Tasks, actors and garbage collection

[bachmeier]

On Monday, 26 April 2021 at 17:34:41 UTC, russhy wrote:
> On Monday, 26 April 2021 at 15:38:48 UTC, Adam D. Ruppe wrote:
>> On Monday, 26 April 2021 at 15:30:48 UTC, russhy wrote:
>>> You guys never made a game or a game engine and you are 
>>> telling me GC is fine
>>
>> Which games have you made?
>>
>> I made a buggy version of Minesweeper once. You can play it 
>> online actually http://webassembly.arsdnet.net/minesweeper
>
> I'm working on a 3D game engine
>
> yeah minesweeper can have a 5 second GC pause, users won't 
> notice

If you want to be precise, this is a matter of implementation of 
GC. There are *some games* for which *current GC implementations* 
result in unacceptable performance issues. Many will do just 
fine. There's also no reason to pretend manual memory management, 
reference counting, or garbage collection is a silver bullet. 
Allocation/freeing of memory always results in a performance hit 
no matter what name you apply. As [stated in this blog 
post](https://www.sebastiansylvan.com/post/on-gc-in-games-response-to-jeff-and-casey/):

> Jeff and Casey’s ranting is not actually about the GC itself, 
> but about promiscuous allocation behavior, and I fully agree 
> with that, but I think it’s a mistake to conflate the two. GC 
> doesn’t imply that you should heap allocate at the drop of a hat

I'm not arguing that the GC will work for all games because it 
won't. You give the impression of having a fundamentalist 
position though that "real games" (not Minesweeper) cannot have a 
GC and that GC is obviously going to be slow (uncontrollable 5 
second pauses). That's not a helpful starting point.