Performance of global access versus thread local

[Will Newton <will dot newton at linaro dot org>]

Hi,

I've been trying to build a benchmark to measure the performance of
accessing a thread pointer relative variable versus a global variable,
in order to examine the trade-offs in implementing stack guard and
pointer guard.

Using the attached code[1] I get the following numbers:

Core i5 (x86_64):

TLS ticks per 1000 loops: 0.0000101905 Global ticks per 1000 loops: 0.0000100481

(which is understandable, as the two code sequences are practically identical)

Cortex-A15 (arm):

TLS ticks per 1000 loops: 0.0000052731 Global ticks per 1000 loops: 0.0000064143

Does this test look valid? I would be interested to see if anyone else
gets different numbers on different platforms.

[1] gcc -shared -fPIC -O2 tlsvglobal.c -o libtlsvglobal.so
   gcc -O2 main.c -o main -L. -ltlsvglobal -lrt
   LD_LIBRARY_PATH=. ./main

-- 
Will Newton
Toolchain Working Group, Linaro

#include <stdio.h>
#include <time.h>

#define LOOPS 10000000

int main(void)
{
  struct timespec start, end;
  unsigned int i, loops = LOOPS;
  double tls_elapsed, global_elapsed;

  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &start);
  for (i = 0; i < LOOPS; i++)
    {
      tls_access();
    }
  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &end);

  tls_elapsed = (end.tv_sec - start.tv_sec) + (end.tv_nsec - start.tv_nsec) * 1e-9;

  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &start);
  for (i = 0; i < LOOPS; i++)
    {
      global_access();
    }
  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &end);

  global_elapsed = (end.tv_sec - start.tv_sec) + (end.tv_nsec - start.tv_nsec) * 1e-9;

  printf("TLS ticks per 1000 loops: %.10f Global ticks per 1000 loops: %.10f\n",
         (tls_elapsed / loops) * 1000, (global_elapsed / loops) * 1000);

}

static __thread int tlsvar __attribute__((tls_model ("initial-exec")));
//int globalvar __attribute__ ((visibility ("hidden")));
int globalvar;

int tls_access(void)
{
  return tlsvar;
}

void set_tls(int v)
{
  tlsvar = v;
}

int global_access(void)
{
  return globalvar;
}