Java's new AOT cache feature is in a rough state

Comments

Comment by lazystone 3 hours ago

> In theory, the whole AOT cache idea sounds like something that should only affect the class data and not perform any native compilation

https://openjdk.org/jeps/544

> Non-Goals

> It is not a goal to support cross-compilation. Code compiled in a training run must run on the same CPU architecture, with the same set of features, in subsequent production runs.

Comment by crschnick 3 hours ago

Your link is about a completely separate feature, called the AOT code cache (targeted for JDK28, not released yet). The post is about the AOT cache, which forms the foundation for this feature, but is separate and already available since JDK24. The JEP for the feature the post is about is https://openjdk.org/jeps/483

Comment by florieger 3 hours ago

I wonder if this should trigger a warning (or even an error) at application startup whenever the CPU architectures differ.

IMHO, that would be much better than running into crashes and having to figure out the cause yourself.

Comment by odo1242 1 hour ago

Yea, it should probably warn and fallback to normal JIT

Comment by jayd16 2 hours ago

Should probably just kick in the JIT at that point, no?

Comment by paulddraper 3 hours ago

Yeah, there are some strange expectations going into this.

"This is what I would like AOT cache to mean."

Comment by pron 2 hours ago

For the past several years, Java features are released piecemeal, with each piece released as soon as it can be useful to some significant portion of users or users who can offer useful feedback (and when it's not expected to restrict future pieces). The AOT cache is already used in production despite not being as user-friendly as it could be, because the companies using it find it valuable.

The first use case is the very common one (for Java) where developers control deployment (internal deployment). The last basic building block, caching machine code, will be released next year in JDK 28, and then the feature is likely to evolve to better support other use cases, such as external deployment.

Comment by jbellis 2 hours ago

Yes. And this is the right tradeoff vs the ossification you saw previously in Java (and we're starting to see now in Python, see the recent ejection of the CPython JIT), especially now when you can say "GPT, source dive the JDK and figure out which cutting edge is causing the bleeding."

Comment by deepsun 3 hours ago

> What would be needed is some sort of compatibility switch that disables all optional CPU features. Sure, you will get some reduced performance of the compiled code, but in exchange, a faster startup time.

I think other AOT-compiled languages don't opt in for the lowest set of CPU features -- they just detect them at runtime and use whenever present, all in compiled machine code.

Comment by mikepurvis 3 hours ago

It's normally an opt-in thing I think, since it's not an unconditional win: you have to figure out which code paths are worth it, and then where to trade off binary bloat vs the runtime cost of dynamic dispatch. In rust for example:

https://doc.rust-lang.org/beta/core/arch/#dynamic-cpu-featur...

Comment by pregnenolone 31 minutes ago

I saw this on Reddit, and as expected, the replies from the Java community were quite hostile. Sometimes it feels like Java developers have never left their own ecosystem and have never seen how easy some things (like deployment) are on other platforms. Especially Java's way of deploying applications has hindered not only Java but also alt JVM languages like Kotlin and Scala.

Comment by wokkel 6 minutes ago

If you are regering to the mess that is python or javascript in beg to differ

Comment by pjmlp 2 hours ago

Note this is all about OpenJDK and Project Leyden.

There are plenty of other AOT/JIT cache approaches in the Java ecosystem already, with more mature deployment experience of several years, namely:

OpenJ9, Azul, PTC, Aicas, the Android cousin since Android 7

Technology based on process snapshots CRaC for Amazon OpenJDK distro (Correcto), and OpenJ9, is also another approach available for some time.

Comment by _old_dude_ 3 hours ago

The Inside Java podcast episode on AOT caching covers the issue in this segment:

https://youtu.be/U2j4bKgKr-s?t=3577

Comment by emoII 3 hours ago

What are the most common ways to improve warmup times, actually? Looked a bunch into this, primarily native image, aot and crac, all of them super messy to get to work with a normal JVM webapp. I understand the warmup concept is inherent in JIT compilation, but how are big companies like Netflix handling this to quickly scale microservices horizontally etc?

Comment by hocuspocus 2 hours ago

Native images shouldn't be too messy anymore, Oracle put a lot of effort into DX improvements and all major web frameworks are now aware it's an important and growing use case.

Struggling to build native images should be a sign you're reliant on libraries that are dangerously outdated and abandoned.

Comment by jcurtis 1 hour ago

They can definitely be pretty messy in some ways; even the standard library doesn't work out of the box perfectly. I have certainly struggled to get image manipulation with java.awt (e.g. thumbnail generation for a fairly standard web use case) to work in a native image.

Comment by emoII 1 hour ago

So it is an antipattern nowadays to manipulate the classpath after building? Iirc native image require static classpath, but it is quite common to inject e.g. environment specific data via the classpath. Should one just read from the filesystem instead, or what’s the strat there?

Comment by crschnick 31 minutes ago

With the AOT cache, you can actually modify the classpath (or module path as is now used for modern applications) after it has started up. The names and classes for the APIs are different, but you can still load jars at runtime even with the AOT cache enabled and functional.

The AOT cache compatibility issues come when your boot module path differs between training and actual run. As long as that is not the case, you can use it without problem and load additional modules at runtime after it has started up.

With GraalVM, that is not possible. It still has the closed-world assumption.

Comment by jiehong 3 hours ago

When AOT came to be, I kinda hoped for GraalVM's solution would make it relatively as-is to open jdk.

Instead, this AOT feature is not very ergonomic to use at all, with too many pitfalls to be of broad use IMO.

Comment by JMKH42 3 hours ago

I wonder if you could do the training run on install, so that each user gets a build tied to their specific cpu?

Comment by pron 2 hours ago

Yes (or something along those lines), it's just that this isn't the first use case addressed by this capability, which is being delivered gradually (as we normally do in the JDK), and not even all the basic building blocks are out yet.

Comment by phi_node_48 1 hour ago

[flagged]