GraalVM 20.3 Released With Many Small Performance Optimizations

[Ironmask]

But maybe someday I can write useful programs that don't need 6, 10, or 35MB of RAM just to print "Hello World".

I know that was just a quip, but it's really nitpicky to critique a runtime for costing a few measly megabytes of memory. You know that memory doesn't scale up or anything, the runtime will cost 30MB whether you're writing "hello world" or calculating weather patterns. You probably have images on your hard drive that are bigger than that.

[blacknova]

Are there any tests of GraalVM agains OpenJ9? Both are oriented for faster startup and smaller footprint.

[Michael_S]

I know that was just a quip, but it's really nitpicky to critique a runtime for costing a few measly megabytes of memory. You know that memory doesn't scale up or anything, the runtime will cost 30MB whether you're writing "hello world" or calculating weather patterns. You probably have images on your hard drive that are bigger than that.

I've just been cranky about the efficiency of software in general recently. At work I have to use Microsoft Teams (Electron), Slack (Electron), Jira (single browser tabs in Firefox or Chrome running Jira consume over 1GB of RAM if I leave them open a few days), and IntelliJ (which rarely dips below 5GB of RAM when I work). My computer hits swap from time to time, and I've got 16GB of RAM.

[bug77]

More capable how? My impression of Kotlin was that someone looked at Scala and then tried to adopt the 50-80% of its features that wouldn't terrify someone coming from Java 7.

I was thinking of coroutines, Kotlin Native... sbt I managed to avoid so far.
Though I'll admit, I have fonder memories of Scala than I do of Kotlin.

[elatllat]

...GraalVM at least supports almost all of the use cases of JVM so its easier to migrate.

GraalVM native is still missing support for ARM64

[pal666]

Andrew Kelley is working on the Zig language.
Alex Medvednikov is working on the V language.
Jonathan Blow is working on the JAI language.
Drew DeVault has said he's working on his own C replacement.

thousands of them, languages nobody uses

[Michael_S]

thousands of them, languages nobody uses

You're probably right, but I hope you're wrong.

[mdedetrich]

I was thinking of coroutines, Kotlin Native... sbt I managed to avoid so far.
Though I'll admit, I have fonder memories of Scala than I do of Kotlin.

Scala has Future/Task/IO, it doesn't need co-routines (also Scala's abstractions for async IO are much better than co-routines). Scala actually had a compiler plugin that provided co-routines called scala-async https://github.com/scala/scala-async which basically did CSP style transformation on Scala's Future. Having both used, co-routines is usable for very trivial async code however once you start doing anything remotely complex using Future/IO/Task is much better (this is also the reason why the scala-async plugin in Scala kinda died).

There is also scala-native https://scala-native.readthedocs.io/en/v0.3.9-docs/ which is analogous to kotlin-native, it uses llvm as a backend to create native binaries (although with Graal this may not be as big of an advantage). Note that Scala native actually predates kotlin native, so kotlin wasn't first here either.

[Sin2x]

Relatively little known there is a project to create a safer C subset at Microsoft until a better alternative is created if it ever is:

https://www.microsoft.com/en-us/research/project/checked-c/