search

PFZheng / psutil

Automatically exported from code.google.com/p/psutil
Other
0 stars 0 forks source link

cpu_percent() always returning 100% on osx #120

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
My machine is essentially idle, but cpu_percent() always returns exactly 100%.

Here is copy/paste from terminal:

Python 2.6.1 (r261:67515, Feb 11 2010, 00:51:29) 
[GCC 4.2.1 (Apple Inc. build 5646)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import psutil
>>> psutil.cpu_percent()
100.0
>>> 

What version of psutil are you using? On what operating system?
using psutil-0.1.3

Installed from source. Used the command:
python setup.py install 
--prefix=${HOME}/Library/Frameworks/Python.framework/Versions/2.6

Please provide any additional information below.

Not sure if it helps, but here is the output from when I installed it:
ph-jeffkline-l01:psutil-0.1.3 kline$ python setup.py install 
--prefix=${HOME}/Library/Frameworks/Python.framework/Versions/2.6
running install
running build
running build_py
creating build
creating build/lib.macosx-10.6-universal-2.6
creating build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/__init__.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/_psbsd.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/_pslinux.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/_psmswindows.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/_psosx.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/error.py -> build/lib.macosx-10.6-universal-2.6/psutil
copying psutil/wmi.py -> build/lib.macosx-10.6-universal-2.6/psutil
running build_ext
building '_psutil_osx' extension
creating build/temp.macosx-10.6-universal-2.6
creating build/temp.macosx-10.6-universal-2.6/psutil
creating build/temp.macosx-10.6-universal-2.6/psutil/arch
creating build/temp.macosx-10.6-universal-2.6/psutil/arch/osx
gcc-4.2 -fno-strict-aliasing -fno-common -dynamic -DNDEBUG -g -fwrapv -Os -Wall 
-Wstrict-prototypes -DENABLE_DTRACE -arch i386 -arch ppc -arch x86_64 -pipe 
-I/System/Library/Frameworks/Python.framework/Versions/2.6/include/python2.6 -c 
psutil/_psutil_osx.c -o 
build/temp.macosx-10.6-universal-2.6/psutil/_psutil_osx.o
In file included from psutil/_psutil_osx.c:21:
/usr/include/mach/shared_memory_server.h:48:2: warning: #warning 
"<mach/shared_memory_server.h> is deprecated.  Please use 
<mach/shared_region.h> instead."
psutil/_psutil_osx.c: In function ‘pid_exists’:
psutil/_psutil_osx.c:182: warning: unused variable ‘idx’
psutil/_psutil_osx.c:181: warning: unused variable ‘num_processes’
psutil/_psutil_osx.c:180: warning: unused variable ‘procList’
psutil/_psutil_osx.c: In function ‘get_process_cpu_times’:
psutil/_psutil_osx.c:305: warning: unused variable ‘thread_list’
psutil/_psutil_osx.c:304: warning: unused variable ‘thread_count’
psutil/_psutil_osx.c:303: warning: unused variable ‘t_cpu’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime_ms’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime_ms’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime’
psutil/_psutil_osx.c: In function ‘get_memory_info’:
psutil/_psutil_osx.c:451: warning: implicit declaration of function 
‘vm_region_64’
psutil/_psutil_osx.c:418: warning: unused variable ‘system_time’
psutil/_psutil_osx.c:418: warning: unused variable ‘user_time’
psutil/_psutil_osx.c: In function ‘get_avail_phymem’:
psutil/_psutil_osx.c:512: warning: implicit declaration of function 
‘host_statistics’
psutil/_psutil_osx.c:516: warning: implicit declaration of function 
‘mach_error_string’
psutil/_psutil_osx.c:516: warning: format ‘%s’ expects type ‘char *’, 
but argument 3 has type ‘int’
In file included from psutil/_psutil_osx.c:21:
/usr/include/mach/shared_memory_server.h:48:2: warning: #warning 
"<mach/shared_memory_server.h> is deprecated.  Please use 
<mach/shared_region.h> instead."
psutil/_psutil_osx.c: In function ‘pid_exists’:
