Profiling, recursive func slower than imperative, normal?

[skanemupp]

the 0.409 vs 0.095 is the total times right?
so the imperative function is >4 times faster than the recursive.
or what does tottime stand for?

is this always the case that the recursive function is slower?
the gain is less code?

are some functions only implementable recursively?



def power(nbr, po):
if po==0:
return 1
if po>0:
return nbr*power(nbr, po-1)
if po<0:
return 1/power(nbr, -1*po)

109992 function calls (10002 primitive calls) in 0.409 CPU seconds

Ordered by: standard name

ncalls tottime percall cumtime percall
filename:lineno(function)
1 0.015 0.015 0.409 0.409 <pyshell#217>:1(test1)
1 0.000 0.000 0.409 0.409 <string>:1(<module>)
109989/9999 0.394 0.000 0.394 0.000 myMath.py:39(power)
1 0.000 0.000 0.000 0.000 {method 'disable' of
'_lsprof.Profiler' objects}



def power2(nbr, po):
acc=1
if po >= 1:
acc=nbr
for x in range(1, po):
acc=acc*nbr
if po < 0:
if nbr!=0:
acc=1
for x in range(0, po, -1):
acc=acc/nbr
else:
return "Division by zero"
return acc


20001 function calls in 0.095 CPU seconds

Ordered by: standard name

ncalls tottime percall cumtime percall
filename:lineno(function)
1 0.026 0.026 0.095 0.095 <pyshell#221>:1(test1)
1 0.000 0.000 0.095 0.095 <string>:1(<module>)
9999 0.051 0.000 0.069 0.000 myMath.py:47(power2)
1 0.000 0.000 0.000 0.000 {method 'disable' of
'_lsprof.Profiler' objects}
9999 0.017 0.000 0.017 0.000 {range}

 

[Terry Reedy]

| the 0.409 vs 0.095 is the total times right?
| so the imperative function is >4 times faster than the recursive.
| or what does tottime stand for?
|
| is this always the case that the recursive function is slower?
| the gain is less code?
|
| are some functions only implementable recursively?

Computers can be viewed as linear iteration machines implementing

while True:
execute next instruction and set next pointer appropriately

But algorithms most naturally expressed with multiple recursion usually
look more ungainly when linearized.