State of Loom

[Ron Pressler]

A branch in a blocking method should have a negligible effect, even if it 
isn’t elided altogether. There are bigger implementation changes all the
time, so that this kind of detail is below the noise level.

— Ron


On 27 May 2020 at 15:11:19, Mike Rettig (mike.rettig at gmail.com) wrote:

On Wed, May 27, 2020 at 7:08 AM Ron Pressler <ron.pressler at oracle.com> wrote:  
>There is no impact Of any kind on code that does not block virtual threads.  
>  
> — Ron  

Can you elaborate on how that is achieved? In the jvm when a blocking  
call occurs, there are two different behaviors for an OS thread and a  
virtual thread. Is the cost avoided by the os thread via dynamic  
dispatch? How do you prevent even a small performance impact for OS  
threads?  

Mike