psutil/_psutil_osx.c:182: warning: unused variable ‘idx’
psutil/_psutil_osx.c:181: warning: unused variable ‘num_processes’
psutil/_psutil_osx.c:180: warning: unused variable ‘procList’
psutil/_psutil_osx.c: In function ‘get_process_cpu_times’:
psutil/_psutil_osx.c:305: warning: unused variable ‘thread_list’
psutil/_psutil_osx.c:304: warning: unused variable ‘thread_count’
psutil/_psutil_osx.c:303: warning: unused variable ‘t_cpu’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime_ms’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime_ms’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime’
psutil/_psutil_osx.c: In function ‘get_memory_info’:
psutil/_psutil_osx.c:451: warning: implicit declaration of function 
‘vm_region_64’
psutil/_psutil_osx.c:418: warning: unused variable ‘system_time’
psutil/_psutil_osx.c:418: warning: unused variable ‘user_time’
psutil/_psutil_osx.c: In function ‘get_avail_phymem’:
psutil/_psutil_osx.c:512: warning: implicit declaration of function 
‘host_statistics’
psutil/_psutil_osx.c:516: warning: implicit declaration of function 
‘mach_error_string’
psutil/_psutil_osx.c:516: warning: format ‘%s’ expects type ‘char *’, 
but argument 3 has type ‘int’
In file included from psutil/_psutil_osx.c:21:
/usr/include/mach/shared_memory_server.h:48:2: warning: #warning 
"<mach/shared_memory_server.h> is deprecated.  Please use 
<mach/shared_region.h> instead."
psutil/_psutil_osx.c: In function ‘pid_exists’:
psutil/_psutil_osx.c:182: warning: unused variable ‘idx’
psutil/_psutil_osx.c:181: warning: unused variable ‘num_processes’
psutil/_psutil_osx.c:180: warning: unused variable ‘procList’
psutil/_psutil_osx.c: In function ‘get_process_cpu_times’:
psutil/_psutil_osx.c:307: warning: cast from pointer to integer of different 
size
psutil/_psutil_osx.c:305: warning: unused variable ‘thread_list’
psutil/_psutil_osx.c:304: warning: unused variable ‘thread_count’
psutil/_psutil_osx.c:303: warning: unused variable ‘t_cpu’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime_ms’
psutil/_psutil_osx.c:302: warning: unused variable ‘t_stime’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime_ms’
psutil/_psutil_osx.c:301: warning: unused variable ‘t_utime’
psutil/_psutil_osx.c: In function ‘get_memory_info’:
psutil/_psutil_osx.c:417: warning: cast from pointer to integer of different 
size
psutil/_psutil_osx.c:451: warning: implicit declaration of function 
‘vm_region_64’
psutil/_psutil_osx.c:418: warning: unused variable ‘system_time’
psutil/_psutil_osx.c:418: warning: unused variable ‘user_time’
psutil/_psutil_osx.c: In function ‘get_avail_phymem’:
psutil/_psutil_osx.c:512: warning: implicit declaration of function 
‘host_statistics’
psutil/_psutil_osx.c:516: warning: implicit declaration of function 
‘mach_error_string’
psutil/_psutil_osx.c:516: warning: format ‘%s’ expects type ‘char *’, 
but argument 3 has type ‘int’
gcc-4.2 -fno-strict-aliasing -fno-common -dynamic -DNDEBUG -g -fwrapv -Os -Wall 
-Wstrict-prototypes -DENABLE_DTRACE -arch i386 -arch ppc -arch x86_64 -pipe 
-I/System/Library/Frameworks/Python.framework/Versions/2.6/include/python2.6 -c 
psutil/arch/osx/process_info.c -o 
build/temp.macosx-10.6-universal-2.6/psutil/arch/osx/process_info.o
gcc-4.2 -Wl,-F. -bundle -undefined dynamic_lookup -arch i386 -arch ppc -arch 
x86_64 build/temp.macosx-10.6-universal-2.6/psutil/_psutil_osx.o 
build/temp.macosx-10.6-universal-2.6/psutil/arch/osx/process_info.o -o 
build/lib.macosx-10.6-universal-2.6/_psutil_osx.so
running install_lib
copying build/lib.macosx-10.6-universal-2.6/_psutil_osx.so -> 
/Users/kline/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site
-packages
running install_egg_info
Removing 
/Users/kline/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site
-packages/psutil-0.1.3-py2.6.egg-info
Writing 
/Users/kline/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site
-packages/psutil-0.1.3-py2.6.egg-info

