I have been trying to make my hash template faster for a few days now but I can't seem to find any more optimizations. I noticed that my single hash runs slower than some people's double hash. Can someone tell me where I'm going wrong?

typedef long long ll;

int powhash1[ 1000000+ 10]={};
int powhash2[ 1000000+ 10]={};
int prefhash1[1000000 + 10];
int prefhash2[1000000 + 10];

int add(ll x,ll y,ll mod)
{
    return (x+y>=mod)?(x+y-mod):(x+y);
}
int subtract(ll x,ll y,ll mod)
{
    return (x-y<0)?(x-y+mod):(x-y);
}

int multp(ll x,ll y,ll mod)
{
    return (x*y)%mod;
}


ll BASE1 = 125;
ll MOD1 = 1e9 + 9;

ll BASE2 = 250;
ll MOD2 = 1e9 + 7;

void prefhashcalc(string s,ll base,ll mod,int*prefhash)
{
    ll sum = 0;
    int ns = s.size();

    for(int i=0;i<ns;i++)
    {

        sum = add((sum*base)%mod,s[i]-'A'+1,mod);
        prefhash[i]=sum;
    }
}


ll strhash(string s,ll base,ll mod)
{
    ll sum = 0;
    int ns = s.size();

    for(int i=0;i<ns;i++)
    {
        sum = add((sum*base)%mod,s[i]-'A'+1,mod);
    }
    return sum;

}

void powhashfill(ll base,ll mod,int*powhash)
{
    for(int i=0;i<1000000 + 10;i++)
    {
        if(i==0)
        {
            powhash[0]=1;
            continue;
        }

        powhash[i] = multp(powhash[i-1],base,mod);
    }

}

ll substrhash(int l,int r,ll mod,int*prefhash,int*powhash)
{

   ll x = subtract(  prefhash[r],  multp(prefhash[l-1],powhash[r-l+1],mod),  mod  );

    return x;
}