The metric value is low, which may signal a poor logical CPU cores utilization caused by load imbalance, threading runtime overhead, contended synchronization, or thread/process underutilization. Explore sub-metrics to estimate the efficiency of MPI and OpenMP parallelism or run the Locks and Waits analysis to identify parallel bottlenecks for other parallel runtimes.
| Sync Object | Wait Time with poor CPU Utilization | (% from Object Wait Time)(%) | Wait Count |
|---|---|---|---|
| Condition Variable 0x0d43fa13 | 6.377s | 64.0% | 1 |
| Condition Variable 0xd9e74274 | 6.270s | 63.6% | 1 |
| Futex 0x251f10cb | 4.865s | 100.0% | 56 |
| Condition Variable 0x17c7c503 | 1.201s | 100.0% | 439 |
| Mutex 0xe462291b | 0.225s | 100.0% | 229 |
| [Others] | 0.453s | 100.0% | 355 |
| Function | Module | Spin and Overhead Time | (% from CPU Time)(%) |
|---|---|---|---|
| __sched_yield | libc.so.6 | 0.245s | 0.5% |
| [MKL SERVICE]@get_dynamic | libmkl_intel_thread.so.2 | 0.030s | 0.1% |
| _PyFunction_Vectorcall | python3.9 | 0.028s | 0.1% |
| PyObject_GetAttr | python3.9 | 0.012s | 0.0% |
| std::__atomic_base<int>::load | libtbb.so.12 | 0.012s | 0.0% |
| [Others] | N/A | 0.018s | 0.0% |