Add HIP launch bounds

libtcc/kernel/TCCorrelator.cu

@@ -764,6 +764,7 @@ extern "C" __global__
 #ifdef __HIP__
 // ToDo: add HIP launch bounds. Second parameters is different:
 // min warps per execution unit instead of min blocks per SM
+__launch_bounds__(NR_WARPS * warpSize, 1)
 #else
 __launch_bounds__(NR_WARPS * 32, NR_RECEIVERS_PER_BLOCK == 32 ? 4 : 2)
 #endif