is +=1 thread safe

[Gary Herron]

Alexander Schmolck wrote:
> Gary Herron <gherron at islandtraining.com> writes:
>
>   
>> But... It's not!  
>>
>> A simple test shows that.   I've attached a tiny test program that shows this
>> extremely clearly.  Please run it and watch it fail.
>>     
>
> In [7]: run ~/tmp/t.py
> final count: 2000000
>   should be: 2000000
>
> (I took the liberty to correct your test to actually do what I said, namely
> use a numpy.array; just replace ``count = 0`` with ``import numpy; count =
> numpy.array(0)``).
>   

The test was meant to simulate the OP's problem, but even with your 
suggestion of using numpy, it *still* fails!  Just start increasing the 
number of increments by a factor of 2, 4, 10 until it fails:

final count: 1999998
  should be: 2000000

final count: 5999997
  should be: 6000000

final count: 5995068
  should be: 6000000

(Surprisingly, using numpy makes this test *much* slower, meaning the 
increment executes many more instructions, which implies hitting a 
thread context switch at exactly the critical point is much less 
common.   But it can happen if you run the test long enough.)

I am running this on a Core2 Duo CPU, but the GIL should prevent that 
from affecting the result of the run.

I've reattached the file with the numpy change (and a larger loop 
counter) for others to try.

Gary Herron

>
> 'as
>
> p.s. please don't send me copies to on-list replies, at least not without
> explicitly mentioning the fact -- I've got better things to do then guessing
> whether something was meant to be off-list or not.
>   
My apologies.
> --
> http://mail.python.org/mailman/listinfo/python-list
>   

-------------- next part --------------
A non-text attachment was scrubbed...
Name: t.py
Type: application/python
Size: 525 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20080504/65db1ba0/attachment-0001.bin>