I tried running nvbit on the SimpleStream benchmark found under the NVIDIA SDK (cuda-samples/Samples/0_Introduction/simpleStreams.cu). However, the stream id dumped by the tool seems unusually large. Any pointers as to what could be causing this issue?
Here's an example of running the mem trace tool (which dumps the stream id):
I tried running nvbit on the SimpleStream benchmark found under the NVIDIA SDK (cuda-samples/Samples/0_Introduction/simpleStreams.cu). However, the stream id dumped by the tool seems unusually large. Any pointers as to what could be causing this issue?
Here's an example of running the mem trace tool (which dumps the stream id):
The command I've used is:
LD_PRELOAD=./nvbit_release/tools/mem_trace/mem_trace.so ~/cuda-samples/Samples/0_Introduction/simpleStreams/simpleStreams