HTTomo is a user interface (UI) written in Python for fast big data processing using the MPI protocol. It orchestrates I/O data operations and enables processing on CPUs and/or GPUs. HTTomo utilises other libraries, such as TomoPy, ToMoBAR and HTTomolibGPU as backends for data processing. The methods from the libraries are exposed through YAML templates to enable fast task programming.
HTTomolibGPU is a collection of image processing methods in Python for computed tomography. The methods are GPU-accelerated with the open-source Python library CuPy. Most of the methods migrated from TomoPy and Savu software packages. Some of the methods also have been optimised to ensure higher computational efficiency, before ported to CuPy.
Our objective is to enhance the performance and efficiency of the Log-polar reconstruction algorithm and the broader HTTomo framework. As part of this effort, we have already achieved a significant performance boost — improving the Log-polar reconstruction speed by approximately 60% and reducing memory consumption by around 40%. These optimizations involved extensive work with Python, CuPy, and CUDA, focusing on fine-tuning GPU-accelerated computations and memory handling to ensure the framework is scalable and efficient for large-scale tomographic data processing.
