I estimate that developing OpenJDK + Graal/Truffle costs ~$100M a year (hundreds of full-time developers). That's a lot of money.
Re:GraalVM 19.0
by aardvark179 on Saturday May 11, 2019 @11:17AM
(#19886582)
Making a VM isn’t too difficult. Making s VM with an optimising JIT is a lot more work, and adding things like tooling, instrumentation and debugging increases the work needed.
Re:GraalVM 19.0
by chrisseaton on Saturday May 11, 2019 @10:39AM
(#19886401)
Nobody did the hard work to add advanced optimising backends to those projects.
They were all new ideas for the frontend - great but someone somewhere has to do the work to get it to perform well, and few people know how to do that so it never got done.
Re:GraalVM 19.0
by pawelmurias on Saturday May 11, 2019 @11:22AM
(#19886608)
Parrot never reached a point where it didn't suck. GraalVM has the futurama projection tech to make the universal vm dream possible.
Re:GraalVM 19.0
by tempguy9999 on Saturday May 11, 2019 @11:32AM
(#19886656)
futamura? :)
I recall there being several contenders 6-10 years ago. Like the Parrot VM [1], which was driven by Perl but aimed to support a large variety of languages.
[1] https://en.wikipedia.org/wiki/Parrot_virtual_machine