[Good First Issue]: if operation is incorrectly reporting profiling data in GPU

Context

Performance counter numbers works incorrectly with if operation. Using models with if layer, the reported time is almost zero, and there is a big delta between the some of the latencies of the operations vs. the end-to-end latency reported by benchmark_app.

Counter statistics are critical for performance analysis and optimizations

Below "total time" is reported by -pc, whereas "Latency" is computed as end-to-end inference latency by benchmark_app.

What needs to be done?

Fix performance counters in if operation to see adequate layer statistics

Example Pull Requests

No response

Resources

• Contribution guide - start here!
• Intel DevHub Discord channel - engage in discussions, ask questions and talk to OpenVINO developers
• How to link your Pull Request to an issue

Contact points

sshlyapn, p-durandin

Ticket

No response