Re: realloc() - frequency, conditions, or experiences about relocation?

Am 20.06.2024 um 22:08 schrieb Vir Campestris:

On 17/06/2024 10:22, Bonita Montero wrote:

>
realloc() is just a convenience funciton. Usually the reallocation
can't happen in-place and a second malloc() followed by a copy and
a free() does the same.
For large data it would be nice if the pages being deallocated later
would be incrementally marked as discardable after copying a portion.
This would result in only a small portion of additional physical
memory being allocated since the newly allocated pages become asso-
ciated with phyiscal pages when they're touched first. Windows has
VirtualAlloc() with MEM_RESET for that, Linux has madvise() with
MADV_DONTNEED.

 "Usually can't happen in place"?

Usually you don't resize the block with a few bytes and if you have
a large memory area it's unlikely that you can allocate a few additional
pages in-place.
C++23 has a nice function with the allocator interface called allocate
_at_least. It returns the pointer to the allocated memory with the size
actually allocated. This fits with modern allocators like mimalloc, je-
malloc or tcmalloc, which round up the allocation to the next 2 ^ N
boundary up to 8kB. So a vector's capacity directly includes the blend.