Original issue reported on code.google.com by jeffery....@gmail.com on 22 Oct 2010 at 9:08

GoogleCodeExporter commented 9 years ago

Original comment by g.rodola on 22 Oct 2010 at 9:15

GoogleCodeExporter commented 9 years ago 
Is that a 64 bit system?

Original comment by g.rodola on 22 Oct 2010 at 9:15

GoogleCodeExporter commented 9 years ago 
Yes, here is output from system profiler:

  Model Name:   MacBook Pro
  Model Identifier: MacBookPro6,2
  Processor Name:   Intel Core i7
  Processor Speed:  2.66 GHz
  Number Of Processors: 1
  Total Number Of Cores:    2

Original comment by jeffery....@gmail.com on 22 Oct 2010 at 9:25

GoogleCodeExporter commented 9 years ago 
Can you show the output of the below command as well? 

>>> print psutil.cpu_times(); time.sleep(5); print psutil.cpu_times()
system=3.395108; idle=37.254002; user=4.461708; nice=0.0
system=3.395162; idle=37.254894; user=4.461761; nice=0.0

I'm getting similar results here (not 100% but close, 99.99% which probably 
means the issue is the same). Giampaolo, can you check this on any other 
platforms you have handy and see if this is OS X only? I'd like to figure out 
if this is in the C code or the Python code that handles the CPU calculations.

Thanks

Original comment by jlo...@gmail.com on 22 Oct 2010 at 9:35

GoogleCodeExporter commented 9 years ago 
Here is my output:

>>> import psutil, time
>>> print psutil.cpu_times(); time.sleep(5); print psutil.cpu_times()
system=1e-06; idle=4294.967295; user=2.112688; nice=0.0
system=1e-06; idle=4294.967295; user=2.112688; nice=0.0
>>>

Original comment by jeffery....@gmail.com on 22 Oct 2010 at 9:42

GoogleCodeExporter commented 9 years ago 
I'll have to look at this some more later but I think the 
get_system_cpu_times() is not returning valid results for some reason: 

>>> _psutil_osx.get_system_cpu_times(); time.sleep(5);  
_psutil_osx.get_system_cpu_times()
(4.4710289999999997, 0.0, 3.4054090000000001, 37.401888)
(4.4710669999999997, 0.0, 3.4054660000000001, 37.402794999999998)

After 5 seconds it's giving almost identical values, and I suspect any 
differences in the values are probably from floating point math...

Original comment by jlo...@gmail.com on 22 Oct 2010 at 9:49

GoogleCodeExporter commented 9 years ago 
Linux 64bit, 2 cores:

giampaolo@ubuntu:~/svn/psutil$ python -c "import psutil, time; print 
psutil.cpu_times(); time.sleep(5); print psutil.cpu_times()"
softirq=33.91; iowait=1355.06; system=3944.62; idle=73988.76; user=5297.51; 
irq=0.01; nice=339.7
softirq=33.91; iowait=1355.06; system=3945.14; idle=73997.73; user=5298.01; 
irq=0.01; nice=339.72

FreeBSD 7.0 32bit, 1 core:

[root@freebsd ~/psutil]# python -c "import psutil, time; print 
psutil.cpu_times(); time.sleep(5); print psutil.cpu_times()"
system=696.9296875; irq=7197.453125; idle=32646.4296875; user=81.4296875; 
nice=0.0078125
system=696.984375; irq=7200.296875; idle=32646.4296875; user=83.9765625; 
nice=0.0078125

Windows 7 64 (py 32 bit), 1 core:

C:\Users\user\Desktop\svn\psutil>C:\python27\python -c "import psutil, time; 
print psutil.cpu_times(); time.sleep(5); print psutil.cpu_times()"
idle=16220.328125; system=1174.375; user=81.71875
idle=16224.90625; system=1174.796875; user=81.734375

Original comment by g.rodola on 22 Oct 2010 at 10:34

GoogleCodeExporter commented 9 years ago 
Thanks giampaolo, what about cpu_percent, does it look accurate?

