GraalVM is a hugely under-appreciated piece of technology. A lot of my initial interest came from being able to blend different languages into a single runtime via the polyglot APIs, but the combination of performance, strong sandboxing support, and multi-language support has helped it emerge as a key primitive for anyone wishing to build extensible platforms without sacrificing speed, security, or developer ergonomics.
It's fascinating technology, but in my opinion too little, too late.
Compiling with GraalVM is still a pain and will not work out-of-the-box without modifications for any non-trivial project.
This is mostly, I believe, not GraalVM's fault, because it can only do so much within the confines of the existing ecosystem.
If only AoT compilation to native code would have been taken seriously from the start, if only gcj would have gotten more attention, we could have ended up with an ecosystem that is somewhat amenable to what GraalVM tries to do.
As much as I like GraalVM and wish it success, my prediction is that in a few years it will be in the same state gcj has been quite some years now.
gcj had a lot of problems beyond needing configuration of reflection metadata. It used a full reimplementation of the standard library, and it was never adopted by the wider Java community being largely just Red Hat's strategy for creating a fully open source Java implementation rather than something offering specific benefits to Java developers. In particular people thought it'd lead to faster code, but GCC was never designed for Java and the results were actually a fair bit slower iirc.
Native image is quite different. With this new release the compiled images can not only be faster than JIT compiled Java (wow) but also use way less memory and start instantly. At a stroke this is resolving one of the biggest complaints people have always had against JVM languages.
And as a consequence you're seeing adoption by the wider community. All the modern Java web frameworks support it now, and there's a metadata repository where it's collected for projects that haven't accepted it upstream yet .
Valhalla and the vector API for SIMD, yes. In theory, at that point, the only major performance difference between them would be GC vs manual memory management. Also: code size.
One interesting experiment that'd be worth trying is generating value types that wrap memory segments, as that's the new manual memory allocation API. You can do pretty sophisticated manual allocation with arenas and stuff, but you can't allocate Java objects into those segments. What you can do though, is create a "struct" using VarHandles that you can cast the segment to which then reads/writes through to the segment. This leads to the question of whether you can make a nice wrapper around such value types that yield C++ style manual memory allocation.
I'm not sure any of those are actually quite right.
Valhalla supports flattening, that's the whole point of it. To get that you need to have a type declared as a value type, and then put it into a non-null variable, field or parameter. At least that's how the current prototype works. If you do that then the compiler will fully inline the allocation into arrays, containing objects/structs, or the stack.
Valhalla started with monomorphization prototypes, although that went silent years ago. I don't know if it'll be delivered in the first version. But, it's an intended goal of the project to deliver.
As for the final point, could you evidence that? It's very hard to do direct comparisons here because they're usually compiling very different languages and Graal hasn't been optimized for C++. For example, Graal can compile many languages LLVM just doesn't even try to. Also, you'd really need to compare against the Oracle GraalVM (née GraalVM Enterprise) to see the best of what it can do. I'm curious what comparison you're thinking of when you say that, as Graal is an extremely advanced compiler by any measure.
I think the reality is that faster startup, smaller deployments, and lower memory usage is considered much more valuable today than it was when GCJ was an actively maintained project. It simply did not offer things people needed at that time. But now, the tradeoffs are different, and so the effort of going back and modifying existing code to support AOT compilation has a greater overall payoff than it did in the past, which is why people are pursuing it. It's the same reason .NET is adding full AOT support - because people want it now, and they want it now more than they did before.
I worked on gcj (and GNU Classpath) for a number of years. gcj's origins are in the embedded systems space, where our customers valued the small memory footprint and portability across processor architectures that were unlikely to get performant JITs. As we eased out of the embedded space, it was clear that jumping on the OpenJDK train was the right choice for Free Software-minded gcj/Classpath devs, as Sun adopted the GPL and GNU Classpath exception license. gcj was unlikely to ever catch up, and many of us were motivated by the idea of a Free Software Java solution, which Sun just handed to us.
Good point. It is fascinating technology, but to get development resources in the long run it needs management/sponsor attention and for that it needs a practical use case today. That, I think is AoT compilation and I believe that, unfortunately, it doesn't cut it there.
Everything has a cost though. The one thing we lose is the ability for hotspot to change its mind and recompile code based on usage patterns I've toyed with graalvm for some uses and ended up with lost performance. Also a key note is that it only supports a pretty naive serialgc right now which is a big limitation. Bellsoft has a parallelgc which probably plays a little better, but we're always going to need to play trade-offs until they can support the more robust GC impls
As someone on the outside looking at GraalVM with curiosity, is it even possible to use something like Poetry to manage the Python part? Would you do that? Would I even need a venv anymore? Where would its edges be? What are the limits?
How are we people practically putting this polyglot stuff into action?
Not sure I have questions, just generally interested in making it easier/lighter weight/built-in to constrain ambient authority (for example, to mitigate supply chain risks), thus have it be done more.
The Polyglot sandfox feels very loosely analogous to Deno per process (and thus subprocess) permissions, though it looks like ISOLATED and UNTRUSTED can limit a bunch of things not possible with Deno.
