Re: Merging scratch/no-purespace to remove unexec and purespace

[Helmut Eller]

On Sun, Dec 22 2024, Stefan Kangas wrote:

> Helmut Eller <eller.helmut@gmail.com> writes:
[...]
>> between the versions with and without the patch.  The results was:
>>
>>    without-pure-section: 0.006251480181
>>       with-pure-section: 0.003986384857
>
> This is interesting, thanks.  Would this experiment easily transfer to
> and be relevant to the MPS branch?

To be useful for MPS, we would probably need sections that can be loaded
to different/non-contiguous memory blocks that are allocated by MPS.
(One section per MPS pool).  That would require some work on the
pdumper.

It's less relevant for MPS because MPS is incremental and generational
and so pure objects contribute much less to the average GC pause time.

Helmut