On May 27, 8:27 am, rahul wrote:
> How is the memory allocated for structures? I need to optimize the
> memory usage and bit fields are not doing the trick.
>
> Any details about the memory allocation for the structures would be a
> great help.

More context would be a great help too.
Besides padding and memory alignment:

http://en.wikipedia.org/wiki/Data_structure_alignment

There isn't much I can think of without more context (and even better,
code).

-- paulo

On May 27, 5:27 am, rahul wrote:
> How is the memory allocated for structures? I need to optimize the
> memory usage and bit fields are not doing the trick.
>
> Any details about the memory allocation for the structures would be a
> great help.

The best reference will be the C standard ... try getting a copy of
that as well if you want to know what is guaranteed by the language
and what is system dependent in this regard.

Ivan Novick

In article
<7e91d38b-4728-44a7-ba17-b10a14816ecb@t12g2000prg.googlegroups.com>,
ivan@novickmail.com wrote:

> On May 27, 5:27 am, rahul wrote:
> > How is the memory allocated for structures? I need to optimize the
> > memory usage and bit fields are not doing the trick.
> >
> > Any details about the memory allocation for the structures would be a
> > great help.
>
> The best reference will be the C standard ... try getting a copy of
> that as well if you want to know what is guaranteed by the language
> and what is system dependent in this regard.

I think the C specification leaves most of this
implementation-dependent, so it's not going to be much help for him.

--
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***

>How is the memory allocated for structures? I need to optimize the
>memory usage and bit fields are not doing the trick.
>
>Any details about the memory allocation for the structures would be a
>great help.

Structure members are allocated *in order*. Any member may have
padding after it (thus, there may be no padding before the first
member). That's it for what the C Standard guarantees. Padding
is *internal*, that is, it counts as part of sizeof(struct whatever).

Although not guaranteed by the C Standard, usually padding is added
only if the next element would be misaligned.

One strategy to minimize memory taken by a structure is to order
members with the likely strictest alignment first:

long double
double
long long
pointers
long
float
int
short
char

Now, the above list is a guess based on typical implementations,
and it will vary somewhat with the implementation. Because of size
limits in the C Standard, (and alignment often requires alignment
to a multiple of the size, but the C Standard does not require
that), the ordering:

long double >= double >= float
and
long long >= long >= int >= short >= char
will likely stay put.

Gordon Burditt wrote:
>[... good stuff ...]
>
> alignment often requires alignment
> to a multiple of the size, but the C Standard does not require
> that [...]

The alignment requirement for a type T must be a *divisor*
of sizeof(T) -- you can see this by considering the addresses
of the two elements of `T array[2]'.

So Gordon's right, sort of: C does not require that the
alignment be a multiple of the sizeof, but in the case where
it does happen to be a multiple the multiplier must be unity!

--
Eric Sosman
esosman@ieee-dot-org.invalid