Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  

Poll

After experimenting with the options, how is 40d13? Problems only count if the defaults don't work.

Faster than 40d, no problems
- 42 (26.1%)
Faster than 40d, problems
- 72 (44.7%)
No slower than 40d, no problems
- 14 (8.7%)
No slower than 40d, problems
- 16 (9.9%)
Slower than 40d, no problems
- 2 (1.2%)
Slower than 40d, problems
- 3 (1.9%)
Doesn't work (please explain)
- 12 (7.5%)

Total Members Voted: 160


Pages: 1 ... 61 62 [63] 64 65 ... 147

Author Topic: FotF: Help test the output code for the next version of DF (40d13)  (Read 373776 times)

Baughn

  • Noble Phantasm
  • The Haruhiist
  • Hiss
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #930 on: February 19, 2009, 06:07:38 pm »

Don't suppose you found any solutions?

snooptodd: Try disabling compiz, or better yet disable compositing entirely. That could be interfering; ISTR there was some issue with double-buffering not working under compiz.

Oh, and you could /turn off/ double-buffering in init.txt by turning single-buffer on, which might also work. Actually, try that first.
Logged
C++ makes baby Cthulhu weep. Why settle for the lesser horror?

numerobis

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #931 on: February 19, 2009, 06:16:57 pm »

You don't seem to be using the system keyrepeat rate, which indicates you're actually figuring out keyrepeat all by your lonesome.  Also, sometimes a key 'sticks' and I can unstick it by hitting the key again, which indicates etc etc.  Is there a strong reason to do this -- i.e. isn't that something OpenGL or SDL or whatever should be dealing with?

Another thing, it seems like a game frame goes by every time I hit a key -- so I get meetings occurring while paused and so on.  Once again, I'm not entirely convinced of my diagnosis, but if so, that seems like a bug.
Logged

Exponent

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #932 on: February 19, 2009, 06:37:47 pm »

Another thing, it seems like a game frame goes by every time I hit a key -- so I get meetings occurring while paused and so on.  Once again, I'm not entirely convinced of my diagnosis, but if so, that seems like a bug.

That's been a long-standing problem which doesn't seem to have anything to do with these graphic-code versions.  (Bug 768)
Logged

Baughn

  • Noble Phantasm
  • The Haruhiist
  • Hiss
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #933 on: February 20, 2009, 08:17:15 am »

You don't seem to be using the system keyrepeat rate, which indicates you're actually figuring out keyrepeat all by your lonesome.  Also, sometimes a key 'sticks' and I can unstick it by hitting the key again, which indicates etc etc.  Is there a strong reason to do this -- i.e. isn't that something OpenGL or SDL or whatever should be dealing with?

Yes, SDL doesn't have quite the right behaviour when dealing with key-repeat. I used to have it do all the work, but then if you pressed a, pressed b and then released a, b would never repeat - annoying for arrow keys.

The code in 40d9 is still somewhat buggy, though. Patches were sent to Toady a very long time ago; I guess he'll get around to them once the next release is ready.
Logged
C++ makes baby Cthulhu weep. Why settle for the lesser horror?

snooptodd

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #934 on: February 20, 2009, 10:08:09 am »

Don't suppose you found any solutions?

snooptodd: Try disabling compiz, or better yet disable compositing entirely. That could be interfering; ISTR there was some issue with double-buffering not working under compiz.

Oh, and you could /turn off/ double-buffering in init.txt by turning single-buffer on, which might also work. Actually, try that first.
ok, it works this morning.
there must have been a fix uploaded yesterday unfortunately im not getting local mail so i cant tell what was updated.
i did see a post over at the ubuntu formus that seems to confirm my suspicion.

edit:
yep sdl was upgraded
[UPGRADE] libsdl1.2debian 1.2.13-4ubuntu2 -> 1.2.13-4ubuntu3
[UPGRADE] libsdl1.2debian-alsa 1.2.13-4ubuntu2 -> 1.2.13-4ubuntu3
« Last Edit: February 20, 2009, 10:35:08 am by snooptodd »
Logged

MuonDecay

  • Bay Watcher
  • Say hello to my little μ
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #935 on: February 23, 2009, 02:04:38 am »

Wow, I just now get around to trying this.

