cudaMalloc vs VBO

Hey guys , i would like to know if is there any difference between cudaMalloc and VBO when it comes to huge data on card allocation and then move back to HOST memory…

i mean i need to use memory which will be handeled and calculated over the GPU card but , printed(drawn) over gpu using the CPU proc,

should i use VBO alloction and then move back to host mem

or use old cudamemcopy and cudamalloc?