Re: [AD] SF.net SVN: alleg:[11840] allegro/branches/4.9

[Milan Mimica <milan.mimica@xxxxxxxxxx>]

Elias Pschernig wrote:
> How much do you get when using (int)x? (Edit it in
> include/allegro/internal/ainterh_float.h). I would expect int-conversion
> to be slower with anything using the 387 floating point instructions,
> but faster with MMX instructions.

That's what I tested, but I didn't link statically and there are many 
liballegros laying around.
Statically linked this time. I've got some interesting results:

gcc-4.3.2 -march=pentium-m
  unmodified SVN:
  Time = 1.11218 s, FPS = 269.74
  #define _al_fast_float_to_int(f)  ((int)f):
  Time = 1.06996 s, FPS = 280.384

gcc-4.3.2 -march=i586
  unmodifed SVN:
  Time = 1.23826 s, FPS = 242.275
  #define _al_fast_float_to_int(f)  ((int)f):
  Time = 3.23983 s, FPS = 92.5973

gcc-4.3.2 (no extra flags, still have to figure what that means)
  unmodifed SVN:
  Time = 1.10411 s, FPS = 271.711
  #define _al_fast_float_to_int(f)  ((int)f):
  Time = 1.05332 s, FPS = 284.813


$ cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 13
model name      : Intel(R) Pentium(R) M processor 1.73GHz
stepping        : 8
cpu MHz         : 800.000
cache size      : 2048 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov 
pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe nx bts est tm2
bogomips        : 1597.32
clflush size    : 64



--
Milan Mimica
http://sparklet.sf.net