Created on 2010-11-09 19:15 by exarkun, last changed 2010-11-20 17:09 by orsenthil. This issue is now closed.
| Messages (3) | |||
|---|---|---|---|
| msg120890 - (view) | Author: Jean-Paul Calderone (exarkun) *  |
Date: 2010-11-09 19:15 | |
Consider this transcript:
>>> cProfile.run("import time; time.sleep(1)")
4 function calls in 1.012 CPU seconds
Ordered by: standard name
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.011 0.011 1.012 1.012 <string>:1(<module>)
1 0.000 0.000 1.012 1.012 {built-in method exec}
1 1.001 1.001 1.001 1.001 {built-in method sleep}
1 0.000 0.000 0.000 0.000 {method 'disable' of '_lsprof.Profiler' objects}
>>>
It is not the case that the profiled code uses >1 CPU seconds. It spends the entire time sleeping. The default timer for cProfile is a wallclock timer. The output should reflect this.
|
|||
| msg121101 - (view) | Author: Senthil Kumaran (orsenthil) * |
Date: 2010-11-13 01:18 | |
On Tue, Nov 09, 2010 at 07:15:57PM +0000, Jean-Paul Calderone wrote:
> >>> cProfile.run("import time; time.sleep(1)")
> 4 function calls in 1.012 CPU seconds
>
> It is not the case that the profiled code uses >1 CPU seconds. It
> spends the entire time sleeping. The default timer for cProfile is
> a wallclock timer. The output should reflect this.
So, It should just be 1.012 seconds, instead of 1.012 CPU seconds. Correct?
Or is there any other technically more accurate way to say?
|
|||
| msg121685 - (view) | Author: Senthil Kumaran (orsenthil) * |
Date: 2010-11-20 17:09 | |
Fixed in r86580 (py3k), r86581 and r86582. |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2010-11-20 17:09:31 | orsenthil | set | status: open -> closed messages: + msg121685 resolution: fixed components: + Library (Lib) |
| 2010-11-13 01:18:06 | orsenthil | set | nosy:
+ orsenthil messages: + msg121101 |
| 2010-11-11 15:23:23 | eric.araujo | set | versions: - Python 2.6 |
| 2010-11-09 19:15:52 | exarkun | create | |