Hacker News Reader

Comment by tjoff | Hacker News Reader

That statement has been said for pretty much every release of any GC.

I'd say the opposite is true. Relying on GC requires significant extra work. Because you always need to think about memory (exception: small script-like applications). The only thing a GC does is that it enable you to not think about it, but the moment you don't you will write bad code and realize it was a disservice all along. And by then it is too late.

So in a GC language you need to constantly be aware of when you take something for granted. Which is more work than just doing it manually yourself.

View on HN

First, a GC allows you to abstract over memory allocations (i.e. hide them). This makes maintaining code over time significantly easier, as allocation becomes a hidden implementation detail.

Second, while there have been similar claims made in the past, they always apply for a certain target throughput and latency. GCs are constantly making great strides in that regard. Next year, ZGC will have a <1ms worst-case latency for pretty serious allocation rates, and with an acceptable hit to throughput. As you can see in this post and previous ones, G1 offers great throughput with acceptable latencies.

This GC you speak of that hides allocations as an implementation detail sounds excellent! Every GC I've worked with - and my day job is orchstrating several thousand JVMs - simply moves the time when you need to think of allocations somewhere else.

Either it moves it into PagerDuty, like G1, or it moves it into your GCP bill like Shenandoa and ZGC.

The trade-off has always been latency-throughput-footprint, nothing I've seen yet has changed that. The innovation in Rust is realizing you can do all the tracing work at compile time.

An implementation detail doesn't mean that you don't think about it. We think about implementations all the times. It means that it's a detail, and changing a subroutine's allocation pattern does not create a ripple-effect throughout the code in its transitive consumers.

> The innovation in Rust is realizing you can do all the tracing work at compile time.

This is spectacularly false. For all non-trivial allocation/deallocation patterns, Rust also uses a runtime reference-counting GC, which is significantly slower than the tracing GCs you find in OpenJDK. The benefit comes from not relying on it too much, but this comes at a considerable cost of lost abstraction, which means more costly maintenance over the years. This is the same for all low-level languages; the difference Rust brings is that it (conservatively!) checks for memory access errors.

Another difference is that most people who talk about Rust haven't actually written a significant application in it and had to maintain it for years. I'm not saying it's impossible -- people do this for C and C++, which make a similar tradeoff in this regard -- but it does come at a substantial cost.

At this point I think we are arguing somewhat subjective things.

I used to be a hard-core zealot for the JVM as the performance platform of the future - and did talks arguing just as you are here why HotSpot outperforms $LANGUAGE in real applications. I feel like I'm hearing myself in your argument..

I wrote a signficant portion of the Neo4j storage engine, which is in Java. Now I'm writing another database engine in Rust (sidenote: not for replacing the Neo4j engine, just because it's interesting). Arguably database engines qualify as "significant applications".

I find - subjectively:

- Maintaining performant code in Rust is easier. I do the same patterns as I did in Java, except it doesn't rely on easy-to-break assumptions of how HotSpot happens to work (ex: stack allocations)

- Like you said elsewhere, the issue is often about fragmentation, ultimately stemming from object churn. I find that Rust makes it, culturally perhaps, easier to maintain low allocation code than Java. (sidenote: I think these two points is also why Go code generally has "better GC behavior"; Go doesn't have a better GC, it has a language that encourages less heap allocation)

- The engine I'm writing in Rust is faster than anything I've written in Java and - critically - runs for days and months without notable stalls.

Come to the dark side!

I personally like both languages and think they both have a place. My 2 cents.

The JVM works exceptionally well when you are dealing with long lived apps which deal with a lot of allocations and the machine running it has a substantial amount of memory. Things like webservers, for example, are near perfect fits for the JVM.

Rust does really well when you need high performance, low memory, and ultra fast startup times. It won't necessarily outperform the JVM when you talk about doing a lot of heap allocations (due to heap fragmentation) and it unfortunately suffers from the same heap fragmentation if you are dealing with a long lived server that does a lot of allocations. But then, maybe that performance loss is acceptable for the ability to very quickly scale up and down servers.

Now, the JVM is making great strides towards getting faster startup times and even fast performance at startup (AppCDS). However, those strides often involve trade offs with either build complexity or performance losses (such as Graal's AOT). The benefit for rust is that it is as fast as it ever will be without any special build steps or tweaks.

Oh, and let's not forget diagnostics. Flight recorder for the JVM is simply AMAZING. The ability to hook up to a poorly behaving production server, start flight recording, and getting detailed information about things like "where are allocations happening" or "what are the hot methods" is simply amazing. No other platform that I know of has the level of detail you can get right out of the box with flight recorder. Certainly not without restarting the application with additional configuration. For example, you'd need a special build of rust with profiling turned on to even start to get the same level of info, doing such also significantly negatively impacts performance.

Just a small correction, AOT and JIT caches like AppCDS were already available in commercial JVMs like J/Rockit or IBM WebSphere Real Time, as two examples from a couple of possible ones.

What is happening now is that the free beer Java users are also getting those features on the package.

