Upgrading Hardware Across a Global Workforce

Was the performance improvement worth the investment? Yes!Īnd so the question moved from if to how. Was the build time performance better? Yes! The faster development times lead to the break-even timeline occurring in three months and potential savings of 36% in engineering costs per team.

Overall: Apple silicon is about 43% faster.Īndroid: Apple silicon is about 50% faster:įinancial Benefits in Upgrading to the M1s What metric will Spotify consider successful?ĪvgLocalBuildTime(M1) < AvgLocalBuildTime(non-M1) Will the performance improvements justify the costs associated with the investment needed for the upgrade? This blog post written by a Reddit staff engineer inspired Spotify to evaluate the upgrade as a financial use case as well. Were the new machines “good enough” to be used by Spotify developers? Good enough, in this case, was defined as a statistically significant improvement in Android/ iOS local build times. The recent blog post written by Spotify Product Manager Anna Granhed and Spotify Senior Data Scientist Nandini Singh explained that two questions needed sufficient answers before the upgrade was approved. Spotify spent 2022 upgrading developer hardware. Spotify ran multiple build tests on different hardware and found that the build times on machines with Apple silicon (the M1 chip and higher) were 43% faster overall (the M1 Pro and M1Max yielded slightly better results than the M1). Spotify conducts quarterly Engineering Satisfaction surveys, and one such survey revealed that longer build times reduced developer productivity and satisfaction.

Music streaming service Spotify recently upgraded to the Apple M1 gear for its Client Platform (CLiP) team after internal testing proved the machines are 43% faster than Intel-based Mac systems overall: 50% faster on Android builds and 40% faster on iOS builds.

Spotify mac client