Doing less really means less code totally at the current compilation target, essentially feeding fewer total instructions to the compiler. This means no frameworks and minimal abstractions. It means having a clear appreciation for the APIs you are writing to. It means minimizing use of nested loops, which exponentially increase statement count.
Sometimes caching groups of instructions in functions can allow for cleaner code with a positive performance impact.
V8 cannot compile arithmetic assignment operators, which it calls left-side expressions, so you can see a rapid speed boost in V8 when you replace something like a += 1 with a = a + 1.
The side benefit of less code is generally clearer and cleaner code to read. There isn't any wizardry or black magic. No tricks or super weapon utilities.
As an example I wrote a new diff algorithm last year that I thought was really fast. https://news.ycombinator.com/item?id=13983085 This algorithm is only fast because it does substantially less than other algorithms. I only wrote it because I could not wrap my head around the more famous Myers' O(ND) algorithm. A side benefit, in this case, of doing less is an algorithm that produces substantially more accurate results.
Is there a reason it can't? I'm not familiar with Javascript, but aren't the two expressions equivalent?
Nesting two loops has an n^2 cost and nesting 3 levels deep costs n^3. At no point does it ever cost 2^n or any x^n.
It's polynomial, not exponential.
I know in reality the frequency of iterations varies considerably but for simplicity of discussion let's remove variability.
Say we have a loop with 1000 iterations. That is at minimum 1000 statements in the loop body plus expression overhead from the loop itself. If this loop is nested once with a same sized loop there are now 1,000,000 iterations plus some expression overhead per iteration. If it is nested twice deep there are now 1 billion iterations.
That example is exponential of 1000. Given that there is overhead associated with operation of a loop it is actually greater than exponential. It may not be quite so dramatic as a logarithmic growth curve though.
I completely concede that in reality loops vary in iteration count and so nested loops aren't likely perfectly exponential unless their iteration counts are identical. The increase of iterations from nesting loops does increase more dramatically than a simple multiplicative operation as the depth of loop nesting increases, such that the growth of total iterations is a curve on a graph. A polynomial growth operation when graphed should present a straight diagonal line without the presence of a third variable.
That's only true if your n is referring to the number of iterations of each loop. If your n instead refers to the number of nested loops, you will indeed get a runtime of O(2^n).
Seeing as in this context we were talking about varying the number of nested loops, it makes sense that we would define our variable n = number of nested loops.
a += 1 would only compute the address of "a" only once and then duplicate it, whereas
a = a + 1 would compute the address of "a" twice. For more complicated examples the first should see an even faster speedup.
So, from my perspective, there is a serious problem here in the optimiser.
Where did I neglect maintainability as a factor? The only optimization that potentially affects maintainability is manually allocating Mapping-s in the typed array. And there I openly acknowledged that it affects readability and makes the code error prone. All other optimizations are not in any way affecting maintainability.
Even typed array optimization is purely confined in the library internals... On the other hand WASM spills out of the library by requiring users to explicitly destroy SourceMapConsumer.
But "completely neglected code maintainability" is definitely unfair. While you made changes that reduced maintainability, you weren't neglectful.
Circling back; both authors have their biases - it's best to read both articles skeptically, and IMO, take away the sage bits of advice they both echo and not anything about a specific technology: You should make the right choices for your project and your design, performance, and maintenance needs (including making your own evaluations), rather than jumping on some bandwagon without being properly informed. Also that profile-guided algorithmic improvements are usually the easiest place to make sweet, sweet perf gains (in any language) before you have to get into hairy maintainability trade-off decisions.
You get dynamic weak typing making reuse more complex, unpredictable (nonlinear in performance with coffee changes) GC and JIT, weird type system. No error handling facilities in the language either.
Heck, compared to JS even modern Java (which shares the GC and JIT unpredictability) or C++ (incl. arcane syntax and less safety if you like living dangerously) seem easy to achieve predictable results with.
Rust takes more work you front - but not that much more.
Not that the WASM version of the library would've helped, since Nashorn doesn't do WASM at all. But maybe the performance would've been decent if it had.
I'm not sure that is true. Having worked/interacted with a lot of people working with Rust on different experience levels, most of them (that includes me) don't have a deep knowledge of what Rust concept maps to a specific concept with which performance implications. And if they do it's often only partial. I'd say that right now, only very few people that don't work on the Rust compiler have a broad knowledge in that area. Sure, it's much better to have to Result of the optimization expressed in code itself, but I'd say that the amount of knowledge and effort required to get to such a level of optimization is similar to optimizing Javascript.
