The rationale behind opt-level = 's' is that it's a good compromise between code size and code performance. On our chips, there is even a relationship between the two, as our caches are quite small so the smaller the code the more we can fit in the cache for better performance. Users can of course update this for their specific needs.
The rationale behind
opt-level = 's'
is that it's a good compromise between code size and code performance. On our chips, there is even a relationship between the two, as our caches are quite small so the smaller the code the more we can fit in the cache for better performance. Users can of course update this for their specific needs.