We do not do process isolation yet, although we have plans to implement that as well as a fallback strategy.
The advantage of the native-image-isolate-based isolation is that it is much more lightweight. For example, calls from and to the host application are much faster. There is no copying or expensive synchronization necessary. The disadvantage is that we need to do our own protections against attacks, as the OS protections between processes don't apply to such isolates. By default, we deploy software/compiler-based protections but are also very close to supporting hardware like Intel MPK.
I've only done one project with GraalVM and I've been pretty happy with it in regards to faster startup time.
I was doing stuff in Clojure. Clojure is a great language but it tends to have very slow startup times, even by JVM standards (it's not weird for a large Clojure program to take 5-6 seconds to start. Even a "hello world" can take upwards of a second or two). Graal mostly Just Worked with the standalone uberjar produced by Leiningen and created an executable that started in about 3 milliseconds. It was amazing.
While the lack of proper reflection support was a little annoying, it actually wasn't as horrible with Clojure as you might think; most problems were fixed with basic type hinting, and all but one Clojure library I used (http-kit) worked flawlessly.
The recurring problem I've seen that would cause slow startup is cramming many (require) and (import) in the project core.clj file - for non-clojure devs here, this means including a lot of dependencies in the file called on the startup.
By reordering things to be loaded smartly or lazily, startup speed can be improved a lot.
Another reason for a slow startup is that Clojure must load its own runtime every time. The current compiler is not a tree shaker, nor can it rewrite complex expressions in compile-time, although many clojure.core functions were rewritten over time to accommodate JVM optimizers. However, Graal is here precisely to do at the bytecode level, as long as you take care of all type hint warnings and avoid things Graal can't see during compilation (e.g. deferring evaluation of some expressions for runtime).
Without getting too involved with details (I was working for a certain company that doesn't like people talking in detail about anything I did there), I had a program that routinely started other processes. I was required to use Java or Clojure, and we had a sort of quasi-RPC thing that would spawn process on the shell, get a result, and return that result as a string. This application wasn't super performance-oriented, but it was enough to where a 5 second start time was untenable.
We tried to use it to improve AWS lambda startup times but desisted as it was a pain to use it with an existing app.
It required too many tweaks as there are waay too many things that rely on reflection :(
Things that broke include:
JSON (de)serialization using Jackson, validations using hibernate, validator, AWS SDK, and even simpler libs like picocli...
It could be quite useful for a set of simpler apps though
Yes, Jackson really was a bummer. To this day I can't understand how a project like Springboot advertises Graalvm readiness when Jackson is not supported without tweaks. What do the Springboot devs think we are using Springboot for, Hello World blog posts?
this was our experience as well. Both with GraalVM/Quarkus and .NET CrossGen/Native/CoreRT. It's almost a different platform that's far less supported and less stable. Even if you put in the work to make things work, they'll eventually break. Quarkus tries to give a full experience like Kotlin Native, but both are also not really there.
It's really unfortunate how Java and C# are perceived as slow because of their painful startup time.
Look at the Quarkus framework. It's designed with Graalvm native in mind, so it knows how to handle Jackson, AWS, etc. and compile them to a native binary. I'm using it in production with a native-compiled Lambda that calls other AWS services. It works GREAT!
Polyglot experimentation just got a whole lot easier! Kudos to everyone involved in pushing the JVM ecosystem forwards.
TruffleRuby runs Rails and is compatible with many gems, including C extensions. TruffleRuby is not 100% compatible with MRI 3.2 yet. Please report any compatibility issues you might find. TruffleRuby passes around 97% of ruby/spec, more than any other alternative Ruby implementation.
It is "Ahead of Time Compilation for the Java Virtual Machine". The proposal details of using AOT code at startup and then JIT taking over. So you get best of both worlds with fast application launch and then throughput once things are warmed up.
Graal is cool, but it will not be as throughput performant compared to JIT. Only for fast startup, serverless or low memory environments.
Lots of comments about issues with dependencies and Graal. I suggest looking into Quarkus. Have had great success with and mandrel (a patched version of Graal for Quarkus). Quarkus focuses on creating Web APIs so wont cover it all but if thats what you want to do its great and builds native images quite easily.
I was considering this, working on a new javafx app for a side project. Using java 17, and generating binaries with fxlauncher/jpackage/wix is not working great imo. Any good recommendations on tooling for using the graal route? I develop on linux but most users will be on windows. It would be great to not have to use windows vm for building the binaries (but i can if needed ofcourse).
What trouble are you having with jpackage? It's easy. Anyway, in order to AOT compile your JavaFX compilation you will have to run(or test) your application with Graal's tracing agent to collect the metadata and then compile it. You can use Graal's native plugins for Gradle or Maven.
Sun themselves (not Oracle) also "opened up" their Java implementation as GPL 15 years ago. It didn't stop Oracle spending years sowing their own uncertainty and doubt and trying to claim billions of dollars in damages from Google in a lawsuit where McNealy sat on the stand and testified that the GPL licensing didn't allow commercial use.