Allegro.cc - Online Community

Allegro.cc Forums » Programming Questions » draw_trans_sprite, Devcpp and optimizations

This thread is locked; no one can reply to it. rss feed Print
draw_trans_sprite, Devcpp and optimizations
Paul whoknows
Member #5,081
September 2004
avatar

I am writting a game using only sprites with alpha channel, obviously I am using only draw_trans_sprite for blitting.
The weird thing is that without optimization my game runs at 60fps, but when I use the compiler optimization options(devcpp): -O1, -O2 or -O3 the fps drops to 35 or less! what´s wrong here? code compiled with optimization options should run faster, I think.

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Milan Mimica
Member #3,877
September 2003
avatar

It should run faster, unless there is a compiler bug. What is the version of gcc? It doesn't happen often to hit a bug with just a -O1.

Paul whoknows
Member #5,081
September 2004
avatar

Dev-C++ 5.0 beta 9.2 (4.9.9.2) with Mingw/GCC 3.4.2
I get it from here
The frame drop happens with draw_trans_sprite only, I am also using Fladimir´s blender routines and they run faster(as expected) with -O3, so I am not sure about the compiler bug.

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Milan Mimica
Member #3,877
September 2003
avatar

Sounds like a compiler bug that is triggered by allegro's draw_trans_sprite(). But someone with the same compiler version should confirm.

Fladimir da Gorf
Member #1,565
October 2001
avatar

Could you try out with -O2? -O3 may turn on some experimental stuff, that may or may not speed things up...

OpenLayer has reached a random SVN version number ;) | Online manual | Installation video!| MSVC projects now possible with cmake | Now alvailable as a Dev-C++ Devpack! (Thanks to Kotori)

Paul whoknows
Member #5,081
September 2004
avatar

Quote:

Could you try out with -O2?

draw_trans_sprite runs slower with optimizations enabled, I tried -O1, -O2, -O3, there is no difference it works slower, really slower, 60fps went to 35 fps.

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Fladimir da Gorf
Member #1,565
October 2001
avatar

Maybe there's something fishy in your FPS calculations? After all, I bet you're not compiling Allegro with different settings...

OpenLayer has reached a random SVN version number ;) | Online manual | Installation video!| MSVC projects now possible with cmake | Now alvailable as a Dev-C++ Devpack! (Thanks to Kotori)

Dustin Dettmer
Member #3,935
October 2003
avatar

Do you do a complete recompile each time you switch the -On?

Try recompiling allegro with all asm turned off and see if the difference still occurs.

Paul whoknows
Member #5,081
September 2004
avatar

I compiled exalpha.c example 3 times, with -O3, -O2, and without optimization, these 3 files are attached, try them in your PCs.

Here is the exalpha source (slightly modified) I used.

