Dear All,
In order to get information of my application memory bandwidth between GPU chip
and Global Device memory on my GPU using pgprof command,
I typed “pgprof --query-metrics” and I found out as follows.
dram_read_throughput: Device memory read throughput
dram_write_throughput: Device memory write throughput
gst_throughput: Global memory store throughput
gld_throughput: Global memory load throughput
What is the difference between “dram_read_throughput” and “gld_throughput” ?
Which should I use to get the information ?
Sincerely yours,