JVM support for arbitrary machine code loading and execution

[Paul Sandoz]

> On 4 Dec 2015, at 18:20, Vladimir Ivanov <vladimir.x.ivanov at oracle.com> wrote:
> 
> 
> On 12/4/15 8:11 PM, Vitaly Davidovich wrote:
>> Wow, that's awesome.  Could something like this ever be exposed to user
>> code (java's version of inline asm)? (ducks under the desk :))
> Thanks, Vitaly.
> 

Really nice work Vladimir.


> Why not? That's exactly what we try to find out in Project Panama :-)
> 

See also here for some teasers (slide 42):

  http://cr.openjdk.java.net/~psandoz/conferences/2015-JavaOne/j1-2015-unsafe-CON7076.pdf <http://cr.openjdk.java.net/~psandoz/conferences/2015-JavaOne/j1-2015-unsafe-CON7076.pdf>

For some context i used an early prototype of Vladimir’s excellent work to experiment with SIMD array mismatch and that fed nicely into the JavaOne Unsafe presentation.

I had some fun playing around this this and would encourage others to do the same. Now this is public perhaps i can make my example public too.

Paul.