![]() ![]() We then asked teams across the organisation to make sure dependencies and relevant tooling were accurately updated. We set up a workstream dedicated to assist our developers while in the process of reaching full support. To enable this upgrade, we needed to add support for the new machines to our development infrastructure. How does one prioritise and scale in a distributed-first world? ![]() Were the machines actually better in build time performance? And was the performance improvement worth the investment required to procure the machines?īased on the empirical findings and our financial analysis, upgrading to Apple silicon machines definitely meets our criteria and is a viable solution to reduce developer build time woes. We had two questions to answer to be completely satisfied. In our analysis, we found the upgrade to be cost-effective as well - we break even in about three months, and we could potentially save up to 36% in engineering costs per team. Are there any financial benefits in upgrading to M1s? Overall: Apple silicon is about 43% faster.Īndroid: Apple silicon is about 50% faster. Given the processing capabilities of M1s, we hypothesised that M1 machine local build times would improve significantly.ĪvgLocalBuildTime(M1) < AvgLocalBuildTime(non-M1) Empirical findings *What is “good enough”? For the purpose of our analysis, we define “good enough” as a statistically significant performance improvement in Android/iOS local build times.
0 Comments
Leave a Reply. |