Parallelize with function ParallelizeUntil and formally deprecate the Parallelize. ([#68403](https://github.com/kubernetes/kubernetes/pull/68403), [@wgliang](https://github.com/wgliang))

- Removed the function Parallelize, please convert to use the function ParallelizeUntil. ([#76595](https://github.com/kubernetes/kubernetes/pull/76595), [@danielqsj](https://github.com/danielqsj))

### Apps

- Failed pods associated with a job with `parallelism = 1` are recreated by the job controller honoring exponential backoff delay again. However, for jobs with `parallelism > 1`, pods might be created without exponential backoff delay. ([#115022](https://github.com/kubernetes/kubernetes/pull/115022), [@nikhita](https://github.com/nikhita)) [SIG Apps]

### Est-ce que la concurrence est mieux que le parallélisme ? { #is-concurrency-better-than-parallelism }

Nope ! C'est ça la morale de l'histoire.

La concurrence est différente du parallélisme. C'est mieux sur des scénarios **spécifiques** qui impliquent beaucoup d'attente. À cause de ça, c'est généralement bien meilleur que le parallélisme pour le développement d'applications web. Mais pas pour tout.

### Feature

- Parallelize calls to conversion webhooks when watching from etcd to prevent watchcache from lagging and helps clusters with slow conversion webhooks or transformers. 

## Concepts de déploiement { #deployment-concepts }

Ici, vous avez vu comment utiliser plusieurs workers pour paralléliser l'exécution de l'application, tirer parti de plusieurs cœurs du CPU et être en mesure de servir davantage de requêtes.