I also found the hint to `#[inline]` suggestions, a bit disingenuous. In the end they are just _suggestions_, and your are just as much at the mercy of the Rust/LLVM optimizer to accept them, as you are with a Javascript JIT.
I'm a big fan of Rust, and I'm a big fan of Rust+Webassembly (working with it is the most fun I had programming in a long time!). Generally I think that Rust has one of the better performance optimzation stories, I just don't a gree with some of the sentiments in the post. There are also enough other reasons to love Rust+WebAssembly than just the peformance!
Really? After reading all of the articles in this series (the original about porting to Rust, the rebuttal about optimizing JavaScript, and this one)?
I'm more familiar with Rust than JavaScript, but I found that other than the algorithmic optimization, the rest of the JavaScript optimizations were very non-obvious in order to achieve an effect that is entirely natural in Rust.
There's no need to be careful to avoid particular types of function calls, since there are no dynamic variadic function calls in Rust. There's no need to resort to using the equivalent of `eval` for monomorphization, it's a natural feature of the language. There's no need to do manual memory management by encoding values into typed arrays of integers; you can simply allocate vectors of structs, borrow references, and so on.
These are all things that had significant cost in JS, and needed to be worked around via some intensive profiling and knowledge of how JIT engines work, but just doing things naturally in Rust leads to a pretty much zero-cost solution.
It's true that if you want to really get into the nitty-gritty of micro-optimization in Rust, you need to learn some more and do things like profiling and inspecting the generated code to see what optimizations the compiler was able to apply or not.
But the Rust rewrite of source maps did none of that; they just rewrote it in fairly idiomatic Rust, and achieved a similar speedup to what a JIT engineer armed with a profiler, a deep knowledge of what affects how well a JIT works, and a willingness to do manual memory management in typed arrays was able to do.
> Having worked/interacted with a lot of people working with Rust on different experience levels, most of them (that includes me) don't have a deep knowledge of what Rust concept maps to a specific concept with which performance implications
What parts of Rust do you feel like you don't understand the performance implications of? I feel like the mental model for performance is relatively similar to C or C++, with relatively few things that would surprise you if you're familiar with modern C++ (in fact, a lot fewer performance surprises than modern C++ offers, in addition to the extra safety).
The main point it makes is, again "He perfectly demonstrates one of the points my “Oxidizing” article was making: with Rust and WebAssembly we have reliable performance without the wizard-level shenanigans that are required to get the same performance in JavaScript."
This doesn't make a lot of sense as a claim.
Why? Because underneath all that rust .... is an optimizing compiler, and it happens the author has decided to stay on the happy path of that. There is also an unhappy path there. Is that happy path wider? Maybe. It's a significantly longer and more complex optimization pipeline just to wasm output, let alone the interpretation of that output. I have doubts it's as "reliable" as the author claims (among other things, WebAssembly is still an experimental target for LLVM). Adding the adjective "reliable" repeatedly does not make it so.
Let's ignore this though, because there are easier claims to pick a bone with.
It also tries to differentiate optimizations between the two in ways that don't make sense to me: "In some cases, JITs can optimize away such allocations, but (once again) that depends on unreliable heuristics, and JIT engines vary in their effectiveness at removing the allocations."
I don't see a guarantee in the rust language spec that these allocations will be optimized away. Maybe i missed it. Pointers welcome.
Instead, i have watched plenty of patches to LLVM go by to try to improve it's heuristics (oh god, there's that evil word they used above!) for removing allocations for rust. They are all heuristic based, they deliberately do not guarantee attempting to remove every allocation (for a variety of reasons). In general, it can be proven this is a statically undecidable problem for a language like rust (and most languages), so i doubt rustc has it down either (though i'm sure it does a great job in general!)
The author also writes the following: "WebAssembly is designed to perform well without relying on heuristic-based optimizations, avoiding the performance cliffs that come if code doesn’t meet those heuristics. It is expected that the compiler emitting the WebAssembly (in this case rustc and LLVM) already has sophisticated optimization infrastructure,"
These two sentences literally do not make sense together. The "sophisticated optimization infrastructure" is also using heuristics to avoid expensive compilation times, pretty much all over the place. LLVM included. Even in basic analysis, where it still depends on quadratic algorithms in basic things.
