Python language support for HPC debugging and data analytics
Python’s popularity has grown so much and so fast, that most reports now list it as one of the top five most popular programming languages, along with C and C++. As its popularity grew for building data structures and algorithms, our users requested Python support in our products. Both IMSL and TotalView for HPC support users who have complex algorithmic work, such as scientists and data analysts across industries from finance to science. They use Python due to its power and ease of use while also maintaining the underlying speed and existing knowledge base in C/C++.
Python and IMSL data analytics
IMSL is a collection of mathematical and statistical functions used in science, technical, and business environments. Users of IMSL needed support for calling the IMSL C library from Python, and with our published wrappers, Python applications can easily call the library directly. Combining the ease of Python with the depth and breadth of IMSL allows users to easily create custom algorithms.
Python and debugging with TotalView for HPC
TotalView is a leading debugger for multithreaded and multiprocess applications. This makes it an ideal tool for scientists and analysts to understand and debug Python data structures in their applications. Developers like the speed of C/C++ on the backend but want the ease of use for Python for creating their front-end algorithms and logic. The issue is they want to see both languages in the same debugger and with a shared stack frame. For further details on this functionality please look at this other blog entry.
Both methods make sense for their individual tool but soon, the IMSL team plans to open source their wrappers to allow customers to build their own additional functionality. Not only does this help create more powerful applications, this also provides the opportunity for IMSL users to take advantage of the Python mixed language debugging functionality in TotalView for HPC.