#SelectExpand
1/* 2 * Example program for the Allegro library, by Shawn Hargreaves. 3 * 4 * This program demonstrates how to use the 32 bit RGBA 5 * translucency functions to store an alpha channel along with 6 * a bitmap graphic. Two images are loaded from disk. One will 7 * be used for the background and the other as a sprite. The 8 * example generates an alpha channel for the sprite image, 9 * composing the 32 bit RGBA bitmap during runtime, and draws 10 * it at the position of the mouse cursor. 11 */ 12 13 14#include <allegro.h> 15 16 17 18int main(int argc, char *argv[]) 19{ 20 char buf[256]; 21 BITMAP *background; 22 BITMAP *alpha; 23 BITMAP *sprite; 24 BITMAP *buffer; 25 int bpp = -1; 26 int ret = -1; 27 int x, y, c, a; 28 29 if (allegro_init() != 0) 30 return 1; 31 install_keyboard(); 32 install_mouse(); 33 install_timer(); 34 35 /* what color depth should we use? */ 36 if (argc > 1) { 37 if ((argv[1][0] == '-') || (argv[1][0] == '/')) 38 argv[1]++; 39 bpp = atoi(argv[1]); 40 if ((bpp != 15) && (bpp != 16) && (bpp != 24) && (bpp != 32)) { 41 allegro_message("Invalid color depth '%s'\n", argv[1]); 42 return 1; 43 } 44 } 45 46 if (bpp > 0) { 47 /* set a user-requested color depth */ 48 set_color_depth(bpp); 49 ret = set_gfx_mode(GFX_AUTODETECT, 640, 480, 0, 0); 50 } 51 else { 52 /* autodetect what color depths are available */ 53 static int color_depths[] = { 16, 15, 32, 24, 0 }; 54 for (a=0; color_depths[a]; a++) { 55 bpp = color_depths[a]; 56 set_color_depth(bpp); 57 ret = set_gfx_mode(GFX_AUTODETECT, 640, 480, 0, 0); 58 if (ret == 0) 59 break; 60 } 61 } 62 63 /* did the video mode set properly? */ 64 if (ret != 0) { 65 set_gfx_mode(GFX_TEXT, 0, 0, 0, 0); 66 allegro_message("Error setting %d bit graphics mode\n%s\n", bpp, 67 allegro_error); 68 return 1; 69 } 70 71 /* load the background picture */ 72 replace_filename(buf, argv[0], "allegro.pcx", sizeof(buf)); 73 background = load_bitmap(buf, NULL); 74 if (!background) { 75 set_gfx_mode(GFX_TEXT, 0, 0, 0, 0); 76 allegro_message("Error reading %s!\n", buf); 77 return 1; 78 } 79 80 /* make a copy of it */ 81 set_color_depth(32); 82 sprite = create_bitmap(background->w, background->h); 83 blit(background, sprite, 0, 0, 0, 0, background->w, background->h); 84 85 /* load the alpha sprite image. Note that we specifically force this 86 * to load in a 32 bit format by calling set_color_depth(). That is 87 * because the disk file is actually only a 256 color graphic: if it 88 * was already a 32 bit RGBA sprite, we would probably want to use 89 * set_color_conversion(COLORCONV_NONE) instead. 90 */ 91 replace_filename(buf, argv[0], "mysha.pcx", sizeof(buf)); 92 alpha = load_bitmap(buf, NULL); 93 if (!alpha) { 94 set_gfx_mode(GFX_TEXT, 0, 0, 0, 0); 95 allegro_message("Error reading %s!\n", buf); 96 return 1; 97 } 98 99 /* normally we would have loaded an RGBA image directly from disk. Since 100 * I don't have one lying around, and am too lazy to draw one (or I could 101 * rationalise this by saying that I'm trying to save download size by 102 * reusing graphics :-) I'll just have to generate an alpha channel in 103 * code. I do this by using greyscale values from the mouse picture as an 104 * alpha channel for the Allegro image. Don't worry about this code: you 105 * wouldn't normally need to write anything like this, because you'd just 106 * get the right graphics directly out of a datafile. 107 */ 108 drawing_mode(DRAW_MODE_TRANS, NULL, 0, 0); 109 set_write_alpha_blender(); 110 111 for (y=0; y<sprite->h; y++) { 112 for (x=0; x<sprite->w; x++) { 113 c = getpixel(alpha, x, y); 114 a = getr(c) + getg(c) + getb(c); 115 a = MID(0, a/2-128, 255); 116 117 putpixel(sprite, x, y, a); 118 } 119 } 120 121 destroy_bitmap(alpha); 122 123 set_color_depth(bpp); 124 125 /* darken the background image down a bit */ 126 drawing_mode(DRAW_MODE_TRANS, NULL, 0, 0); 127 set_multiply_blender(0, 0, 0, 255); 128 rectfill(background, 0, 0, background->w, background->h, 129 makecol(32, 16, 128)); 130 solid_mode(); 131 132 /* create a double buffer bitmap */ 133 buffer = create_bitmap(SCREEN_W, SCREEN_H); 134 135 /* scale the background image to be the same size as the screen */ 136 stretch_blit(background, buffer, 0, 0, background->w, background->h, 137 0, 0, SCREEN_W, SCREEN_H); 138 139 textprintf_ex(buffer, font, 0, 0, makecol(255, 255, 255), -1, 140 "%dx%d, %dbpp", SCREEN_W, SCREEN_H, bpp); 141 142 destroy_bitmap(background); 143 background = create_bitmap(SCREEN_W, SCREEN_H); 144 blit(buffer, background, 0, 0, 0, 0, SCREEN_W, SCREEN_H); 145 146 set_alpha_blender(); 147 while (!keypressed()) { 148 /* draw the alpha sprite */ 149 x = mouse_x - sprite->w/2; 150 y = mouse_y - sprite->h/2; 151 152 // set_alpha_blender(); 153 draw_trans_sprite(buffer, sprite, x, y); 154 draw_trans_sprite(buffer, sprite, x, y); 155 draw_trans_sprite(buffer, sprite, x, y); 156 draw_trans_sprite(buffer, sprite, x, y); 157 draw_trans_sprite(buffer, sprite, x, y); 158 draw_trans_sprite(buffer, sprite, x, y); 159 draw_trans_sprite(buffer, sprite, x, y); 160 draw_trans_sprite(buffer, sprite, x, y); 