Print

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

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

[Baughn]

As you may be aware, Dwarf Fortress' approach to OpenGL leaves something to be desired. To the point of my writing terrible hacks like DF Accelerator, even..

At any rate, Toady has decided to do something about that, and did it by releasing the relevant source code in the form of Battle Champs for others to fix.

This has been done: The code has been converted to use SDL by Bernard Helyer, and the OpenGL code has been rewritten by yours truly. It is almost complete. However, for Toady to accept the code into the next version of DF, your assistance is required: It must be tested. For this, you are required.

Download links
Windows: http://www.bay12games.com/dwarves/df_28_181_40d13_win.zip
Linux: http://www.bay12games.com/dwarves/df_28_181_40d13_linux.tar.bz2
OS X: http://www.bay12games.com/dwarves/df_28_181_40d13_osx.tar.bz2
For linux users, the following link contains a tarball with the newest fixes, which will be officially released in 40d14. It must be unpacked on top of the 40d13 tarball.
Linux 40d13 HEAD: http://brage.info/~svein/df-40d13-head.tar.bz2

For users on Linux and OS X, run the df script, don't run the dwarfort.exe executable directly.

New features in 40d13 (vs. 40d)
- Translucent tilesets are supported, using PNG.
- The DF window can be resized on the fly.
- The DF window can be zoomed on the fly using the mouse wheel. There are two modes, it defaults to mode 1. Pressing mouse wheel switches modes. Pressing F12 resets mode 1, not 2. In mode 1, the zoom from mode 2 is not visible. (And yes, you can use both at once)
- The keyboard input system has been totally rewritten:
* You can bind multiple keys to one command
* You can create keyboard macros
* Input bindings are divided into sections
- Lots of bugfixes and performance improvements.
Upcoming features in 40d14 or 40d13 HEAD
Grid size autogenerated
Fullscreen resolution autogenerated

Please download it, run it, and test with all the different tilesets and grid sizes your hearts desire. Graphics sets, too.

Bug-tracker
There should be a list of known bugs at http://github.com/Baughn/Dwarf-Fortress--libgraphics-/issues. Feel free to add yours there when you report it.
- Tiles may show black borders when zooming. Fixed, d14/d13-head
- The mouse buttons get stuck when designating. Fixed, d14/d13-head
- Various keyboard issues. Fixed, d14/d13-head

Tips and tricks
- Make sure you test all the combinations for buffers/partial printing that Toady suggests in init.txt.
- Make sure to try the debug version (not up yet) before posting about problems, if you run linux. If you don't run linux, try running linux.

Source code
For the perennially curious, or programmers, the source code (that is still under active development) can be found at git://github.com/Baughn/Dwarf-Fortress--libgraphics-.git. It uses the git version-control system, and can also be seen in a browser at http://github.com/Baughn/Dwarf-Fortress--libgraphics-/tree/master.

Compiling this code gives a functional DF, but it can only be compiled on Linux.

[Duke 2.0]

 So what does this provide that the normal DF does not? Just a speed up?

 Link to relevant page?

 Despite these questions, I have a creeping sense of awesomeness in the back of my mind. Like something grand will happen to DF soon. Like 500 dwarf fortresses and 12x12 areas. I know nothing drastic like that will happen...

 Will it?

 Edit:

 ...

...

 Wha... You...

 Alright, time to take off the 100 fps cap...

 Edit 2:

 Alright, it is now hovering around 110 FPS with ALL of the speed-lowering things on. With the highest grid size for a 9x9 tileset on a 1024 X 768 monitor.

 I... I may be a man, but I wish to have your babies. That is how staggeringly awesome this is.

[Baughn]

Just a speed-up, but for some people it's a *large* speed-up.

Especially if you want to use a 200x200 grid.

[Baughn]

Well, and a Linux version, and a properly functioning OS X version

[Davion]

Pretty big improvement. Several weird anomalies that many people have been getting, including myself, seem to have been fixed with this; chiefly minimizing doesn't make everything in the window disappear, and if you move the window around it doesn't "bleed". That's with partial print on, at least.

