News for hackers. Stuff that matters.

A comment on Making WebAssembly better for Rust and for all languages. Parent comment

Re:Making WebAssembly better for Rust and for all languages by ZenPsycho on Thursday March 15, 2018 @12:13AM (#16590362)
you're reminding me of this 30 minute talk from 2014 "the birth and death of javascript", in which he supposes a future where CPUs basically run webassembly directly https://www.destroyallsoftware.com/talks/the-birth-and-death...
[ Parent | on HN ]
Re:Making WebAssembly better for Rust and for all languages by icebraining on Thursday March 15, 2018 @02:08AM (#16590748)
Reminds me of Jazelle[1], which allowed some ARM CPUs to run Java bytecode directly. AFAIK it never really caught on.

[1] https://en.wikipedia.org/wiki/Jazelle

[ Parent | on HN ]
Re:Making WebAssembly better for Rust and for all languages by baobrien on Thursday March 15, 2018 @11:23AM (#16593617)
It also probably didn't help that Jazelle was locked behind an NDA.
[ Parent | on HN ]
Re:Making WebAssembly better for Rust and for all languages by ZenPsycho on Thursday March 15, 2018 @08:06PM (#16597603)
not exactly, the idea is to have a normal cpu like arm or x86 but the kernel doesn't ever deal with native binaries, only web assembly, and the performance penalty of running stuff in a VM gets offset by disabling memory protection and its performance penalty, since memory protection is already guaranteed by the VM.
[ Parent | on HN ]