Re: [v3][PATCH] Framework for performance benchmarking of functions

[Richard Henderson <rth at twiddle dot net>]

On 2013-02-20 02:17, Siddhesh Poyarekar wrote:
> +  for (i = 0; i < ITER; i++)
> +    {
> +      for (j = 0; j < NUM_SAMPLES; j++)
> +	{
> +	  clock_gettime (CLOCK_MONOTONIC_RAW, &start);
> +	  BENCH_FUNC(j);
> +	  clock_gettime (CLOCK_MONOTONIC_RAW, &end);
> +
> +	  int64_t cur = (end.tv_nsec - start.tv_nsec
> +			 + (end.tv_sec - start.tv_sec) * 1000000000);

The 1e9 constant needs to be cast to int64_t, lest this expression 
simply overflow on 32-bit hosts.

There are quite a few hosts for which the resolution of this clock isn't 
good enough to make measuring in the inside loop work.  Can we have a 
look at what clock_getres returns and perhaps measure the outside loop? 
 At least then we'll get *some* sort of answer...

E.g. ARM Cortex A9 can only measure at 1kHz, but Cortex A15-mp can 
measure at the cpu frequency (1.6GHz, reported by getres as 1ns).


r~