global memory coalescing using nvprof

Hello,
I’m usng nvprof metrics to characterize CUDA workloads. Is there a way to measure the global memory coalescing using the metrics that are reported by nvprof command line profiler?