MEMORY FRAGMENTATION

Discuss.

Discuss.

>>1
Can be alleviated by proper allocating.

>>2
Can be alleviated by proper haxing.

Memory, it gets fragmented.

allocate static amount of memory each time then

Compacting garbage collector solves it.

>>3

Can be alleviated by proper allocating.

Be more specific.

>>6
GC isn't shit?

>>7
For static allocation only, it's an optimization problem in NP somewhere. For dynamic allocation it's impossible to optimize completely the general case. If you remove dependance on user input I'm not sure whether it's doable... it might be reducible to the halting problem.

Compacting (moving) collectors are pretty rad, yeah. Note: using one moves the allocation strategy out of the general case, and into one that guarantees all pointers to allocated segments can be identified and therefore the segments can be moved. Normally the problem of identifying the pointers is halting equivalent (actually; it's worse without being able to further modify the program logic in other hard, possiby halting equivalent, ways) and can't be applied to the general dynamic case.

Do not allocate to a memory region if it is much larger than the requested allocation (unless all memory regions are much larger, in which case you allocate to the smallest one).

Problem solved.

>>9
That solution is called “living with memory fragmentation (and a naïve allocator)”

_{(stopping the dubsfaggot from dubsbumping)}