Print

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

[Kogut]

I think that it will require a LOT of work and will help only people using CUDA.

[Phmcw]

Fps death will alway be an issue. The map is just huge, and the dwarf will get cleverer and cleverer. (which mean more and more processing power will be needed).
You'll just have to take a map large enough, and you'll bring any processors to his knees.

Now there is lot of room for improvement, and toady will get to that eventually.
I don't think gpu will be used efficiently, as toady would have to learn and implement both cuda and ati's equivalent (and df would require a powerful graphic card, wouldn't work on any competitor's graphic card, and would be threatened by any change by nvidia and ati).
Maybe if he release it open source, but he won't do that until it's finished, and even then it's not sure.

[Durin Stronginthearm]

It won't always be an issue, because Toady will die one day.

Or he'll stop adding features. But perhaps that's a fate worse than death?

[Kogut]

Or he'll stop adding features. But perhaps that's a fate worse than death?

For me "fate worse than death" in this topic means great game with bugs (and low FPS) ruining everything.

===

On topic:
"Will we ever get to a point where forts don't die FPS deaths?" - no, always we can start 25x25 embark with 100 z-levels of air, 900 of rock, 500 z levels waterfall. And maybe 3900 dwarves and 100k goblins on top of it.

But constant lowering of FPS for no apparent reason*, multiple CPU eating bugs - it is other story.

[ * ] I have this in second year - FPS from 80 to 40 with constant number of items and fort inhabitants.

[Encased in burning magma]

Hi, i've been paging trough this thread a little and i feel i bit out of place(Math breaks my brain for some reason) But you know the nvidia has a little side project running using CUDA for repettetive tasks and that the cuda source is open for develepors (look at boinc for example) I think Df is a prime example of a program that could use the brute force of recently releaced video cards?

Or am i making a REALLY dump statement here 

There are portability problems. I didn't look into it since this summer, but running it on a radeon card was hell of a bitch.
The code is possibly different for mac/linux/windows, too. Ah, and it can get obsolete.
And it's horribly complicated, and more appropriate for matrix operations anyway.

[Qinetix]

It won't always be an issue, because Toady will die one day.

Or he'll stop adding features. But perhaps that's a fate worse than death?

Are you sure? Someone will take his place... hopefully and won't transform DF into something bad

[Kogut]

It won't always be an issue, because Toady will die one day.

Or he'll stop adding features. But perhaps that's a fate worse than death?

Are you sure? Someone will take his place... hopefully and won't transform DF into something bad

It is unlikely that his testament includes what should be done with DF source code.

[Encased in burning magma]

It is unlikely that his testament includes what should be done with DF source code.

He said in an interview or something he would release it on death.

[Naros]

It is unlikely that his testament includes what should be done with DF source code.

He said in an interview or something he would release it on death.

Morbid, but true. And considerate of him.

As for FPS death. *RAGE!*
Buying parts for a new PC is largely influenced by how well DF will run on it now, I find. o.o;

[Mel_Vixen]

Toady releases it on his death yes but only if it was a "natural" one - thus if nobody tried to murder him or arranged a unfortunate accident. He did so just in case some nutjob fan gets ideas.

[Lemi]

Even when using the most powerful intel processor available to commercial markets, I can only get the game to maintain 100 FPS up to 80 dwarves. Even it will become unbearable if I embark anything larger then the default area and get more then 150 dwarves. DF needs serious optimization and/or multithreading before FPS deaths become less common, but for now I just cap how many dwarves I get so it remains comfortable to play.

[Tellemurius]

Hi, i've been paging trough this thread a little and i feel i bit out of place(Math breaks my brain for some reason) But you know the nvidia has a little side project running using CUDA for repettetive tasks and that the cuda source is open for develepors (look at boinc for example) I think Df is a prime example of a program that could use the brute force of recently releaced video cards?

Or am i making a REALLY dump statement here 

There are portability problems. I didn't look into it since this summer, but running it on a radeon card was hell of a bitch.
The code is possibly different for mac/linux/windows, too. Ah, and it can get obsolete.
And it's horribly complicated, and more appropriate for matrix operations anyway.

we would have to use openCl if you want to use the graphics cards for calculations without losing compatibility. thing is that new radeon chips has a massive number of cores over nvidia (MADD cores from 2008 are from 800 to i think now 1700), while nividia cores is more than half that number (80 to 512).

[abadidea]

This thread has been extremely informative - on FPS improving tips, DF internals, and coding in general.

And of course it's informed me that "someone" is an immature skiddie who can't distinguish being a good programmer from knowing how to twiddle bits in IDA. A totally legal copy of IDA, I'm sure. Sorry, just had to get that off my chest after reading the entire thread straight through.

[RiotHouse]

If you think you have FPS problems, try all of the above and then throw FRAPS into the mix.

One of the things i've noticed is that if you want to get rid of stone for FPS you can set your craftsdwarves to repeat rock crafts.  Then dump all that onto the merchants when they come.  Gets rid of a ton of stone and gives you something else (maybe more useful) for your fortress.  It won't solve the problem but there are a few in game methods that are not as inventive as quantom smashing etc.

The best we can do is try to streamline our Forts, which means designing them to be efficient.  Instead of mining out sprawl we can have a planned setup where Dwarves don't path all over the place in order to do something.  Keep everything centralized and functional.

Just my 2 cents