Original comment by jlo...@gmail.com on 23 Oct 2010 at 1:12

GoogleCodeExporter commented 9 years ago

Original comment by jlo...@gmail.com on 23 Oct 2010 at 1:12

GoogleCodeExporter commented 9 years ago 
Aside from an initial "100%" which perhaps we could adjust someway results look 
ok and match top/taskmgr on BSD/Linux/Windows on both single and double CPUs.
The script:

import time, psutil, threading

def foo():
    while 1:
        pass
t = threading.Thread(target=foo).start()

while 1:
    print psutil.cpu_percent()
    time.sleep(1)

In r724 I added a test based on the results you pasted in msg #4 which should 
fail on OSX so you can at least use that to test results.

Original comment by g.rodola on 23 Oct 2010 at 10:52

GoogleCodeExporter commented 9 years ago 
> In r724 I added a test based on the results you pasted in msg #4 which 
> should fail on OSX so you can at least use that to test results.

Thanks... I took a look at this some more last night but I am still working on 
it. 

I can see the problem (times not updating properly) but I don't know if the 
system call is just returning strange data or if something has changed in 10.6 
in the way statistics are reported/calculated.

Original comment by jlo...@gmail.com on 23 Oct 2010 at 3:50

GoogleCodeExporter commented 9 years ago

Original comment by jlo...@gmail.com on 23 Oct 2010 at 3:50

GoogleCodeExporter commented 9 years ago 
Point is times are updated, but not as much as expected, so I agree with you 
that this is likely to be a math problem. Maybe we shouldn't divide per 
CLOCKS_PER_SEC as we do on BSD:

- (double)r_load.cpu_ticks[CPU_STATE_USER] / CLOCKS_PER_SEC,
+ (double)r_load.cpu_ticks[CPU_STATE_USER],

Original comment by g.rodola on 23 Oct 2010 at 5:02

GoogleCodeExporter commented 9 years ago 
I suspect it's a math problem but I don't know why, or what to use instead of 
CLOCKS_PER_SEC... we can't just return it as is because it'd be returning clock 
ticks for the values, and our math in the cpu_percent() function depends on 
seconds. 

In theory, dividing by CLOCKS_PER_SEC is the correct thing to do but it's 
evidently not working correctly, so something is out of whack :-/

Original comment by jlo...@gmail.com on 23 Oct 2010 at 5:17

GoogleCodeExporter commented 9 years ago 
Should be fixed in r725 now. I've no idea what changed, if anything, but 
CLOCKS_PER_SEC (1000000) isn't working correctly and OS X 10.6 seems to be 
expecting you to divide CLK_TCK (100) instead. Replacing the division with 
CLK_TCK I'm getting results that match Activity Monitor when using the 
following test: 

python -c 'import time,psutil; time.sleep(5); print psutil.cpu_percent(); 
time.sleep(5); print psutil.cpu_percent()'

Original comment by jlo...@gmail.com on 23 Oct 2010 at 6:08

GoogleCodeExporter commented 9 years ago

Original comment by g.rodola on 23 Oct 2010 at 6:45

GoogleCodeExporter commented 9 years ago 
Using the svn just checked out a few minutes ago, I still receive the 
following output:

python -c "import psutil as p, time;  print p.__version__; print 
p.cpu_percent(); time.sleep(5); print p.cpu_percent()"
0.2.0
100.0
100.0

Is there a minimal example I could run  that will expose the low-level behavior?

Original comment by jeffery....@gmail.com on 25 Oct 2010 at 6:27

GoogleCodeExporter commented 9 years ago 
> Is there a minimal example I could run  that will expose the low-level 
behavior?

