What is PGO?
Profile-guided optimization (PGO) collects execution data while the application runs and uses it to optimize performance.
- Faster execution by optimizing frequently used code paths
- Smaller memory footprint due to better inlining decisions
- Better CPU utilization for improved performance
Types of PGO in .NET
a) Static PGO (Ahead-of-Time Compilation - AOT)
Uses profiling data collected during testing and applies optimizations before deployment.
b) Dynamic PGO (JIT Optimization at Runtime)
Uses real-time execution data and optimizes code while the application is running.
How to Enable PGO in .NET 9 / C# 13?
a) Enabling Dynamic PGO
DOTNET_TieredPGO=1
DOTNET_TC_QuickJitForLoops=1
Or enable it in C# code.
AppContext.SetSwitch("System.Runtime.TieredPGO", true);
b) Enabling Static PGO in .NET 9
Collect Profile Data.
dotnet pgo collect --application-path MyApp.dll
Optimize using the collected data
dotnet pgo optimize --input profile_data.pgo
Example: PGO in Action
public int Compute(int x)
{
if (x == 0) return 10;
if (x == 1) return 20;
if (x == 2) return 30;
return 40;
}
With Dynamic PGO, the JIT compiler will optimize the most frequently used branches for better performance.
When Should You Use PGO?
- Web applications (ASP.NET Core)
- Cloud-native applications
- Game engines (Unity with .NET 9)
- High-performance computing (HPC) applications
Conclusion
PGO in .NET 9 / C# 13 helps applications run faster and more efficiently by analyzing execution patterns in real time.