# Performance improvements via AWS or Azure?

**URL:** <https://discourse.pymc.io/t/performance-improvements-via-aws-or-azure/17128>\
**Category:** v5\
**Created:** [June 23, 2025, 2:00pm UTC](https://discourse.pymc.io/t/performance-improvements-via-aws-or-azure/17128 "2025-06-23T14:00:08Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![bridgeland](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/bridgeland/32/2773_2.png) [@bridgeland](https://discourse.pymc.io/u/bridgeland)\
**Post date:** [June 23, 2025, 2:00pm UTC](https://discourse.pymc.io/t/performance-improvements-via-aws-or-azure/17128/1 "2025-06-23T14:00:08Z")

</div>

I am running PyMC5 models on my MacOS laptop. My laptop has plenty of memory and an M2 processor.

As of June 2025, is it possible to get better sampling performance using AWS or Azure or maybe Google Cloud? In other words, is there a way to configure a cloud platform to sample faster than can be achieved locally with a 2022 Apple Silicon device?

---

<div class="post-metadata">

**Author:** ![sam-walsh](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/sam-walsh/32/9243_2.png) [@sam-walsh](https://discourse.pymc.io/u/sam-walsh)\
**Post date:** [June 23, 2025, 3:47pm UTC](https://discourse.pymc.io/t/performance-improvements-via-aws-or-azure/17128/2 "2025-06-23T15:47:07Z")

</div>

I’ve tested AWS r7a and r7i clusters. They’re fast but not quite on par with my M4 chip. Probably competitive with or faster than some of the older apple silicon.