How many grids in CUDA

To elaborate on Robert's answer, here's an example of how you could use streams to make your two instances of Kernel1 run concurrently:

cudaStream_t stream1; cudaStreamCreate(&stream1);
cudaStream_t stream2; cudaStreamCreate(&stream2);

Kernel1<<gridDim, blockDim, 0, stream1>>(dst1, param1);
Kernel1<<gridDim, blockDim, 0, stream2>>(dst2, param2);

A few more notes about concurrent execution with streams:

• If we launch a kernel without specifying a stream Kernel1<<<g, b>>>(), and then launch a kernel with a specific stream Kernel2<<<g, b, 0, stream>>>(), then Kernel2 will wait for Kernel1 to finish.
• When a kernel is launched without a stream (Kernel1<<<g, b>>>()), Nvidia calls this "using the NULL stream."
• If you use cudaEvents, your work can sometimes get serialized even if you distribute the kernels over several streams.