I'm already on the dark side. Rust isn't my cup of tea, but I do a lot of programming in C++, and hope to one day use Zig for low-level programming. But this does come at a considerable cost, so I do it only when I must -- i.e. I'm targeting a constrained environment or I really need low-level control. The cost isn't so much up front (aside from the big loss in observability), when you first write the code, but low-level languages result in rigid programs that are much harder to evolve over years, certainly with large teams, that with Java. Separate from that, JDK 14 doesn't behave like JDK 8. If you need performance and observability and you are not constrained, then Java is an excellent choice that can have a big positive impact on costs.

> This is spectacularly false. For all non-trivial allocation/deallocation patterns, Rust also uses a runtime reference-counting GC, which is significantly slower than the tracing GCs you find in OpenJDK.

Are you talking about `Rc` and similar smart pointer types? If so, the twist is that in Rust almost all allocations in Rust are trivial in this sense.

There's no "twist". There was a claim that Rust does tracing at compile time; it does not. It's allocation patterns and costs are similar to all other low-level languages. The cost and benefits of such languages are well known.

My point was that Rust's compile-time machinery and conventions allow you to avoid using referencing counting smart pointers like `Rc` in almost all cases. What you call trivial patterns of allocation is in fact the prevailing type of allocation in Rust.

> For all non-trivial allocation/deallocation patterns, Rust also uses a runtime reference-counting GC...

Lest anyone read this and think it's true, it's not. Using Rc is a design choice, and not one that is a given. I have written tens of thousands of lines of Rust code doing very heavy data processing and used Rc only a handful of times. In fact, I find using Rc without a very good reason to usually be a bad idea that enables lazy thinking.

I on the other hand have used Rc quite a lot, it is either that or having to twist the application architecture to somehow fit into Gtk-rs expectations.

>This is spectacularly false. For all non-trivial allocation/deallocation patterns, Rust also uses a runtime reference-counting GC, which is significantly slower than the tracing GCs you find in OpenJDK.

This is not really a good comparison because you can't write the same code in either language. In Java practically everything gets allocated on the heap barring some optimizations. Meanwhile Rust programs can selectively allocate memory on the stack when it makes sense to do so. Reference counting is is just one of many different allocation strategies available to Rust. It is not the first tool you grab when you want to allocate memory in Rust, therefore absolute throughput of reference counting might not be as relevant in Rust as the absolute performance of the GC in Java.

> ZGC will have a <1ms worst-case latency

In our benchmarks we never saw a GC pause of more than 2 ms on either ZGC or Shenandoah, but the end-to-end latency, the one the user cares about, is impacted by much more than a single GC pause. Sometimes there would be several pauses in a rapid sequence, or just the background GC thread would do too much work at once.

Even after dedicating a core or two to the GC, you still face the issues of cache pollution and RAM throughput stealing that heap walking incurs.

Yep, that's because concurrent stack processing (https://openjdk.java.net/jeps/376) hasn't been merged yet, nor has the changes to the VM's internal references. Both of those are coming within a year. Until then, there's still significant work done by the GC at safepoints.

A month ago we benchmarked against the master of the ZGC repo. There wasn't a big difference.

Sorry, I misread your previous comment. The planned improvements that would make the biggest difference are making ZGC generational, and improving scalar-replacement (or doing some sort of stack allocation).

> This makes maintaining code over time significantly easier, as allocation becomes a hidden implementation detail.

You say that, but there seem to be no end to the stories of people spending enormous amount of time fighting the gc. It is not difficult to avoid heap allocations in other languages as well as freeing them deterministically.

First, you say it as if there aren't event bigger struggles without a GC. So the comparison is not "I'm fighting with the GC vs. I'm not" but "I'm fighting with the GC vs fighting other things, which is better?" So you hear those stories because those are the stories people who use a GC can tell.

Second, as this blog post series shows, the "fight" is not what it used to be. You don't really need to control allocation any more until your rates are really high. Java's GCs have just gotten so much better in JDK 14 and beyond.

> First, you say it as if there aren't event bigger struggles without a GC.

I'm extremely skeptical about that. My experience is that with modern C++ you lose very little elegance and gain a huge amount of control by giving up a garbage collector. Memory management becomes a very minor problem. The vast majority of memory allocations are avoided and those that need to be there can be done ahead of time.

I have never heard anyone writing a latency sensitive program in C++ (games, trading etc.) say that their life would be easier if they were using a gc or that they wished they could do it in java. I have however seen decades of people talking about all the extreme lengths and rabbit holes they go down to deal with the java gc.

From a broader perspective, pretty much any language with a gc ends up having a perpetual conversation around how the next gc will solve the problems with the current gc. You can see it in java, go, julia, and D. The only one I never hear about is LuaJIT, but maybe I just haven't seen it or maybe the expectations are lower.

Lua without a JIT is just a glue language. The expectations are low and LuaJIT is easily smashing them because a lot of Lua code is/was written for the original interpreter.

> From a broader perspective, pretty much any language with a gc ends up having a perpetual conversation around how the next gc will solve the problems with the current gc. You can see it in java, go, julia, and D.

