Welcome to the Becoming a Sage podcast, hosted by Dr. Jann. This week’s episode features Dr. Jann’s conversation with Pat Hoertdoerfer.
Rev. Patricia Hoertdoerfer is a professional educator and a retired Unitarian Universalist minister who practiced her leadership in academic institutions, congregations, UUA staff and camps/conference centers, and interfaith communities over the past forty years. As a certified Sage-ing® Leader she is currently sharing her ministry with elders while engaging in service to future generations. Partner to Manfred for fifty-plus years, mother to four adult children, and Oma to seven grandchildren, Pat enjoys retirement living in the Lakes Region of New Hampshire. Through her many decades she has lived her passions of sharing stories around a campfire, engaging multigenerational spiritual growth and participating in cross-cultural adventures and service learning.
So to things:1) It depends entirely on the task you are trying to parallelize “Nick’s class” or “NC” is the category of algorithms considered to parallelize well, while “P” is the class of algorithms that are generally considered tractable to implement and run in a single threaded manner That is the same P as the P versus NP debate, and it’s generally considered that P is not likely equal to NP, and NC is not likely equal to P That whole there are no problem concrete examples, it’s expected that some tractable single threaded algorithms do not parallelize well 2) The number of lines in a program tells you literally nothing about it’s runtime If they only run once each in order, even an e-core of a older budget chip will handle 2000 lines of script faster than you can blink On the other hand, even a few hundred lines would be more than sufficient to (try to) print out the Goodstein sequence of an input number For an input of 3, this runs quickly, outputting 6 numbers For an input of 4 it will eventually to output over 1e100,000,000 numbers, most with over 100 million digits each Determining how long a program can run based on just on the program size is actually so difficult that it is *literally impossible*