161 draw_trans_sprite(buffer, sprite, x, y); 162 draw_trans_sprite(buffer, sprite, x, y); 163 draw_trans_sprite(buffer, sprite, x, y); 164 draw_trans_sprite(buffer, sprite, x, y); 165 draw_trans_sprite(buffer, sprite, x, y); 166 draw_trans_sprite(buffer, sprite, x, y); 167 draw_trans_sprite(buffer, sprite, x, y); 168 draw_trans_sprite(buffer, sprite, x, y); 169 draw_trans_sprite(buffer, sprite, x, y); 170 draw_trans_sprite(buffer, sprite, x, y); 171 draw_trans_sprite(buffer, sprite, x, y); 172 draw_trans_sprite(buffer, sprite, x, y); 173 draw_trans_sprite(buffer, sprite, x, y); 174 draw_trans_sprite(buffer, sprite, x, y); 175 draw_trans_sprite(buffer, sprite, x, y); 176 draw_trans_sprite(buffer, sprite, x, y); 177 draw_trans_sprite(buffer, sprite, x, y); 178 draw_trans_sprite(buffer, sprite, x, y); 179 draw_trans_sprite(buffer, sprite, x, y); 180 draw_trans_sprite(buffer, sprite, x, y); 181 draw_trans_sprite(buffer, sprite, x, y); 182 draw_trans_sprite(buffer, sprite, x, y); 183 draw_trans_sprite(buffer, sprite, x, y); 184 draw_trans_sprite(buffer, sprite, x, y); 185 draw_trans_sprite(buffer, sprite, x, y); 186 draw_trans_sprite(buffer, sprite, x, y); 187 draw_trans_sprite(buffer, sprite, x, y); 188 draw_trans_sprite(buffer, sprite, x, y); 189 draw_trans_sprite(buffer, sprite, x, y); 190 draw_trans_sprite(buffer, sprite, x, y); 191 draw_trans_sprite(buffer, sprite, x, y); 192 draw_trans_sprite(buffer, sprite, x, y); 193 draw_trans_sprite(buffer, sprite, x, y); 194 draw_trans_sprite(buffer, sprite, x, y); 195 draw_trans_sprite(buffer, sprite, x, y); 196 draw_trans_sprite(buffer, sprite, x, y); 197 draw_trans_sprite(buffer, sprite, x, y); 198 draw_trans_sprite(buffer, sprite, x, y); 199 draw_trans_sprite(buffer, sprite, x, y); 200 draw_trans_sprite(buffer, sprite, x, y); 201 draw_trans_sprite(buffer, sprite, x, y); 202 draw_trans_sprite(buffer, sprite, x, y); 203 draw_trans_sprite(buffer, sprite, x, y); 204 draw_trans_sprite(buffer, sprite, x, y); 205 draw_trans_sprite(buffer, sprite, x, y); 206 draw_trans_sprite(buffer, sprite, x, y); 207 draw_trans_sprite(buffer, sprite, x, y); 208 draw_trans_sprite(buffer, sprite, x, y); 209 draw_trans_sprite(buffer, sprite, x, y); 210 draw_trans_sprite(buffer, sprite, x, y); 211 draw_trans_sprite(buffer, sprite, x, y); 212 draw_trans_sprite(buffer, sprite, x, y); 213 draw_trans_sprite(buffer, sprite, x, y); 214 draw_trans_sprite(buffer, sprite, x, y); 215 draw_trans_sprite(buffer, sprite, x, y); 216 draw_trans_sprite(buffer, sprite, x, y); 217 draw_trans_sprite(buffer, sprite, x, y); 218 draw_trans_sprite(buffer, sprite, x, y); 219 draw_trans_sprite(buffer, sprite, x, y); 220 draw_trans_sprite(buffer, sprite, x, y); 221 draw_trans_sprite(buffer, sprite, x, y); 222 draw_trans_sprite(buffer, sprite, x, y); 223 draw_trans_sprite(buffer, sprite, x, y); 224 draw_trans_sprite(buffer, sprite, x, y); 225 draw_trans_sprite(buffer, sprite, x, y); 226 draw_trans_sprite(buffer, sprite, x, y); 227 draw_trans_sprite(buffer, sprite, x, y); 228 draw_trans_sprite(buffer, sprite, x, y); 229 draw_trans_sprite(buffer, sprite, x, y); 230 draw_trans_sprite(buffer, sprite, x, y); 231 draw_trans_sprite(buffer, sprite, x, y); 232 draw_trans_sprite(buffer, sprite, x, y); 233 draw_trans_sprite(buffer, sprite, x, y); 234 draw_trans_sprite(buffer, sprite, x, y); 235 draw_trans_sprite(buffer, sprite, x, y); 236 draw_trans_sprite(buffer, sprite, x, y); 237 draw_trans_sprite(buffer, sprite, x, y); 238 draw_trans_sprite(buffer, sprite, x, y); 239 draw_trans_sprite(buffer, sprite, x, y); 240 draw_trans_sprite(buffer, sprite, x, y); 241 draw_trans_sprite(buffer, sprite, x, y); 242 draw_trans_sprite(buffer, sprite, x, y); 243 draw_trans_sprite(buffer, sprite, x, y); 244 draw_trans_sprite(buffer, sprite, x, y); 245 draw_trans_sprite(buffer, sprite, x, y); 246 draw_trans_sprite(buffer, sprite, x, y); 247 draw_trans_sprite(buffer, sprite, x, y); 248 draw_trans_sprite(buffer, sprite, x, y); 249 draw_trans_sprite(buffer, sprite, x, y); 250 draw_trans_sprite(buffer, sprite, x, y); 251 draw_trans_sprite(buffer, sprite, x, y); 252 draw_trans_sprite(buffer, sprite, x, y); 253 draw_trans_sprite(buffer, sprite, x, y); 254 draw_trans_sprite(buffer, sprite, x, y); 255 draw_trans_sprite(buffer, sprite, x, y); 256 draw_trans_sprite(buffer, sprite, x, y); 257 draw_trans_sprite(buffer, sprite, x, y); 258 draw_trans_sprite(buffer, sprite, x, y); 259 draw_trans_sprite(buffer, sprite, x, y); 260 draw_trans_sprite(buffer, sprite, x, y); 261 draw_trans_sprite(buffer, sprite, x, y); 262 draw_trans_sprite(buffer, sprite, x, y); 263 draw_trans_sprite(buffer, sprite, x, y); 264 draw_trans_sprite(buffer, sprite, x, y); 265 266 267 268 269 270 /* flip it across to the screen */ 271 blit(buffer, screen, 0, 0, 0, 0, SCREEN_W, SCREEN_H); 272 273 /* replace the background where we drew the sprite */ 274 blit(background, buffer, x, y, x, y, sprite->w, sprite->h); 275 } 276 277 clear_keybuf(); 278 279 destroy_bitmap(background); 280 destroy_bitmap(sprite); 281 destroy_bitmap(buffer); 282 283 return 0; 284} 285 286END_OF_MAIN()

