tottime

tottime is a profiling metric commonly used in program profilers, notably Python’s cProfile and the pstats module. It denotes the total time spent executing the code of a specific function itself, excluding time spent in functions that it calls (subcalls). The value is typically reported in seconds and is often shown alongside other columns such as ncalls (number of calls), percall (tottime divided by ncalls), and cumtime (cumulative time including subcalls).

Interpreting tottime helps identify functions that are intrinsically expensive: a high tottime indicates that the body of the function consumes substantial CPU time per call, whereas a low tottime combined with a high cumtime suggests the function’s cost comes mainly from the functions it invokes. Sorting profile results by tottime is a common first step in performance tuning to find hotspots that can benefit from algorithmic improvements or micro-optimizations.

Limitations of tottime include measurement overhead and granularity: deterministic profilers add instrumentation cost, and built-in or native code may be attributed differently or not at all. I/O wait, blocking, or time spent in the operating system may not be reflected accurately. Interpreting tottime alongside cumtime, call counts, and knowledge of program structure yields the most reliable guidance for optimization efforts.