What's wrong with "cudaMemset( d_arr, 10, 10 * sizeof( int ) ); " in CUDA code?

I have a toy code below:

#include <stdio.h>
#include <stdlib.h>

__global__ void add1InGPU( int *devArr )
{
    int i = threadIdx.x;
    devArr[i] += 1;
}

int main( void )
{
    int *h_arr = (int*)malloc( 10 * sizeof( int ) );
    
    int *d_arr = NULL;
    cudaMalloc( &d_arr, 10 * sizeof( int ) );
    cudaMemset( d_arr, 10, 10 * sizeof( int ) ); 

    add1InGPU <<<1, 10>>> ( d_arr );
    cudaDeviceSynchronize();
    cudaMemcpy(h_arr, d_arr, 10 * sizeof(int), cudaMemcpyDeviceToHost);
    cudaFree(d_arr);

    for (int i = 0; i < 10; i++) {
        printf("%d\n", h_arr[i]);
    }
    free( h_arr );
    return 0;
}

Why running it I get 168430091 in each index of the array? Seems to me a corrupted number.

However, if I change cudaMemset( d_arr, 10, 10 * sizeof( int ) ); to cudaMemset( d_arr, 0, 10 * sizeof( int ) );, I get the correct result, which is all 1 in the array.

Kindly teach where I got wrong, thanks in advance.

Alright I got why… cudaMemset, like a plain memset, only set each byte to the target value, while int is, in most machines, 4 bytes. Hence the issue.