Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Silverionmox]

"Training dummies" is marked as [resolved] by its owner, probably because safe training was more important to him than the actual dummies.

Gender differentiation is partly done (castes, appearances), although nowhere near finished indeed. It might benefit from some reformulation.

[Footkerchief]

Ah, didn't see that about the dummies.

[Draco18s]

"Training dummies" is marked as [resolved] by its owner, probably because safe training was more important to him than the actual dummies.

I own it, and that's pretty much why.  It's not my suggestion, but one that I added to the database.  The reason for the suggestion was averted with the new combat drills and training weapons, so its not needed anymore.

[Beeskee]

I had an idea for the "vein mining" suggestion but couldn't find the thread for it. How about something that works the way designating mining works, but has settings like a stockpile? So you could select an area and then set which things to actually mine. A few handy presets like 'metal and gems' and some space for custom presets would make this a great tool to have.

[Hondo]

Simplest way would be to add an "automine" designation, which when dug out would autodesignate anything of the same type in a 3x3 search square to also be automined.

I'm wary of solutions which involve more menus with weird interfaces.

[SoulSkorpion]

I wish Toady would add multithreading. It's really not that hard to do if you can split up the work to be done (at a guess pathfinding would be a good candidate), and would make performance so much better. Especially with a nice clean library like boost::thread. Hell, I'd even share code I've written if it'd help.

[Osmosis Jones]

I wish Toady would add multithreading. It's really not that hard to do if you can split up the work to be done (at a guess pathfinding would be a good candidate), and would make performance so much better. Especially with a nice clean library like boost::thread. Hell, I'd even share code I've written if it'd help.

The fact that the game has not been built to be multithreaded means any attempt to do so would require a complete overhaul of the code, which would take a very long time. Time which could be better spent doing other things.

Also, Toady doesn't know enough about multithreading to do it himself, and he is very much unlikely to share his code to others to let them do so.

[SoulSkorpion]

The fact that the game has not been built to be multithreaded means any attempt to do so would require a complete overhaul of the code, which would take a very long time

Well, yes and no. It depends how the code is structured already. Say for example there's a loop that looks like this:

Code: [Select]

//pseudocode
for each dwarf
{
  dwarf.update();
}
Say update() takes a while to run, maybe it's working out the dwarf's current activity and/or pathfinding to get there.

With some decent thread management code, you can change it to this:

Code: [Select]

//pseudocode
for each dwarf
{
  threadManager.QueueTask(dwarf, dwarf::update);
}
wait for threads to finish

...where ThreadManager::QueueTask gives the function to one of the threads.

Of course you need to write/use-a-library-of the thread management code, but my point is that you don't necessarily have to make sweeping changes to the code. You just parallelise slow bits to speed them up.

[edit]Oh, and it doesn't take long to write a thread management library either. Mine's taking me about a week all up.[/edit]

[Shades]

[edit]Oh, and it doesn't take long to write a thread management library either. Mine's taking me about a week all up.[/edit]

Alternatively, rather than reinventing the wheel you could just use something like TBB, your example in particular is well suited for it.

Of course which well tested, fast and efficient library to use would depend on how you structure your code.

[SoulSkorpion]

[edit]Oh, and it doesn't take long to write a thread management library either. Mine's taking me about a week all up.[/edit]

Alternatively, rather than reinventing the wheel you could just use something like TBB, your example in particular is well suited for it.

Well, yes. I just brought it up to illustrate that multithreading stuff doesn't necessarily take ages to implement. I'm rolling my own as an exercise. Obviously if Toady's ever going to use multithreading he should use someone else's library.

[Draco18s]

multithreading

For the love of....

It's been discussed, in threads more relevant to it than this one.  Go post there.  Or better yet, read those posts about why it won't happen and STFU because it's not going to happen.

[Footkerchief]

Cool it down please.

[Shades]

Well, yes. I just brought it up to illustrate that multithreading stuff doesn't necessarily take ages to implement. I'm rolling my own as an exercise. Obviously if Toady's ever going to use multithreading he should use someone else's library.

Always the best way to learn something I find

It's been discussed, in threads more relevant to it than this one.  Go post there.  Or better yet, read those posts about why it won't happen and STFU because it's not going to happen.

And it will be discussed again and again and again like many other subjects as new people bring it up, and each time I'll make the same points I did last time it was discussed. Also it will happen eventually as individual cores aren't getting faster quickly enough but I agree it won't happen any time soon.

Cool it down please.

We are all cool and happy bunnies don't mind us.

[Draco18s]

It's been discussed, in threads more relevant to it than this one.  Go post there.  Or better yet, read those posts about why it won't happen and STFU because it's not going to happen.

And it will be discussed again and again and again like many other subjects as new people bring it up, and each time I'll make the same points I did last time it was discussed. Also it will happen eventually as individual cores aren't getting faster quickly enough but I agree it won't happen any time soon.

But not in this thread.  That's my point.

[Shades]

But not in this thread.  That's my point.

Fair enough, you know what threads need though is a way of forking off a conversion sensibly.