If you have a block with 99 stores, and ask LLVM's memory dependence analysis about the dependency between the first and the last, you will get a real answer. If you have 100 stores, it will tell you it has no idea.
What happened to reliable?
Why does this matter? For example: Every time rust emits a memcpy (which is not infrequent), if there are more than 99 instructions in between them in the same block, it will not eliminate it, even if it could. Whoops. Thats' a random example. These things are endless. Because compilers make tradeoffs (and because LLVM has some infrastructure that badly needs rewriting/reworking).
These "sophisticated optimization infrastructures" are not different than JITs in their use of heuristics. They often use the same algorithms. The only difference is the time budget allocated to them and how expensive the heuristics let things get.
There may be good reasons to want to write code in rust and good reasons to believe it will perform better, but they certainly are not the things mentioned above.
Maybe what the author really wants to say is "we expect the ahead of time compiler we use is better and more mature than most JITs and can spend more time optimizing". But they don't.
Maybe it would also surprise the author to learn that their are JITs that beat the pants off LLVM AOT for dynamic languages like javascript (they just don't happen to be integrated into web browsers).
But instead, they make ridiculous claims about heuristics and JITs. Pretending the compiler they use doesn't also depend, all over the place, on heuristics and other things is just flat out wrong. At least to me (and i don't really give a crap about what programming language people use), it makes it come off as rampant fanboism. (Which is sad, because i suspect, had it been written less so, it might be actually convincing)
There are two big differences here: 1) You're comparing "staying on the happy path" of a JIT compiler in the JS case, vs. an optimizing compiler in the Rust case. With the latter you can just compile your code and see what comes out and it tends to be fairly predictable. With the former, I'm not even sure there are tools to inspect the generated JIT code, and you're constantly walking the line of JS engines changing their heuristics and throwing you off the fast path. This was one of the primary motivations for the asm.js/WebAssembly work: the ability to get predictable performance.
2) Many of the optimizations mraleph performed were tricks to avoid allocation (which is normal optimization stuff, but more of a pain in GCed languages). In JS he winds up having to effectively write C-in-JS which looks pretty hairy. In Rust controlling allocation is a built-in language feature, so you can write very idiomatic code without heap allocation.
Predictable as long as you stay on the same version of the compiler (yes, I know that there are crater runs to prevent regressions). Also, how does much can/does the output for different target architetures differ in performance? Couldn't that be likened to trying to optimize for multiple JS engines?
This is in contrast to JavaScript where objects are always allocated on the heap, and are garbage collected.
So Rust gives you a lot more control and flexibility about how memory is managed. This might or might not matter for your use case, of course. There are compiler optimizations and heuristics -on top of that-, to be sure, but you end up with a lot more guarantees about how your code executes, because that’s what the language is designed for.
EDIT: If you want to learn more about memory management in Rust, see https://doc.rust-lang.org/book/first-edition/the-stack-and-t...
Surely you realize that this, and what the author wrote, are basically the same ever rehashed GC vs non-GC language discussion. Performance characteristics of each is not anywhere near as simple as any of these things claim, so i'm going to leave this alone.
"EDIT: If you want to learn more about memory management in Rust, see https://doc.rust-lang.org/book/first-edition/the-stack-and-t...
Again, i see no guarantees here.
I see it saying things like "This program has one variable binding, x. This memory needs to be allocated from somewhere. Rust ‘stack allocates’ by default, which means that basic values ‘go on the stack’."
Can someone point me to an actual language level guarantee in a spec? I looked, and i don't see it.
I can't see anything that makes it non-conformant to build a rust compiler that dynamically allocates and places all of these on the heap, and is very non-constant time for local variable allocation.
But again, i didn't spend more than a few minutes browsing, so i may have missed it (For example, https://doc.rust-lang.org/reference/memory-allocation-and-lifetime.html says nothing here, the requirements i can find in this entire chapter can easily be met by using dynamic allocation. There are not guarantees on space or time usage that would require a real stack :P).
I'm totally willing to believe nobody has written down a good enough spec yet, just saying i don't see it ATM :)
I want to strongly differentiate between what "one implementation of rust does" and what "the language guarantees". Because if you are going to claim it's rust that makes the guarantee, as the author did, you should be able to back the claim up.
Sure, the Rust compiler and the LLVM infrastructure are doing a lot of complicated work internally... but the end users of the compiler aren't having to spend time digging through the guts of it to guess what kind of magic sequences are needed to get fairly good performance.
