Is there any difference between stating the "alignas" keyword in struct definition VS array of struct definition?

Say I have a struct named Particle and I would like it to be aligned by 32bits, should I use the align keyword when I specify the struct definition, like code below:

struct alignas(32) Particle {
    // Total size: 16 bytes, but alignas(32) ensures 32-byte alignment
    float x; // 4 bytes
    float y; // 4 bytes
    float z; // 4 bytes
    float w; // 4 bytes 
};

or should I specify it in the definition for the layouts where the structs reside, like this:

alignas(32) Particle particles[8]; // Array of 8 particles, 32-byte aligned

or should I do both?

>Solution :

You haven’t placed alignas in the correct place in the first example, but:

1. Aligning the first member of the struct will ensure that every instance of struct Particle is aligned according to your specification:

   struct Particle  {
       alignas(32) float x; // 4 bytes
       float y; // 4 bytes
       float z; // 4 bytes
       float w; // 4 bytes 
   };
   
   struct Particle particles[8];
   ... = &particles[0];            // aligned to 32
   ... = &particles[1];            // aligned to 32
   

2. Aligning only the instance will only ensure that the first instance in your array is aligned according to your specification.

   struct Particle  {
       float x; // 4 bytes
       float y; // 4 bytes
       float z; // 4 bytes
       float w; // 4 bytes 
   };
   
   alignas(32) struct Particle particles[8];
   ... = &particles[0];            // aligned to 32
   ... = &particles[1];            // not aligned to 32