Skip to content

rai-project/go-cupti

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

go-cupti

Build Status

Go binding to NVIDIA CUPTI, the CUDA Performance Tool Interface.

Example

The callback functions are publised to a tracing server. You need to have a carml_config.yml and the tracing server running to see the spans. Refer to CarML Config and Starting Tracer Server

Make sure /usr/local/cuda/lib64 and /usr/local/cuda/extras/CUPTI/lib64 are in your LD_LIBRARY_PATH.

cd examples/vector_add
make
cd ../cupti
go run main.go

Then go to TRACER_URL:16686 to see the spans

Note: The CGO interface passes go pointers to the C API. This is an error by the CGO runtime. If you get the following error

panic: runtime error: cgo argument has Go pointer to Go pointer

Then you need to place

export GODEBUG=cgocheck=0

in your ~/.bashrc or ~/.zshrc file and then run either source ~/.bashrc or source ~/.zshrc.

References

The CUDA Profiling Tools Interface (CUPTI) enables the creation of profiling and tracing tools that target CUDA applications. CUPTI provides four APIs: the Activity API, the Callback API, the Event API, and the Metric API.

  1. The CUPTI Activity API allows you to asynchronously collect a trace of an application's CPU and GPU CUDA activity.
  • Activity Record. CPU and GPU activity is reported in C data structures called activity records.
  • Activity Buffer. An activity buffer is used to transfer one or more activity records from CUPTI to the client. An asynchronous buffering API is implemented by cuptiActivityRegisterCallbacks and cuptiActivityFlushAll.
  1. The CUPTI Callback API allows you to register a callback into your own code. Your callback will be invoked when the application being profiled calls a CUDA runtime or driver function, or when certain events occur in the CUDA driver.
  • Callback Domain.
  • Callback ID. Each callback is given a unique ID within the corresponding callback domain so that you can identify it within your callback function.
  • Callback Function Your callback function must be of type CUpti_CallbackFunc. This function type has two arguments that specify the callback domain and ID so that you know why the callback is occurring. The type also has a cbdata argument that is used to pass data specific to the callback.
  • Subscriber. A subscriber is used to associate each of your callback functions with one or more CUDA API functions. There can be at most one subscriber initialized with cuptiSubscribe() at any time. Before initializing a new subscriber, the existing subscriber must be finalized with cuptiUnsubscribe().
  1. The CUPTI Event API allows you to query, configure, start, stop, and read the event counters on a CUDA-enabled device.
  2. The CUPTI Metric API allows you to collect application metrics calculated from one or more event values.

Releases

No releases published

Packages

No packages published

Languages