A typical example on profiling with Yappi, includes at least 3 lines of code:
import yappi
def a():
for i in range(10000000): pass
yappi.start()
a()
yappi.get_func_stats().print_all()
yappi.get_thread_stats().print_all()And the output of running above script:
Clock type: cpu
Ordered by: totaltime, desc
name ncall tsub ttot tavg
deneme.py:35 a 1 0.296402 0.296402 0.296402
name tid ttot scnt
_MainThread 6016 0.296402 1
Let's inspect the results in detail.
The first line:
Clock type: cpu
indicates the profiling timing stats shown are retrieved using the CPU clock. That means the actual CPU time spent in the function is shown.
Yappi provides two modes of operation: CPU and Wall time profiling. You can change the
setting by a call to yappi.set_clock_type().
Read ClockTypes for more.
Second is:
Ordered by: totaltime, desc
This shows the sort order and sort key of the
shown profiling stats. You can see the valid values for this here - YFuncStats.sort().
Now we actually see the statistic of the function call a():
name ncall tsub ttot tavg
deneme.py:35 a 1 0.296402 0.296402 0.296402
Here is what each of these mean -
| Title | Description |
|---|---|
| name | the full unique name of the called function. |
| ncall | How many times this function is called. |
| tsub | How much time this function has spent in total, subcalls excluded. |
| ttot | How much time this function has spent in total, subcalls included. |
| tavg | How much time this function has spent in average, subcalls included. |
The next lines shows the thread stats:
name tid ttot scnt
_MainThread 6016 0.296402 1
Here is what each of these mean -
| Title | Description |
|---|---|
| name | The class name of the Thread. (This is the name of the class inherits the threading.Thread class) |
| tid | The thread id. |
| ttot | How much time this thread has spent in total. |
| scnt | How many times this thread is scheduled. |