Headquartered in London, Improbable is a company dedicated to building technology to enable powerful virtual worlds and simulations designed to help solve previously intractable problems. In gaming and entertainment, this enables the creation of richer, more immersive and persistent virtual worlds.
Improbable has a team of more than 35 developers dedicated to building and refining SDKs and tutorial content to make it easier to use SpatialOS with familiar game development tools such as Unreal Engine. This is complex and technically demanding work, and as a result these developers can encounter slow C++ build times – up to an hour for a full UnrealEngine SDK build.
In addition, Improbable’s work culture encourages flexibility, meaning any acceleration solution had to take this into account.
“Our entire corporate infrastructure is on the cloud. We want our engineers to be able to work effectively even if they are in a different office, or working from home,” says an engineer at Improbable.
How IncrediBuild crunched it:
Since IncrediBuild’s parallel computing acceleration solution can be enjoyed on local machine networks, cloud or hybrid environments, Improbable’s developers leveraged it to accelerate their builds wherever they are.
“We are implementing Incredibuild on Google Cloud Engine so we can spin up agents as required using our VPNs and proxies. This allows people to use and take full advantage of Incredibuild whether they are working from home, their office or elsewhere. Incredibuild allows us to use the cloud and the local network, and to have the flexibility to match our number of agents to our needs. We can scale as we need, without compromising too much on performance.”
On the continuous integration front, Improbable used IncrediBuild to integrate with TeamCity to manage all their CI builds.
“IncrediBuild has reduced C++ compilation times down to manageable levels for our entire team, no matter where they are. By being more flexible and cutting down the time required to build our code, we can iterate faster and produce better software quicker.”