Introducing the WebKit FTL JIT

Re:Introducing the WebKit FTL JIT by lobster_johnson on Tuesday May 13, 2014 @11:04PM (#7741975)

Is this architecture generic enough that one could, say, build a Ruby compiler on top of it?

I imagine even writing a Ruby -> JS transpiler that used the WebKit VM would provide a speedup, similar to how JRuby works on the JVM, but native compilation would be even better.

[ Parent | on HN ]

Re:Introducing the WebKit FTL JIT by sanxiyn on Tuesday May 13, 2014 @11:25PM (#7742025)

Yes.

Moreover, Dropbox is trying roughly similar architecture for their Python implementation.

https://github.com/dropbox/pyston

[ Parent | on HN ]

Re:Introducing the WebKit FTL JIT by steveklabnik on Wednesday May 14, 2014 @01:40AM (#7742376)

You might be able to try that with http://opalrb.org/

Rubinius already does Ruby -> LLVM.

[ Parent | on HN ]

Re:Introducing the WebKit FTL JIT by ksec on Wednesday May 14, 2014 @02:28AM (#7742475)

But as much as i hate to admit Rubinius is still slow.

[ Parent | on HN ]

Re:Introducing the WebKit FTL JIT by om2 on Wednesday May 14, 2014 @03:27AM (#7742618)

Part of this project was building support into LLVM that you would need to build a really great dynamic language JIT. Another key aspect was having lower compilation tiers that don't use LLVM. Using LLVM as a fourth tier works much better than using it as a first or even second tier (we have tried in the past).

[ Parent | on HN ]