Also getting quite a big FPS boost and it's mostly even across partial print on/off, which I can't say the same for the original Battle Champs. Maxed FPS at 10,000 just to see how far it could go:

Optimized Battle Champs:
Partial Print Off at menus: FPS 800 or around there.
Partial Print Off at battles: FPS 2000-7000 or around there.
Partial Print On at menus: FPS is same as above mostly.
Partial Print On at battles: FPS is same as above mostly.

Original Battle Champs with the same init file and settings as above:
Partial Print Off at menus: FPS 260 or around there.
Partial Print Off at battles: FPS 1800 or around there haven't seen it even hit 2000.
Partial Print On at menus: FPS 900-1000 or around there.
Partial Print On at battles: FPS 900-1300 or around there.

[Greiger]

Unfortunately it crashes on startup for me.

Microsoft Windows XP Professional Service pack 3
Intel Celeron 2.4Ghz CPU
Radeon 9250 PCI card (running Omega drivers version 3.8.252)
1.5 GB SDRAM

Drivers and direct x are as recent as I can get the omega driver is the most recent that supports my graphics card.

Fails to run under windowed or fullscreen, partial print on or off.

Is there any other info I could give you that would be helpful?  DF would benefit greatly from a speed increase, but I would like to be able to run it too

[Baughn]

As a matter of fact, there is.

Visit http://www.realtech-vr.com/glview/, install and run the viewer, then click on report. Copy all the text in the report to a post here - preferably in a spoiler.

(The same program does allow me to see supported extensions for /some/ radeon 9250, but those are not necessarily identical to yours.)

[Baughn]

Oh yes. Also run the rendering test, and tell me the results, please.

[Greiger]

Ok doing so now.  I'll post an edit with results soon.

Here we go.  This is the report.

Spoiler (click to show/hide)

Renderer: RADEON 9200 Series DDR x86/SSE2
Vendor: ATI Technologies Inc.
Memory: 128 MB
Version: 1.3.1072 WinXP Release
Shading language version: N/A
Max number of light sources: 8
Max viewport size: 2048 x 2048
Max texture size: 2048 x 2048
Max anisotropy: 16
Max samples: 0
Max draw buffers: 0
Max texture coordinates: 0
Max vertex texture image units: 0


Extensions: 80

GL_ARB_multitexture
GL_ARB_occlusion_query
GL_ARB_point_parameters
GL_ARB_texture_border_clamp
GL_ARB_texture_compression
GL_ARB_texture_cube_map
GL_ARB_texture_env_add
GL_ARB_texture_env_combine
GL_ARB_texture_env_crossbar
GL_ARB_texture_env_dot3
GL_ARB_texture_mirrored_repeat
GL_ARB_transpose_matrix
GL_ARB_vertex_blend
GL_ARB_vertex_buffer_object
GL_ARB_vertex_program
GL_ARB_window_pos
GL_ATI_element_array
GL_ATI_envmap_bumpmap
GL_ATI_fragment_shader
GL_ATI_map_object_buffer
GL_ATI_texture_env_combine3
GL_ATI_texture_mirror_once
GL_ATI_vertex_array_object
GL_ATI_vertex_attrib_array_object
GL_ATI_vertex_streams
GL_ATIX_texture_env_combine3
GL_ATIX_texture_env_route
GL_ATIX_vertex_shader_output_point_size
GL_EXT_abgr
GL_EXT_bgra
GL_EXT_blend_color
GL_EXT_blend_func_separate
GL_EXT_blend_minmax
GL_EXT_blend_subtract
GL_EXT_clip_volume_hint
GL_EXT_compiled_vertex_array
GL_EXT_draw_range_elements
GL_EXT_fog_coord
GL_EXT_multi_draw_arrays
GL_EXT_packed_pixels
GL_EXT_point_parameters
GL_EXT_rescale_normal
GL_EXT_secondary_color
GL_EXT_separate_specular_color
GL_EXT_stencil_wrap
GL_EXT_texgen_reflection
GL_EXT_texture3D
GL_EXT_texture_compression_s3tc
GL_EXT_texture_cube_map
GL_EXT_texture_edge_clamp
GL_EXT_texture_env_add
GL_EXT_texture_env_combine
GL_EXT_texture_env_dot3
GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias
GL_EXT_texture_mirror_clamp
GL_EXT_texture_object
GL_EXT_texture_rectangle
GL_EXT_vertex_array
GL_EXT_vertex_shader
GL_HP_occlusion_test
GL_NV_blend_square
GL_NV_occlusion_query
GL_NV_texgen_reflection
GL_S3_s3tc
GL_SGI_color_matrix
GL_SGIS_generate_mipmap
GL_SGIS_multitexture
GL_SGIS_texture_border_clamp
GL_SGIS_texture_edge_clamp
GL_SGIS_texture_lod
GL_SUN_multi_draw_arrays
GL_WIN_swap_hint
WGL_ARB_extensions_string
WGL_ARB_make_current_read
WGL_ARB_pbuffer
WGL_ARB_pixel_format
WGL_ARB_render_texture
WGL_EXT_extensions_string
WGL_EXT_swap_control

