Open DavidPeicho opened 3 years ago
Hello, i think you can achieve this by using the by_ref
Iterator method?
I'm willing to bet that small f32 and u32 arrays will be optimized into a single SSE register. I'd be pretty surprised to see any measurable performance increase by using references here.
I was wondering if it made sense to add extra implementation of the
Item
trait here, which would allow to iterate over references of items instead of copies.I think this feature may only serve few people (like me), because most users will definitely copy the data extract from accessors into their own data struct only once.