News for hackers. Stuff that matters.

A comment on Going Faster by Duplicating Code.

Re:Going Faster by Duplicating Code by drmikeando on Tuesday October 10, 2023 @09:34PM (#37839958)
IMO the reason the compiler doesn't add special cases for the simplest version is that it doesn't know which of its _many_ special cases to use. If you actually use the unoptimised version of the code like

    void withSwitch(vector<int>& Values, bool v) {
      if (v) {
        multiply1(Values, 2.0);
      } else {
        multiply1(Values, 3.0);
      }
    }
Then it actually inlines the code and optimises each one correctly, as it has context about which special cases are available. (Doesn't even need the `inline` keyword for this at `-O2`)

You can see the code here: https://godbolt.org/z/5beeYe77a

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by voidstarcpp on Tuesday October 10, 2023 @09:52PM (#37840057)
This is possible if the call site can see the implementation, but you can't count on it for separate translation units or larger functions.

My goal was to not rely on site-specific optimization and instead have one separately compiled function body that can be improved for common cases. Certainly, once the compiler has a full view of everything it can take advantage of information as it pleases but this is less controllable. If I were really picky about optimizing for each use I would make it a template.

>Doesn't even need the `inline` keyword for this at `-O2`

The inline keyword means little in terms of actually causing inlining to happen. I would expect the majority of inlining compilers do happens automatically on functions that lack the "inline" keyword. Conversely, programmers probably add "inline" as an incantation all over the place not knowing that compilers often ignore it.

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by chii on Tuesday October 10, 2023 @10:38PM (#37840290)
would it have made a difference if the function was static? The compiler would then be able to deduce that it isn't used anywhere else, and thus could do this inline optimization.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by thesnide on Wednesday October 11, 2023 @02:47PM (#37848616)
Gcc actually totally does that. Even doing it while not static by copy/pasting the function and enabling further optims.

to me modern compilers are like voodoo magic

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by Thorrez on Tuesday October 10, 2023 @11:33PM (#37840571)
>Conversely, programmers probably add "inline" as an incantation all over the place not knowing that compilers often ignore it.

Funnily, the inline keyword actually has a use, but that use isn't to tell the compiler to inline a function. The use is to allow a function (or variable) to be defined in multiple translations units without being an ODR violation.

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by maccard on Wednesday October 11, 2023 @04:08AM (#37842056)
Msvc treats inline as a hint [0] , GCC is ambiguous [1] but I read it as utilising the hint. My under of clang [2] is that it tries to match GCC.

[0] https://learn.microsoft.com/en-us/cpp/build/reference/ob-inl...

[1] https://gcc.gnu.org/onlinedocs/gcc/Inline.html

[2] https://clang.llvm.org/compatibility.html

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by rileymat2 on Wednesday October 11, 2023 @07:06AM (#37843022)
What the OP is saying is that it has a use to satisfy (https://en.cppreference.com/w/cpp/language/definition) and in that case it is not optional and not ignored by the compiler. Whether it will "actually" inline it is another matter, that is optional.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by lifthrasiir on Wednesday October 11, 2023 @12:16AM (#37840823)
You can also force it by using extensions like `[[gnu::always_inline]]` or `__forceinline`. I've actually used this technique to generate an auto-vectorizable function whenever it's possible, without any code duplication [1].

[1] https://github.com/lifthrasiir/j40/blob/252e7987d36d50f617f2...

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by cxr on Wednesday October 11, 2023 @12:52AM (#37841008)
> the reason the compiler doesn't add special cases for the simplest version is that it doesn't know which of its _many_ special cases to use

This is a fact claim. I'm having trouble parsing the "IMO" that's a preface to your comment.

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by deredede on Wednesday October 11, 2023 @02:06AM (#37841459)
The "IMO" here is used instead of "I think" or "I believe" to indicate that this is not a fact claim but a (presumably educated) guess. Not a very correct use of "IMO", technically, but a fairly common one nonetheless.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by cxr on Wednesday October 11, 2023 @08:57AM (#37843803)
Are you sure this is common? I've never seen it before.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by dahart on Wednesday October 11, 2023 @10:34AM (#37844928)
Whoa, that’s wild! IMO == “In My Opinion”, and it’s all over the internet, along with its cousin IMHO (“In My Humble Opinion”). Are you sure you haven’t seen it before? Try searching for it on HN or Google. HN search shows me almost 100k comments with IMO, and the first page of hits is all this meaning.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by cxr on Thursday October 12, 2023 @01:48AM (#37853831)
"IMO" is common. "IMO" used this way, on the other hand, is uncommon enough that I've literally never encountered it before.
[ Parent | on HN ]
Re:Going Faster by Duplicating Code by bunderbunder on Wednesday October 11, 2023 @11:00AM (#37845314)
For all sorts of reasons, I think that being pedantic about the meanings of acronyms probably does more to inhibit understanding each other than it does to facilitate it.

In this particular case, regardless of how often anyone has personally seen the term used to mean something more like "I think that...", taking that to be the intended meaning is the most friendly interpretation, and therefore the preferable one. It's an informal forum, we type fast, and don't necessarily carefully proof-read comments before hitting reply. Nor should we -- this place is a more pleasant space to occupy for everyone involved if we give each other the benefit of the doubt about choice of words.

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by cxr on Thursday October 12, 2023 @01:47AM (#37853822)
> being pedantic about the meanings of acronyms[...] taking that to be the intended meaning is the most friendly interpretation

Why are you responding as if it's a foregone conclusion that my intent was to be uncharitable/unfriendly?

Someone wrote a comment with unusual phrasing, and I asked for help understanding the message. It contains neither an explicit nor an underhanded attempt to be an asshole to someone. The adversarial reading you're going with is exactly that—something that you're bringing to it, and is (perversely) wholly uncharitable in and of itself...

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by yakubin on Wednesday October 11, 2023 @05:10AM (#37842387)
> Doesn't even need the `inline` keyword for this at `-O2`

Not surprising, since inline has nothing to do with inlining. :)

[ Parent | on HN ]
Re:Going Faster by Duplicating Code by bunderbunder on Wednesday October 11, 2023 @10:46AM (#37845103)
On modern hardware, it's also not clear that a special case that makes things faster from a purely CPU-oriented perspective makes things faster overall. Adding additional code to handle special cases makes the code larger, and making the code larger can make it slower because of the memory hierarchy.

There's a CPPCast interview with one of the people who worked on the Intel C Compiler where he talks about the things they do to get it producing such fast binaries. At least from what he said, it sounds like the vast majority of the unique optimizations they put into the compiler were about minimizing cache misses, not CPU cycles.

[ Parent | on HN ]