You could try to run this and paste the output:

import psutil, time
while 1:
    print sum(psutil.cpu_times())
    time.sleep(1)

On my Linux box (2 CPUs) it prints:

1999.74
2001.74
2003.74
2005.75
2007.76
2009.76
2011.76
2013.76
2015.76
2017.76
2019.76
...

Original comment by g.rodola on 25 Oct 2010 at 6:38

GoogleCodeExporter commented 9 years ago 
You can call the cpu times function directly from the _psutil_osx module: 

python -c 'import _psutil_osx,time; print _psutil_osx.get_system_cpu_times(); 
time.sleep(5); print _psutil_osx.get_system_cpu_times()'
(68604.160000000003, 0.0, 47660.730000000003, 490471.07000000001)
(68604.639999999999, 0.0, 47661.330000000002, 490479.98999999999)

Original comment by jlo...@gmail.com on 25 Oct 2010 at 6:53

GoogleCodeExporter commented 9 years ago 
Reply to comments 18 and 19:

>>> import psutil, time
>>> while 1:
...     print sum(psutil.cpu_times())
...     time.sleep(1)
... 
46475.37
16110926.24
16110926.24
16110926.24
16110926.24
16110926.24

 python -c 'import _psutil_osx,time; print _psutil_osx.get_system_cpu_times(); time.sleep(5); print _psutil_osx.get_system_cpu_times()'
(16064133.439999999, 0.01, 327.67000000000002, 45964.080000000002)
(0.050000000000000003, 0.01, 0.0, 45964.080000000002)

Original comment by jeffery....@gmail.com on 25 Oct 2010 at 7:31

GoogleCodeExporter commented 9 years ago 
I'm not sure what to make of the user time in your example above (the first 
value) it looks like it rolled over. The idle time is what we use to calculate 
CPU utilization and you can see in the example that your idle time is the same 
in both iterations. If the idle is the same then you're going to see 100% 
utilization. 

Can you run this and post the output?

python -c "import _psutil_osx,time
for i in range(0, 25):
    time.sleep(2)
    print _psutil_osx.get_system_cpu_times()"

I'd like to see what your idle time looks like over a longer period.

Thanks

Original comment by jlo...@gmail.com on 25 Oct 2010 at 11:28

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
This is a second try to post - apologies if this is a repeat post.

Here is the output from
 python -c "import _psutil_osx,time
> for i in range(0, 25):
>    time.sleep(2)
>    print _psutil_osx.get_system_cpu_times()"
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)
(0.02, 0.01, 0.0, 47930.720000000001)

Original comment by jeffery....@gmail.com on 26 Oct 2010 at 4:27

GoogleCodeExporter commented 9 years ago 
Interesting... your CPU times are reporting identical results every time. It's 
no wonder CPU util is showing 100% then. I'm kind of at a loss, the cpu times 
are being pulled directly from a mach host_statistics() call and the only 
calculation/manipulation done is to divide it by CLK_TCK. It looks to me like 
the host_statics call is returning the same values every time.

I'll put together a small C app you can try it with and see what happens if we 
print the raw idle value from host_statistics()

Original comment by jlo...@gmail.com on 26 Oct 2010 at 4:48

GoogleCodeExporter commented 9 years ago 
Ok r739 now fixes some warnings and adds error checking around 
host_statistics() in get_system_cpu_times() - while making a test C app I 
noticed that if host_statistics() is failing the CPU times struct can still 
return values. I suspect what may have been happening in your case is the call 
to host_statistics is failing and thus not updating the idle time on subsequent 
runs. You can try the newest SVN version and see if that makes any difference. 

