.. title: PAR Homework 2, Thu 2020-01-23
.. slug: homework02
.. tags: homework
.. date: 2020-01-21
.. category: 
.. link: 
.. description: 
.. type: text
.. has_math: true

Rules
-----

#. Due next Thurs, 2020-01-30, noon.

#. Submit the answers to Gradescope.

#. You may do homeworks in teams of 2 students.   Create a gradescope team and make one submission with both names.

#. For redundancy, at the top of your submitted homework write what it is and who you are.  E.g., "Parallel Homework 2, 1/30/20, by Boris Badenov and Natasha Fatale".

Questions
---------

Each is 5 points.

#. Why is one CUDA core less powerful than one Intel Xeon core?

#. If 10% of a program cannot be parallelized, then what is the max speedup even with infinite parallelism?

#. "For short running parallel programs, there can actually be a decrease in performance compared to a similar serial implementation.".  Why?

#. What is the difference between strong scaling and weak scaling?

#. What is cache coherency?

#. Define "embarrassingly parallel".

#. How many CUDA cores does the RTX 8000 have?
   
#. Why have machine cycle speeds stopped increasing?
   
#. Give an advantage and a disadvantage of shared memory versus distributed memory for parallel computation.
   
#. In OpenMP, what's the difference between ATOMIC and CRITICAL?


Total: 50 pts.