Release 3.10.0 corresponding to NGC container 21.04

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.10.0

• Compatibility with NGC 21.04 release

Release 3.9.0 corresponding to NGC container 21.03

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.9.0

• Compatiblity with NGC 21.03 release

Release 3.8.0 corresponding to NGC container 21.02

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.8.0

• Torch.ones and Torch.zeros input can now be a list, tuple, or a sequence of integers

Release 3.7.0 corresponding to NGC container 20.12

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.7.0

• Monkey patching support for APEX libraries.

Release 3.6.0 corresponding to NGC container 20.11

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.6.0

• PyProf overhead was reduced to improve runtime performance
  • Improved database query from Nsight Systems
  • Refactored nvmarker.py

Release 3.5.0 corresponding to NGC container 20.10

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.5.0

• Nsight System database lookup improved to speed up the runtime profile analysis time by 50x.

• Node names will now include class info and can be linked back to the original Python source.

Release 3.4.0 corresponding to NGC container 20.09

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.4.0

• README and User Guide documentation has been updated with more installation
  options and pointers

Release 3.3.0 corresponding to NGC container 20.08

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.3.0

• Capture PyTorch API information and data loading configuration

• Added CUTLASS to the list of GEMM kernels

• Added optional function stack tracing to NVTX markers. Enable with

    pyprof.init(enable_function_stack=True)

Known Issues

• This software only supports PyTorch 1.6.

• Forward-Backward kernel correlation heuristics do not work correctly with PyTorch 1.6. Recommended work arounds include:

  • Use with PyTorch 1.5
  • Use DLProf in the 20.08 NGC PyTorch container
  • Use the 20.03-py3 PyTorch NGC container

 $ docker pull nvcr.io/nvidia/pytorch:20.03-py3

Release 3.2.0 corresponding to NGC container 20.07

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.2.0

• Monkey patch support for APEX, fused Adam, and Layer Norm functions

• PyYAML requirement has been removed

• Error handling for non-existent parse file arguments has been added

Known Issues

• This software only supports PyTorch 1.6.

• Forward-Backward kernel correlation heuristics do not work correctly with PyTorch 1.6. Recommended work arounds include:

  • Use with PyTorch 1.5
  • Use DLProf in the 20.078 NGC PyTorch container
  • Use the 20.03-py3 PyTorch NGC container

 $ docker pull nvcr.io/nvidia/pytorch:20.03-py3

Release 3.1.0 corresponding to NGC container 20.06

PyProf - PyTorch Profiler for GPUs

PyProf is a tool that profiles and analyzes the GPU performance of PyTorch models. PyProf aggregates kernel performance from Nsight Systems or NvProf. The tool can identify the layer that launched the kernel, identify the tensor dimensions and precision, perform forward-backward correlation, determine Tensor Core usage, and correlate the line in the user's code that launched a particular kernel.

What's New In 3.1.0

• Improved NVTX functionality and coverage.
• Add support required by NVIDIA Deep Learning Profiler.

Known Issues

• This software only supports PyTorch 1.6.