Print

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

[Sabin Stargem]

If we were to do that, what kind of machine is required in terms of expense?  I would guess that $400 would be for a minimal prefab, while $1,000 would cover something that is somewhere in the middle tier, depending on if Toady is willing to assemble a machine from custom parts or has a friend to handle it.  Maybe if a Drive is setup, we can have multiple points on the donation threshold for what Toady can get for the money.


01 - Performance profiling
02 - Multi-threading
03 - 64-Bit Dwarf Fortress
04 - Path memorization
05 - SSE1 & SSE2 Instructions
06 - MSX Instructions
07 - Temperatures
08 - Fluids
09 - Data & Item storage
10 - Pathfinding, Hauling Arc, activity prioritization
11 - ?
12 - ?
13 - Profit!

[NW_Kohaku]

Great, just what another multithreading thread needs, a derail into being yet another pathfinding thread. 

This is the problem with constantly rehashing the same topics in different threads: The suggestions are always fundamentally the same, and always warrant the same fundamental response for as long as we are stuck in the rut of not reading the discussions that came before.  It's not like there aren't good answers, but that the simple answers always come first, and the problems with the simple answers never change until you overcome their problems by reading the arguments came before.

Here, let me just pull up a few of the previous pathfinding suggestion threads, first of which was active just a few days ago:

http://www.bay12forums.com/smf/index.php?topic=76278.0
http://www.bay12forums.com/smf/index.php?topic=78413.0
http://www.bay12forums.com/smf/index.php?topic=42678.0
http://www.bay12forums.com/smf/index.php?topic=39067.0

Those are just some of the ones that tried something a little different, and only includes suggestions, not general discussions, which are much more prolific.

Hierarchical ("node-based" or "rectangles") pathfinding is a generally better solution - cutting the fortress into "rooms" as nodes which the game can pathfind through faster than flooding through every individual tile, provided node connection is maintained.  As the first of those threads was discussing, it's just a matter of how you organize this sort of thing.

EDIT: Oh, and Pathfinder Project... for as drama-fueled as that became, it was certainly lengthy and in-depth about pathfinding.

[orbcontrolled]

Hasn't Toady expressed in the past that he has no interest in learning parallel processing? That kind of kills most of these ideas right off the bat.

EDIT: Toady answers the question:
DF Talk #9, at 57 minutes 53 seconds, or Ctrl+F "multiple cores" on the transcript.

The part that isn't on the transcript:
Toady: hehhehhehhehhehhehhehheh
Rainseeker: Uh oh.

[K.I.L.E.R]

Wouldn't it kill future expansion of DF, 10 years later or so, if you do not include multicore processing?

There are millions of methods of adding multicore, even multithreaded agents with message parsing, an example of this would be HTML5 Web Workers.

I am not saying this would be easy, but what would be the monetary requirement for the developers to work on this option? $10,000 to $20,000?

[sockless]

Well email Toady and offer 20k for him to do multi threading. He'd probably say no.

Toady makes DF for the love of it, he makes it because he finds it fun. It isn't "developers" either, it's developer. As it has already been said, there's much more time efficient ways to improve performance that don't include multi-threading. The game isn't going to get much computationally intense than it currently is, the main slow-downs being pathfinding, temperature and fluids. IIRC you can get large increases in FPS by getting rid of temperature, but it pretty much turns magma into a slow water. I can't remember how much faster it is though.

[Jeoshua]

Actually, given that 20k is almost a year's salary... I think he'd consider it for that price tag if it was from one donor.  Because the response when it was released would be alot of donation.  But not so if 10,000 people donated $1 a piece.  Then, they'd feel they already paid for the improvements.

So where is a rich benefactor who loves ASCII when you need him?

[Kogut]

Toady makes DF for the love of it, he makes it because he finds it fun. It isn't "developers" either, it's developer.

  DF is my sole source of incoming, and I'd be assuming a massive risk
  by releasing the source.
 
  Tarn