I've also attached a cpu_idle.c file you can compile with just "gcc cpu_idle.c" 
 if you've got the commandline xcode/gcc tools installed. Then just run ./a.out 
and you should see something like this, updated every two seconds: 

[user@host ]$ ./a.out 
idle=55495234, idle/CLK_TCK=554952.340000
idle=55495585, idle/CLK_TCK=554955.850000
idle=55495925, idle/CLK_TCK=554959.250000
idle=55496269, idle/CLK_TCK=554962.690000
idle=55496604, idle/CLK_TCK=554966.040000
idle=55496927, idle/CLK_TCK=554969.270000

To stop just use CTRL C 

Thanks, 

-Jay

Original comment by jlo...@gmail.com on 26 Oct 2010 at 5:19

Attachments:

GoogleCodeExporter commented 9 years ago 
Here's some output from cpu_idle.c:

gmail:Downloads kline$ ./cpu_idle 
idle=175914686, idle/CLK_TCK=1759146.860000
idle=175915425, idle/CLK_TCK=1759154.250000
idle=175916170, idle/CLK_TCK=1759161.700000
idle=175916911, idle/CLK_TCK=1759169.110000
idle=175917680, idle/CLK_TCK=1759176.800000

Original comment by jeffery....@gmail.com on 26 Oct 2010 at 5:56

GoogleCodeExporter commented 9 years ago 
Okay -  with the latest svn, the output from the code of Comment 21 follows. It 
is different from before:

 python -c "import _psutil_osx,timefor i in range(0, 25):
    time.sleep(2)
    print _psutil_osx.get_system_cpu_times()"
(49156.760000000002, 0.0, 37154.800000000003, 1760902.1299999999)
(49156.980000000003, 0.0, 37155.07, 1760909.6000000001)
(49157.220000000001, 0.0, 37155.349999999999, 1760917.0900000001)
(49157.489999999998, 0.0, 37155.669999999998, 1760924.53)
(49157.720000000001, 0.0, 37155.940000000002, 1760931.97)
(49158.050000000003, 0.0, 37156.349999999999, 1760939.3)
(49158.279999999999, 0.0, 37156.769999999997, 1760946.5600000001)

Output from cpu_percent() is unchanged from before:

 python -c "import psutil; print psutil.cpu_percent()"
100.0

Original comment by jeffery....@gmail.com on 26 Oct 2010 at 6:07

GoogleCodeExporter commented 9 years ago 
> Output from cpu_percent() is unchanged from before:
>  python -c "import psutil; print psutil.cpu_percent()"
> 100.0

Bear in mind you have to have the time.sleep() in there, since the CPU util 
calculation is performed by comparing the amount of idle time reported against 
the amount of clock time reported. The cpu idle values look to be incrementing 
nicely now... can you try 

python -c 'import time,psutil; time.sleep(5); print psutil.cpu_percent(); 
time.sleep(5); print psutil.cpu_percent()'

And see if you get something accurate as compared to Activity Monitor? 

thanks!

Original comment by jlo...@gmail.com on 26 Oct 2010 at 6:29

GoogleCodeExporter commented 9 years ago 
Yes, that's it! Thank you!

Since cpu_percent is often paired with a 'time.sleep' call, I wonder if it 
would be simpler to have cpu_percent accept a "sleep" argument with default 
behavior of, say, 2s.

Original comment by jeffery....@gmail.com on 26 Oct 2010 at 6:44

GoogleCodeExporter commented 9 years ago 
Possibly that would make sense.. at the moment what we do is as soon as you 
import the module we start the clock so to speak, so as long as you wait a 
tenth of a second or more before calling the cpu_percent() code it should be 
able to calculate the CPU percentage. The longer you wait between calls the 
more data we have to work with so it'll be more accurate.

I'll discuss the sleep param with Giampaolo and see if we want to go that 
route, in the meantime just be sure to give it a little time between calls, 
even a tenth of a second or two should be enough.

Original comment by jlo...@gmail.com on 26 Oct 2010 at 7:08

GoogleCodeExporter commented 9 years ago

Original comment by jlo...@gmail.com on 27 Oct 2010 at 3:39

GoogleCodeExporter commented 9 years ago

Original comment by g.rodola on 13 Nov 2010 at 3:14

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Updated csets after the SVN -> Mercurial migration:
r261 == revision 7bda2b4aa692
r724 == revision f621399fbd84
r725 == revision 8be414279d25
r739 == revision 2ba734d2027c

Original comment by g.rodola on 2 Mar 2013 at 11:55