are there any estimations/experiments on how much ...
# gradlep
Piotr Krzemiński
08/03/2023, 11:03 AMare there any estimations/experiments on how much faster Gradle will become once it adopts K2?
Piotr Krzemiński
08/03/2023, 11:09 AMof course, “your mileage may vary”, depending on the complexity of configs, but some ballpark numbers perhaps?
v
Vampire
08/03/2023, 11:15 AMDunno whether it is up-to-date still, but https://github.com/gradle/gradle/issues/21284 is about trying out K2 and says scripting is not yet supported.
No idea whether that changed though, but maybe you can ask there and follow.
👍 2
z
Zac Sweers
08/03/2023, 8:29 PMCompilation would presumably get faster, but that’s amortized considering you compile buildscripts once. Runtime speed is unchanged
👍 1
👍🏾 1
c
Casey Brooks
08/04/2023, 2:43 PMI recall seeing somewhere that the compiler "frontend" is looking to be about 2x faster. For compiling the project code with Gradle, there should be some improvement to compilation time, but I believe more time is spent in the "backend" overall so I don't think raw compilation speed will quite be 2x. The most noticeable performance improvement is expected to be the IDE experience (syntax highlighting, code inspections,etc), since that's built with the same compiler frontend.
For Gradle scripts themselves which are written in Kotlin, there might be some minor improvement, but the Kotlin compiler speed is almost certainly not the limiting factor in your build performance. Kotlin scripts get cached so only the first build is really impacted by compiler speed. If you've got a slow build now, K2 is not going to magically fix it. You should look into optimizing for Gradle features like configuration cache and incremental compilation instead