Bureaucrats, cc_docs_admin, cc_staff
337
edits
Nvprof: Difference between revisions
Jump to navigation
Jump to search
no edit summary
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
Nvprof is a command-line light-weight GUI-less profiler available for Linux, Windows, and Mac OS. | Nvprof is a command-line light-weight GUI-less profiler available for Linux, Windows, and Mac OS. | ||
This tool allows you to collect and view profiling data of CUDA-related activities on both CPU and GPU, including kernel execution, memory transfers, etc. Profiling options should be provided to the profiler via the command-line options. | |||
== Strengths == | |||
It is capable of providing a textual report : | It is capable of providing a textual report : | ||
* Summary of GPU and CPU activity | * Summary of GPU and CPU activity | ||
| Line 23: | Line 26: | ||
Use <code>module load cuda/version</code> to choose a version. For example, to load the CUDA compiler version 10.0, do: | Use <code>module load cuda/version</code> to choose a version. For example, to load the CUDA compiler version 10.0, do: | ||
{{Command|module load cuda/10.0}} | {{Command|module load cuda/10.0}} | ||
== Compile your code == | |||
To get useful information from Nvprof, you first need to compile your code with one of the Cuda compilers (<code>nvcc</code> for C). | |||