Operation Costs in CPU Clock Cycles (2016)

Summary

Author:“No Bugs” Hare Follow: Job Title:Sarcastic ArchitectHobbies:Thinking Aloud, Arguing with Managers, Annoying HRs, Calling a Spade a Spade, Keeping Tongue in Cheek UPDATED: TLB and CAS/atomics (including different NUMA node) added Click to enlarge NB: scale is logarithmic! Premature Pessimization Easy on yourself, easy on the code: All other things being equal, notably code complexity and readability, certain efficient design patterns and coding idioms should just flow naturally from your fingertips and are no harder to write than the pessimized alternatives. This is not premature optimization; it is avoiding gratuitous pessimization.— Herb Sutter, Andrei Alexandrescu —Whenever we need to optimise the code, we should profile it, plain and simple. However, sometimes it makes sense just to know ballpark numbers for relative costs of some popular operations, so you won’t do grossly inefficient things from the very beginning (and hopefully won’t need to profile the program later 🙂 ). So, here it goes – an infographics which should help to estimate costs of certain operations in CPU clocks cycles – and answer the questions such as “hey, how much L2 read usually costs?”. While answers to all these questions are more or less known, I don’t know of a single place where all of them are listed and put into perspective. Let’s also note that while the listed numbers, strictly speaking, apply only to modern x86/x64 CPUs, similar patterns of relative operation costs are expected to be observed on other modern CPUs with large multi-level caches (such as ARM Cortex A, or SPARC); on the other hand, MCUs (including ARM Cortex M) are different enough so some of the patterns may be different there. Last but not least, a word of caution: all the estimates here are just indications of the order of magnitude; however, given the scale of differences between different operations, these indications may still be of use (at least to be kept in mind to avoid “premature pessimisation”). On ...