[Ffmpeg-devel] yet another silly int vs. float benchmark

[matthieu castet]

Michael Niedermayer wrote:
> Hi
> 
> heres another benchmark proggy, advantages over the others
> 1. pure c
> 2. ~40 lines of code, can be easily done in less i know ...
> 3. tries to test both the case where each instruction depends upon the 
> previous one and where the instructions are a little more independant
> 
spar64 TI UltraSparc I   (SpitFire) / UltraSparc I integrated FPU

I have to redefine get_cycle as the code is in the kernel 
(/usr/src/linux-2.4.27/arch/sparc64/kernel/time.c)

#define TICK_PRIV_BIT   (1UL << 63)
static inline long get_cycles() { long ret; \
                      __asm__ __volatile__("rd   %%tick, %0\n\t" \
                              "mov       %0, %0" \
                      : "=r" (ret)); \
                      return ret & ~TICK_PRIV_BIT; }

I hope it is right.


100 ; needed     7 cycles ->     7 cycles per operation
100 iv[0]+=iv[1];iv[1]+=iv[0]; needed   209 cycles ->   104 cycles per 
operation
100 iv[0]*=iv[1];iv[1]*=iv[0]; needed  1011 cycles ->   505 cycles per 
operation
100 fv[0]+=fv[1];fv[1]+=fv[0]; needed   607 cycles ->   303 cycles per 
operation
100 fv[0]*=fv[1];fv[1]*=fv[0]; needed   607 cycles ->   303 cycles per 
operation
100 iv[0]+=iv[1];iv[1]+=iv[2];iv[2]+=iv[3];iv[3]+=iv[4];iv[4]+=iv[5]; 
needed   264 cycles ->    52 cycles per operation
100 iv[0]*=iv[1];iv[1]*=iv[2];iv[2]*=iv[3];iv[3]*=iv[4];iv[4]*=iv[5]; 
needed  2517 cycles ->   503 cycles per operation
100 fv[0]+=fv[1];fv[1]+=fv[2];fv[2]+=fv[3];fv[3]+=fv[4];fv[4]+=fv[5]; 
needed   512 cycles ->   102 cycles per operation
100 fv[0]*=fv[1];fv[1]*=fv[2];fv[2]*=fv[3];fv[3]*=fv[4];fv[4]*=fv[5]; 
needed   512 cycles ->   102 cycles per operation