Problems regarding to performance analyzer

Question

I'm trying to ultilize sun performance analyzer to analyze c codes implementing OpenMP API. I have used a virtual machine which is classified as Linux-32-bit, 2 CPUs, Debian_4.0 system. And I have met two problems:
1. For a large simulator code sets, I could not set more than 2 threads in the code. And from the results of analyzer, it seems there were two threads, but only one cpu is binded.What's the reason for that, do i need to bind cpus to threads?
2. For a simple OpenMP program as follows, after using collect command, in the result nothing is shown in any of the tabs except experiments tab. It does give some warnings, but i don't think that's the key problem.Because the analyzer works for our simulator.(btw. this is the example code of sun thread analyzer.For that one, i also got same kind of problems for simple testing programs i wrote.)
commands used:cc -xopenmp -xO3 openmptest.c
collect ./a.out | sort
analyzer &
Results got: Creating experiment database test.9.er ...
Number of prime numbers between 2 and 3000 430
WARNING: Linux-32-bit, 2 CPUs, Debian_4.0 system is not supported by the Performance tools.
#include <stdio.h>
#include <math.h>
#include <omp.h>
#define THREADS 2
#define N 3000
int primes[N];
int pflag[N];
int is_prime(int v)
     int i;
     int bound = floor(sqrt((double)v) )+1;
     for(i=2;i< bound;++i)
          if(!pflag)
          continue;
          if(v%i==0)
               pflag[v] = 0;
               return 0;
     return(v>1);
int main()
     int i;
     int total = 0;
#ifdef _OPENMP
     omp_set_num_threads(THREADS);
     omp_set_dynamic(0);
#endif
     for(i=0;i<N;++i)
          pflag[i]=1;
     #pragma omp parallel for
     for(i=2;i<N;i++)
          if(is_prime(i)){
               #pragma omp critical
                    primes[total]=i;
                    total++;
     printf("Number of prime numbers between 2 and %d %d\n",N,total);
     for(i=0;i<total;i++)
     printf("%d\n",primes[i]);
     return 0;

Answer 1

I suspect both issues are related to the support of multiple processors in the virtual machine and the guest OS. Would you be able to try it outside a virtual machine? Or provide more details about your setup?
Thanks.
-- Yuan