Nested loops with index locality P&L: -3 (≃ -28 USD)
Can we loop and avoid cache misses for computer performance?
YAML
Проект
Производство
If you iterate over data, the most efficient retrieval is where the data is all in a line.
How do we arrange data so that data is always near the previous data we retrieved?
Годовая IRR: -1.0000
Latest NPV@bank rate=0.1: -3.0000 ħ (-27.91 USD)
-3.0000 ħ (3.0 HUR) (+0.0)
I wrote a nested for loop that iterates in reverse order between iterations to minimize cache misses.
Я потратил на это несколько часов.
Если бы это применялось ко всем циклам программ, мы могли бы иметь еще более эффективные программы.
I spent a few hours on this.
If this was applied to all looping of programs, we could have even more efficient programs.
