RFR[L]: 8237767 Field layout computation overhaul

[Aleksey Shipilev]

On 2/3/20 5:00 PM, Frederic Parain wrote:
> The cause of this regression is the way the new algorithm allocates
> reference fields: it looks for a space big enough to allocate them
> contiguously (so there’s a single oopmap for all oops declared in
> the current class, and less work for the GC to find these oops).

Oh!

Seems to me that "less work for GC to find these oops" is more theoretical than a practical loss of
object density due to this artificial inflation. Single oop map per class is nice too, unless it
leads to this...

As I look through the layouter going through my local Maven Central snapshot, this case does not
look rare for us not to care. Looks like every class with a long and 3+ oop fields is regressing in
instance size, which is in line with your explanation. I believe this needs to be resolved before
releasing the new layouter in the wild.

-- 
Thanks,
-Aleksey