[Distutils] 64-bit vs 32-bit (Re: An observation on how system packagers and developers can be friends)

[Jim Fulton]

On Fri, Nov 5, 2010 at 11:55 AM, Mark Sienkiewicz <sienkiew at stsci.edu> wrote:
>
>> In the same direction, I wish people would understand that 64-bit Python
>> builds just waste memory, and that they were better off with 32-bit
>> implementations.
>
> There are certain irrational reasons to believe that "64 bits is better than
> 32", but there is also some logic to it.
>
> I support an application that can't run in a 32 bit python because there
> literally is not enough memory available.  It needs a larger address space.

Me too. I've found that on 32-bit linux systems, when programs get above
around 2G, mallocs of a few tens of megs will fail.

What's worse -- much much worse -- when a memory error occurs, it can
leave an application in an unpredictable state.  (Think memory errors
in the bowels of your database.)  These errors tend to lead to really
odd symptoms and be be hard to diagnose.

This is bad enough, but I really want to restart an app when a memory
error occurs, but MemoryError is a subclass of Exception so it tends
to be caught a lot.  I really wish MemoryError was one of those errors
you had to go way out of your way to catch.

Because of the havoc that can be caused by even infrequent memory
errors, we're moving most of our apps to 64 bits, even though they
rarely use more than 2G.

Jim

--
Jim Fulton