Impact of Powerful Hardware on Developer Productivity

• Developers want more powerful hardware because it means less time waiting on builds, which means more time to build the next feature or fix a bug.

• However, even if the upfront cost is higher for higher-powered hardware, what’s the actual cost when you consider the impact on developer productivity?

• In my experiment, I used GitHub’s new, larger hosted runners to execute a large build at each compute tier from 2 cores to 64 cores. I then determined how long each build took and its cost on GitHub’s larger runners.

• The results showed that, on average, builds completed faster on more powerful hardware, and that the cost of a build was lower when executed on larger runners.

• This means that, on average, businesses are actually spending less money on developer time when more powerful hardware is used.

• The cost of hardware is much, much less than the total cost for developers, and giving your engineering teams more CPU power means they have more time to develop software instead of waiting on builds to complete.

• Labor is much, much more expensive than compute resources. And that means spending a few more dollars to speed up the build is inconsequential in the long run.

• When you assume the average hourly rate of a developer is $75 (USD), the graph above shows that it almost always makes sense to pay more for more compute power so your developers aren’t left waiting or context switching.

• Even the most expensive compute option—$15 an hour for 64 cores and 256GB of RAM—only accounts for a fifth of the hourly cost of a single developer’s time. As developer salaries increase, the cost of hardware decreases, or the time the job takes to run decreases—and this inverse ratio bolsters the case for buying better equipment.

• So remember: The upfront cost for more CPU power pays off over time.

• ime. And your developers will thank you (trust us).

Explore the impact of using more powerful hardware on developer productivity, and how it affects the overall cost of builds. Discover the results of an experiment comparing build times and costs on different compute tiers, and learn why investing in better equipment pays off in the long run.