Core features
v1.1 (100 % - 7/7)
v1.2 (100 % - 8/8)
v1.3 (100 % - 9/9)
v1.4 (86 % - 13/15)
v1.5 (66 % - 2/3)
v2.0 (0 % - 0/10)
v2.1 (0 % - 0/3)
v3.0 (0 % - 0/21)

OpenGL driver version check (Current: 6.14.10.5819, Latest known: 1.3.1006 WinXP Release):
Latest version of display drivers found
According the database, you are running the latest display drivers for your video card.

Compiled vertex array support
This feature improves OpenGL performance by using video memory to cache transformed vertices.

Multitexture support
This feature accelerates complex rendering such as lightmaps or environment mapping.

Secondary color support
This feature provides an alternate method of coloring specular highlights on polygons.

S3TC compression support
This feature improves texture mapping performance in some applications by using lossy compression.

Texture edge clamp support
This feature improves texturing quality by adding clamping control to edge texel filtering.

Vertex program support
This feature enables a wide variety of effects via flexible vertex programming (equivalent to DX8 Vertex Shader.)

No fragment program support
This feature enables per pixel programming (equivalent to DX9 Pixel Shader.) Some current or future OpenGL programs may require this feature.

Texture anisotropic filtering support
This feature improves the quality of texture mapping on oblique surfaces.

Occlusion test support
This feature provides hardware accelerated culling for objects.

No OpenGL Shading Language support
This may break compatibility for applications using per pixel shading.

No Frame buffer object support
This may break compatibility for applications using render to texture functions.

Extension verification:
GL_ARB_fragment_program was not found, but has the entry point glProgramStringARB
GL_ARB_fragment_program was not found, but has the entry point glBindProgramARB
GL_ARB_fragment_program was not found, but has the entry point glDeleteProgramsARB
GL_ARB_fragment_program was not found, but has the entry point glGenProgramsARB
GL_ARB_fragment_program was not found, but has the entry point glProgramEnvParameter4dARB
GL_ARB_fragment_program was not found, but has the entry point glProgramEnvParameter4dvARB
GL_ARB_fragment_program was not found, but has the entry point glProgramEnvParameter4fARB
GL_ARB_fragment_program was not found, but has the entry point glProgramEnvParameter4fvARB
GL_ARB_fragment_program was not found, but has the entry point glProgramLocalParameter4dARB
GL_ARB_fragment_program was not found, but has the entry point glProgramLocalParameter4dvARB
GL_ARB_fragment_program was not found, but has the entry point glProgramLocalParameter4fARB
GL_ARB_fragment_program was not found, but has the entry point glProgramLocalParameter4fvARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramEnvParameterdvARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramEnvParameterfvARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramLocalParameterdvARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramLocalParameterfvARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramivARB
GL_ARB_fragment_program was not found, but has the entry point glGetProgramStringARB
GL_ARB_fragment_program was not found, but has the entry point glIsProgramARB
GL_EXT_copy_texture was not found, but has the entry point glCopyTexSubImage3DEXT
GL_EXT_polygon_offset was not found, but has the entry point glPolygonOffsetEXT
GL_EXT_subtexture was not found, but has the entry point glTexSubImage3DEXT
GL_ATI_pn_triangles was not found, but is available in driver version 6.14.10.4776
GL_KTX_buffer_region was not found, but is available in driver version 6.14.10.3604
GL_SGI_texture_edge_clamp was not found, but is available in driver version 6.14.10.3604