That's precisely my point: The claims they make about this in this article that claim to prove this are demonstrably false. So that may even be true, but it's definitely not anything in this article that shows that.
If the author had just said "hey, when i wrote this version, it performed better and was easier", that'd be great, and awesome.
Instead, it seems they wanted to make more general claims, and to be honest, don't seem to have a lot of idea what they are talking about on that front, and to anyone who does have experience in this area, like i said, it comes off very badly.
I'll also point out, if you expect to not need to do profiling and algorithmic improvement to anything to get significant speedups, that's also similarly silly. It's just not realistic for any language in the real world. The only question is "which of the code you write will this be true for" not whether it will be true.
>WebAssembly is designed to perform well without relying on heuristic-based optimizations, avoiding the performance cliffs that come if code doesn’t meet those heuristics. It is expected that the compiler emitting the WebAssembly (in this case rustc and LLVM) already has sophisticated optimization infrastructure, that the engine is receiving WebAssembly code that has already had optimization passes applied, and that the WebAssembly is close to its final form.
But really the point of the article is the last part.
>that the engine is receiving WebAssembly code that has already had optimization passes applied, and that the WebAssembly is close to its final form.
The compiler applies heuristics once during the compilation step and then never again. Compare this to JITs which are constantly changing and can pull the rug from under you.
>Maybe it would also surprise the author to learn that their are JITs that beat the pants off LLVM AOT for dynamic languages like javascript (they just don't happen to be integrated into web browsers).
Yes but it gets even better! If you limit yourself to a formalised subset of javascript called asm.js which gives you fine control over memory layout and allocation you can reach even the performance of C! Have you heard of it's successor? I think it's name was Web Assembly and every major browser has integrated it. It's a really cool technology that shows how sandboxed JITs can have the same performance characteristics as AOT compilers.
For optimizations, you're right : LLVM uses a lot of heuristic. But this happens at compile time, when you generate your webassembly blob. When you have it, it will run at a consistent speed across browsers and between several generations of the same browsers. None of this is achievable in plain JavaScript, where JS optimized for old V8 (crankshaft) won't be optimal for SpiderMonkey or newer version of V8 (Turbofan).
A naive javascript interpreter is really slow, spidermoney and V8 do a ton of work to make it reasonably fast. If there's an important piece of code where the compiler doesn't even completely fail to help, but just helps less than normal, the code is no longer reasonably fast.
I think you've got a fair point, that all optimizing compilers and JITs have happy paths and unhappy ones, falling off the happy path for either form will result in slower code. But, the happy path being wider kinda seems like the key thing here?
Additionally, AOT has consistency, in that it doesn't depend on runtime data. This can obviously leave some performance on the table if a JIT ends up optimizing for the data that happens to be used in a given session, but also means performance usually doesn't (accidentally) depend on global state.
> Instead, i have watched plenty of patches to LLVM go by to try to improve it's heuristics (oh god, there's that evil word they used above!) for removing allocations for rust
I recall some that just told LLVM the names of the symbols Rust uses instead of malloc/free, so that it could do its standard escape analysis of completely pointless (that is rarely useful for Rust, IME), but didn't actually touch the rest of LLVM.
In any case, rustc (and the current understanding of Rust as a language) doesn't insert allocations itself, because there's no reason to, unlike JS, where the allocations are essentially semantically required. That is to say, Javascript's language model is so flexible that many things have to allocate by default (as it's the only way to get the required shared-mutation behaviour), where as in Rust, heap allocation is implemented as a standard library construct (it's not necessary for the language: e.g. libcore is a bare-metal subset of the standard library that works with no dependencies) and people won't manually put things on the heap unless they have to.
> Maybe it would also surprise the author to learn that their are JITs that beat the pants off LLVM AOT for dynamic languages like javascript (they just don't happen to be integrated into web browsers).
Are you saying that a JIT for JS beats LLVM for... JS? I don't think that's particularly surprising (and probably especially not for someone who seems to have a lot of history with writing JS engines/Spidermonkey, like the author), given things like webkit's experience: https://webkit.org/blog/5852/introducing-the-b3-jit-compiler... .
This is a common mistake. It should read, "This is a factor of 3 improvement!"
x+x+x+x is an improvement over x of 3x not of 4x. The improvement factor is 3.
The general equation for an increased quantity from A to B is
increase = B - A
increase * 100
(increase + A) * 100