Back in grad school, one of the analysis programs I used dated back to the mid 70s. The original paper gave a performance metric for a test program, which I co

Back in grad school, one of the analysis programs I used dated back to the mid 70s. The original paper gave a performance metric for a test program, which I compared to the runtime on a Chromebook running Linux. I was curious how closely that scaled with Moore's Law, and computed "initial_release + (1.5 years)*log2(initial_runtime/current_runtime)". That is, assuming that the change in program speed has increased due to hardware improvements, and those hardware improvements follow Dennard scaling, what year is it? This (admittedly very rough) measurement ended up giving 2003. It was wrong by over a decade from the actual date, but correctly gave the date at which clock frequencies stopped improving.