Re: [Boost-users] [ Interprocess ] Object allocation performance

Subject: Re: [Boost-users] [ Interprocess ] Object allocation performance
From: Ion Gaztañaga (igaztanaga_at_[hidden])
Date: 2011-10-04 12:28:13

El 03/10/2011 18:35, Yordan Pavlov escribió:
> It looks like this:
>
> |MyType  *  pMy=  my_segment->construct<MyType>(anonymous_instance)();
>
> my_segment->destroy_ptr(pMy);
> |
>
> Where *MyType* is some typical struct and *my_segment* is correctly
> constructed *boost::interprocess::managed_shared_memory ** is around 10
> times slower than the equivalent:
>
> |MyType  *  pMy=  new  MyType();
>
> delete  pMy;
> |
>
> I did not expect this. I though the two allocation algorithms should be
> similar in implementation and performance. Is there some good reason for
> such a huge difference.

I don't expect my algorithm to be extremely fast as you can't use the 
same techniques in heap allocation or shared memory allocation, shared 
memory allocation is intended to be more space-concerned than heap 
allocation, where you can preallocate big chunks of OS memory and 
implement pools. Shared memory allocation is designed avoid any memory 
waste or preallocation. If you need a huge number of objects then think 
about using arrays or vectors.

Best,

Ion