Is it possible to measure memory transfer time and kernel execution time in ITK + GPU?

Rafael_Palomar · July 13, 2021, 7:12pm

When benchmarking GPU algorithms it is important to measure kernel execution time separated from memory transfer times (CPU->GPU and GPU->CPU). Is it possible to measure these times separately for existing algorithms (e.g., GPUBinaryThresholdImageFilter?). Thanks in advance.

matt.mccormick · July 13, 2021, 8:01pm

Hi @Rafael_Palomar ,

Welcome to the ITK community!

To separate the memory transfer times, call UpdateBuffers() on the images.

Here is an example:

github.com

InsightSoftwareConsortium/ITK/blob/fb4fc0168751c21de4b085063272988c45ae9788/Modules/Filtering/GPUSmoothing/wrapping/test/itkGPUMeanImageFilterTest.py

# ==========================================================================
#
#   Copyright NumFOCUS
#
#   Licensed under the Apache License, Version 2.0 (the "License");
#   you may not use this file except in compliance with the License.
#   You may obtain a copy of the License at
#
#          http://www.apache.org/licenses/LICENSE-2.0.txt
#
#   Unless required by applicable law or agreed to in writing, software
#   distributed under the License is distributed on an "AS IS" BASIS,
#   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#   See the License for the specific language governing permissions and
#   limitations under the License.
#
# ==========================================================================*/
import itk
import sys

This file has been truncated. show original

dzenanz · July 14, 2021, 2:22pm

2 posts were split to a new topic: GPU support in Python packages

Rafael_Palomar · July 14, 2021, 8:21am

Thank you, @matt.mccormick! That solved the issue.