Not an expert on open GL but that didn't look promising.

And the render.
Errors occured during the tests.

Spoiler (click to show/hide)

...Legacy PFD creation.:
SetPixelFormat ID:3 failed: The operation completed successfully.


Temporary context creation failed: The operation completed successfully.
 - 1024 x 768, 32 (AAx0)

...DestroyWindow:
ReleaseDC failed: Invalid device context (DC) handle.

Test results show as everything passing.  With 0 FPS.  That doesn't seem right to me, but that's what it shows.

[savanik]

Pretty big improvement.

<snip>

Optimized Battle Champs:
Partial Print Off at menus: FPS 800 or around there.
Partial Print Off at battles: FPS 2000-7000 or around there.
Partial Print On at menus: FPS is same as above mostly.
Partial Print On at battles: FPS is same as above mostly.

I never ran the original, but I'm getting around 8000 fps during battles with partial-print ON, about 4000 with it OFF.

I also have a glitch in this version - running under Linux, if you use 'WINDOWED:NO', the window does not maximize properly, but flickers black breifly before it remains in a window using the Fullscreen resolution. I also got occasional black flickering, as if it were attempting to go fullscreen again but was failing.

If you use 'WINDOWED:PROMPT' and select fullscreen, then you get fullscreen working properly.

This is under Ubuntu 8.10, using an NVidia 8800 GTX with the properitary NVidia drivers loaded. (Not sure if those make a difference, but thought I'd mention.)

Also, take this with a grain of salt, because I'm running in dual-screen and took some heavy-handedness to xorg.conf to achieve this. The program itself runs beautifully (SO much faster than Wine!)... and I don't use fullscreen much, so don't mind it much.

Savanik

[Baughn]

Going fullscreen, it'll actually try to change the screen resolution. If that fails.. something like what you're seeing might happen; I haven't checked that code.

Make sure you've edited init.txt to set the /right/ fullscreen resolution.


EDIT: Hm, actually it just segfaults. I'll have to look at that.

[Baughn]

Ok doing so now.  I'll post an edit with results soon.

*snip*

Test results show as everything passing.  With 0 FPS.  That doesn't seem right to me, but that's what it shows.

Your card is probably fine, but it seems that your drivers are buggy. Considering how large the chance of fixing them is..

I was considering adding a software fallback, but it'd be slow as a dead turtle on that hardware, so not really much point of that either. Sorry.

There is one other option. Linux has an open-source driver for the 9000-series radeons that, miracle of miracles, supports 3D rendering. You should try that.

EDIT: Actually, one last test..
If you run bc.exe from a terminal window (cmd.exe), you should get some debugging output. Does it mention VBOs or arrays?

[Thndr]

Tried this out, got a max of around 4000 breifly, more commonly between 1300 and 2100 depending on size of battle. For some odd reason the slower moving attacks with lesser units go take up framerate, but the initial speedy attack goes the fastest.

With mouse moving//on screen, framerate drops to 600-700

[Christes]

It's ... so ... fast

[userpay]

Hmm I guess I'm getting aroun 1500-2000 during battles but its difficult to tell which side is winning.