Performance Analysis

What is the difference between strong scaling and weak scaling in evaluating parallel computing?

  1. AStrong scaling fixes total problem size while adding processors; weak scaling fixes problem size per processor and grows the total
  2. BStrong scaling uses fast processors, weak scaling slow ones
  3. CBoth are the same measurement under different names
Show answer and explanation

Answer: A. Strong scaling fixes total problem size while adding processors; weak scaling fixes problem size per processor and grows the total

Strong scaling solves the same problem on more processors and measures how much time drops; serial parts and communication tend to make gains plateau as processors increase. Weak scaling keeps work per processor constant while growing the total and checks whether time stays constant. Results cannot be compared correctly without stating which kind of scaling was measured.

  • A. Correct. The former matches Amdahl's view, the latter Gustafson's.
  • B. It is not about processor speed.
  • C. They differ in what is held fixed.
Practice this topic Back to topic

More Performance Analysis questions

  1. What is the most appropriate approach when evaluating computer performance with benchmarks?
  2. Which describes a standard benchmark such as SPEC CPU?
  3. When combining ratios of speed against a reference machine across several benchmarks, why is the geometric mean often used?
  4. Two programs with equal instruction counts run back to back, one at 1000 million instructions per second and the other at 500 million. What is the overall average rate?
  5. Timing a small loop like the one above gave nearly zero time even for large N. What is the most likely cause?