Quote:

After all, I bet you're not compiling Allegro with different settings...

I don't understand what you meant, can you reword it please?

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Evert
Member #794
November 2000
avatar

He means, are you recompiling Allegro along with your program?
If not, then what optimisation options you specify should not influence draw_trans_sprite.

Kauhiz
Member #4,798
July 2004

I think he means, that you only compile your project with different settings, not allegro.

Edit: beaten...

---
It's Ridge Racer! RIIIIIDGE RAAAAACER!

Paul whoknows
Member #5,081
September 2004
avatar

I am only compiling my project, I am using pre-built binaries for MinGW.
here is the compile log:

Building Makefile: "C:\Documents and Settings\pablo\Escritorio\Makefile.win" Finding dependencies for file: C:\Documents and Settings\pablo\Escritorio\allegro-misc-4.2.1\examples\exalpha.c Executing make... make.exe -f "C:\Documents and Settings\pablo\Escritorio\Makefile.win" all g++.exe -c allegro-misc-4.2.1/examples/exalpha.c -o allegro-misc-4.2.1/examples/exalpha.o -I"C:/Dev-Cpp/lib/gcc/mingw32/3.4.2/include" -I"C:/Dev-Cpp/include/c++/3.4.2/backward" -I"C:/Dev-Cpp/include/c++/3.4.2/mingw32" -I"C:/Dev-Cpp/include/c++/3.4.2" -I"C:/Dev-Cpp/include" -O3 g++.exe allegro-misc-4.2.1/examples/exalpha.o -o "draw_alpha.exe" -L"C:/Dev-Cpp/lib" -mwindows ../../../Dev-Cpp/lib/liballeg.a Execution terminated Compilation successful

Am I doing something wrong? if yes, then how should I compile my project with optimizations in devcpp?

