wait for the completeness of a given kernel only

I have kernel_1 and kernel_2
So, logically

call kernel_1
call kernel_2
wait_4_completeness_of_kernel_1
do_host_code_that_depend_on_kernel_1
wait for all kernel
do host code

cudaThreadSynchronize() just wait for the completeness of all kernel. Is it possible to tell the host code to wait for the completeness of a particular kernel, e.g. kernel_1.

Tuan