How do memory requirements change when using a multi-module architecture during build?

Using multi-module architecture can significantly increase memory requirements during the build process.

Main reasons:

• Larger dependency graph: The build system (e.g., Gradle) must process a complex dependency graph between modules. This requires more memory to store information about modules, their dependencies, and build order.
• Parallel build: If parallel module building is enabled, each parallel process or thread may consume its own share of memory.
• Caching: Gradle and other build tools use caching to speed up incremental builds. The more modules, the more data can be stored in cache, increasing memory demands.
• Resource processing: Building resources (XML, images, etc.) in each module, merging, and optimization also require memory.
• KAPT/KSP: Annotation processors (KAPT for Kotlin, KSP) can consume significant memory resources, especially in large projects with many modules.

To reduce memory consumption:

• Allocate more memory to Gradle daemon in gradle.properties:

  org.gradle.jvmargs=-Xmx4g -XX:MaxPermSize=512m -XX:+HeapDumpOnOutOfMemoryError -Dfile.encoding=UTF8
  

• Optimize dependencies between modules.
• Use incremental builds.
• Use Gradle Build Cache.
• Profile the build to identify bottlenecks.