You know what? It's great! It's noticeably faster, even on a computer with a powerful video card. I guess the DF graphics were just shoving more data than necessary through my CPU bottleneck.

And as a bonus side effect, there is no longer a 3-5 second delay when I switch between DF and another windowed game. Now it's instantaneous. Handy.
Logged

numerobis

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #936 on: February 23, 2009, 02:52:08 am »

Yes, SDL doesn't have quite the right behaviour when dealing with key-repeat. I used to have it do all the work, but then if you pressed a, pressed b and then released a, b would never repeat - annoying for arrow keys.
Ewwww.

Do the patches you sent include attempts to figure out the system keyrepeat rate?  It seems to be queryable via SDL_GetKeyRepeat(), maybe.
Logged

Baughn

  • Noble Phantasm
  • The Haruhiist
  • Hiss
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #937 on: February 23, 2009, 04:22:25 am »

No, it's set via init.txt, same as for 40d.
Logged
C++ makes baby Cthulhu weep. Why settle for the lesser horror?

Topace3k

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #938 on: February 23, 2009, 06:24:59 pm »

Just stopping by to say that 40d9 with Yes:0 No No brought my 100+ dwarf multifeature fort currently in the midst of the massive HFS slowdown up from a crippling ~14 fps to a painful but bearable ~25 fps.  Thanks!
Logged

Flok Speargrabber

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #939 on: February 23, 2009, 06:38:15 pm »

So I boot this up and load up a fortress that gave me ~50 - 70 FPS on .40d.
I get ~ 10 - 20 FPS on 40d9.
The fortress itself is a decent one; 2x2 with 101 pop, and ~11 animals.

Computer specs:
Windows XP
Pentium 4 Processor 2.8 GHz
2.12 GB RAM
ATI Raedeon 9800 PRO
I think that's all the relevant information?
Logged

Topace3k

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #940 on: February 23, 2009, 06:52:47 pm »

What kind of frame refresh combination are you using from the init?
Logged

Flok Speargrabber

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #941 on: February 23, 2009, 07:41:13 pm »

What kind of frame refresh combination are you using from the init?

[FPS_CAP:100]
[G_FPS_CAP:50]
Logged

Topace3k

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #942 on: February 23, 2009, 09:11:34 pm »

What kind of frame refresh combination are you using from the init?

[FPS_CAP:100]
[G_FPS_CAP:50]

I meant this section:
[PARTIAL_PRINT:YES:0]
[FRAME_BUFFER:YES]
[SINGLE_BUFFER:NO]
Logged

Random832

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #943 on: February 23, 2009, 11:44:28 pm »

Well, yeah, but you're supposed to run the scripts, not dwarfort.exe directly. :P

I named it that specifically *because* unix users tend to assume executables shouldn't really be named .exe. Perhaps I should put it in a libexec subdirectory instead.

The convention for executable binaries that aren't supposed to be started directly (firefox, openoffice, etc) is to name it ending in -bin (and name the shell script with no extension)
Logged

Flok Speargrabber

  • Bay Watcher
    • View Profile
Re: Future of the fort: Help test the output code for the next version of DF
« Reply #944 on: February 24, 2009, 12:06:15 am »

What kind of frame refresh combination are you using from the init?

[FPS_CAP:100]
[G_FPS_CAP:50]

I meant this section:
[PARTIAL_PRINT:YES:0]
[FRAME_BUFFER:YES]
[SINGLE_BUFFER:NO]
I can't find that in the init text file.
« Last Edit: February 24, 2009, 01:53:15 am by Flok Speargrabber »
Logged
Pages: 1 ... 61 62 [63] 64 65 ... 147