Do you have any references handy on the julia bit there? I actually have seen very little conversation in the julia community about replacing or upgrading the GC. Mostly just the ocassional post from an inexperienced user who thinks that a borrow checker would be a good fit for julia.

Discussion seems to almost always revolve around showing users who need it, how to manually manage memory when necessary by pre-allocating arrays, using in-place operations or writing stack allocated code so that they avoid the GC in performance critical code.

I've never seriously used a language without a GC, but my feeling in Julia has always been that I never really had gripes about the GC because it's so easy to avoid the GC and take memory management into my own hands.

There were plenty of discussions a few years ago. Julia is probably used the least out of those languages for things that might have latency problems though - servers, guis, games, etc.

Interesting, in the three years I've been around I've seen little discussion of it. The main people I was aware of who cared about low-latency were robotics people and they mostly found that the techniques for manually managing memory allowed them to avoid the GC and get the latency they needed quite easily.

One complaint I will say I've heard though is that while these people find they can get the allocation behavior they want (i.e. none), some of them would like semantic guarentees that they will not hit allocations, rather than needing to test and make sure their code doesn't start hitting the GC if they switch Julia versions. Someday we might be able to provide such guarantees, but for now GC behavior is just an implementation detail that can change across minor versions.

I can sympathize with those who find that uncomfortable for sure, but in practice, Julia versions have been consistently better at getting more automatic stack allocations, not less, so it hasn't really been a problem.

> My experience is that with modern C++ you lose very little elegance and gain a huge amount of control by giving up a garbage collector.

The cost of maintaining a large C++ application (>1MLOC) with a large team over years is very significantly higher than a similar Java application. In some cases the footprint and/or performance benefits are worth that extra cost, but in the vast majority of cases they're not.

> I have however seen decades of people talking about all the extreme lengths and rabbit holes they go down to deal with the java gc.

Again, 1. Java's GCs changed dramatically in the last two years -- the GCs described in the post, are brand new/recently revised and 2. that's because that's Java's particular rabbit hole. C++'s rabbit holes, from undefined behaviour, through partial evaluation with templates and constexprs, to compilation times and sheer language complexity are far deeper.

> I have never heard anyone writing a latency sensitive program in C++ (games, trading etc.) say that their life would be easier if they were using a gc or that they wished they could do it in java.

Their lives would be easier if they could do it in Java, but sometimes they can't. I think that the changes in the last couple of years and the upcoming changes in the next few years will make Java more appropriate even in domains where it hasn't been used before, but it's fine if not. Its market reach is so huge as it is. But games are not often maintained for many years, and telemetry isn't that important, so Java's benefits are not as big as for servers.

> The cost of maintaining a large C++ application (>1MLOC) with a large team over years is very significantly higher than a similar Java application.

I am very skeptical of this, I don't know why it would be the case. My experience is that with modern C++ and avoiding inheritance programs end up much more direct and clear since a type isn't fragmented into multiple classes and base objects don't need to be used for generic programming and data structures.

> C++'s rabbit holes, from undefined behaviour, through partial evaluation with templates and constexprs, to compilation times and sheer language complexity are far deeper.

This seems like what would be said by someone who has just read a few comments on C++ here and there but not actually used it for non-trivial projects. These are rarely issues. I don't know what 'partial evaluation with templates' means and constexprs didn't even exist until recently. Compilation times do seem to be a big problem, mostly because many projects don't do anything with their structure to mitigate them.

This is only the case when latency is important below some threshold†. There is a very large class of applications that can work with that limitation, and in that case a GC can free up programmer mind-space by letting him mostly forget about memory management. You're optimizing for productivity of the developer instead of hardware usage.

Rust or other upcoming or future languages might change that.

† The threshold varies between VM and GC, but usually <10ms is easily achievable

ZGC's pause latency is at a point where it rivals, or about to rival, nondeterministic pauses by the OS. So unless you're running on a realtime kernel, GC is not an issue any more as far as latency is concerned. The only real, serious cost for modern GCs is RAM overhead.

I'm not convinced to be honest. Even with tiny heaps like 32MB JVM processes with practically nothing but a hello world micronaut project consume up to 200MB of RAM. There is more to this than just the dumb GC which always predictably hits the configured maximum or if there is no maximum it will just gobble up your most of your system. I once saw an instance of tomcat allocate 30GB to itself on a staging system with at most one user at a time simply because the machine had 128GB RAM. 30GB for nothing. I personally dislike the JVM. However I have no problem with Java or any JVM derived languages but it simply doesn't matter how much you hyper optimize your code if the underlying VM is junk.

The way RAM is managed in Java and in C is very different; neither is better than the other, but it's different, and you don't seem to understand the Java model. In that model, you give an application a maximum heap, which the program will use regardless of the minimum it needs to improve its performance. You want it to use less RAM? Just tell it to. Having said that, very recently Java has started to return RAM to the OS if it's just sitting idle. See https://malloc.se/blog/zgc-softmaxheapsize

This is not really true at all though, nor is it how people who design performance intensive systems go about it.

generally you design your application the way you normally would, do some testing, and then go back through the hot path and make some adjustments.

Reply on news.ycombinator.com