Does CUDA suffer from memory free list fragmentation? May there be a situation were I fail to allocate N byte, but success allocating two times N/2 byte?
– Kuisma
Does CUDA suffer from memory free list fragmentation? May there be a situation were I fail to allocate N byte, but success allocating two times N/2 byte?
– Kuisma
Any comment about this? I want to know that too!