I'm not sure. My colleague explained me that the main gain came from avoiding to do a malloc in kernel space and a data copy. But it's true also that many small

I'm not sure. My colleague explained me that the main gain came from avoiding to do a malloc in kernel space and a data copy. But it's true also that many small reads in a single block implies many system calls. I don't know the real relative processing cost of these operations.