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Fladimir da Gorf
Member #1,565
October 2001
avatar

Paul, by the way, my routines are optimized for anti-aliased sprite rendering, so they perform best when the percentage of translucent pixels is low.

OpenLayer has reached a random SVN version number ;) | Online manual | Installation video!| MSVC projects now possible with cmake | Now alvailable as a Dev-C++ Devpack! (Thanks to Kotori)

Evert
Member #794
November 2000
avatar

Do you know how to use a profiler? That could help determine what's going on.

Paul whoknows
Member #5,081
September 2004
avatar

Quote:

Paul, by the way, my routines are optimized for anti-aliased sprite rendering, so they perform best when the percentage of translucent pixels is low.

Anti-aliased sprites is exactly what I am using.:)
.
591104
.
And your routines(with -O3) are faster than draw_trans_sprite(without optimization), but they are much much much more faster than draw_trans_sprite(with -O3).

Quote:

Do you know how to use a profiler? That could help determine what's going on.

Yes, I'll profile it, however I don't need to profile anything to appreciate the speed lose in draw_trans_sprite with -On.

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Fladimir da Gorf
Member #1,565
October 2001
avatar

In fact, I could implement a run-length encoding of translucent sprites which might be even faster... but it'd be some work. However, it seems that these days the memory speed plays even a bigger role than before, so the MMX logic doesn't speed things up as drastically as in my old Pentium 800Mhz...

OpenLayer has reached a random SVN version number ;) | Online manual | Installation video!| MSVC projects now possible with cmake | Now alvailable as a Dev-C++ Devpack! (Thanks to Kotori)

Dustin Dettmer
Member #3,935
October 2003
avatar

draw_trans_sprite wont change when you change your -On parameter. The problem would have to be somewhere else...

Paul whoknows
Member #5,081
September 2004
avatar

Quote:

In fact, I could implement a run-length encoding of translucent sprites which might be even faster... but it'd be some work. However, it seems that these days the memory speed plays even a bigger role than before, so the MMX logic doesn't speed things up as drastically as in my old Pentium 800Mhz...

That would be great! more speed == more sprites on the screen.
I don't know, you are the expert, but if the RLE works on the alpha channel as I think it should, then blitting a big sprite, with an alpha channel like the one I posted below, should obtain a significant gain in speed, or not?

.
{"name":"591106","src":"\/\/djungxnpq2nug.cloudfront.net\/image\/cache\/2\/6\/26ca77ddda0e128005fa9f427b2fabba.png","w":320,"h":240,"tn":"\/\/djungxnpq2nug.cloudfront.net\/image\/cache\/2\/6\/26ca77ddda0e128005fa9f427b2fabba"}591106
.

Quote:

draw_trans_sprite wont change when you change your -On parameter. The problem would have to be somewhere else...

Yes, but does someone know why -On in Devcpp makes exalpha.c run slower? without optimization it runs faster, what is wrong then?

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Milan Mimica
Member #3,877
September 2003
avatar

I see no difference with gcc-3.4.6.

Paul whoknows
Member #5,081
September 2004
avatar

Quote:

I see no difference with gcc-3.4.6.

I am using devcpp with GCC 3.4.2.
Did you try the binaries I posted in my previous post? they run at different speeds in my P4 2400Mhz, if you have a faster PC probably you would not notice the difference.

____

"The unlimited potential has been replaced by the concrete reality of what I programmed today." - Jordan Mechner.

Fladimir da Gorf
Member #1,565
October 2001
avatar

Quote:

if the RLE works on the alpha channel as I think it should, then blitting a big sprite, with an alpha channel like the one I posted below, should obtain a significant gain in speed

Yes, indeed. The more there's either transparent or solid pixels, the faster it'd be.

OpenLayer has reached a random SVN version number ;) | Online manual | Installation video!| MSVC projects now possible with cmake | Now alvailable as a Dev-C++ Devpack! (Thanks to Kotori)

Go to: