perf: build runtime at opt-3

[AaronO]

This PR improves performance by building runtime at opt-3, it adds ~100kb to bin size whilst improving op performance since all ops are registered & wrapped in runtime this helps build more efficient closures, etc...

Benches

Before (main):
date_now:            	n = 500000, dt = 0.029s, r = 17241379/s, t = 58ns/op
perf_now:            	n = 500000, dt = 0.218s, r = 2293578/s, t = 436ns/op
url_parse:           	n = 50000, dt = 0.325s, r = 153846/s, t = 6500ns/op
read_zero:           	n = 500000, dt = 0.328s, r = 1524390/s, t = 656ns/op
write_null:          	n = 500000, dt = 0.324s, r = 1543210/s, t = 648ns/op

After (runtime opt-3):
date_now:            	n = 500000, dt = 0.031s, r = 16129032/s, t = 62ns/op
perf_now:            	n = 500000, dt = 0.174s, r = 2873563/s, t = 348ns/op
url_parse:           	n = 50000, dt = 0.287s, r = 174216/s, t = 5739ns/op
read_zero:           	n = 500000, dt = 0.275s, r = 1818182/s, t = 550ns/op
write_null:          	n = 500000, dt = 0.275s, r = 1818182/s, t = 550ns/op

[bartlomieju]

LGTM