Metal: Activate bfloat affine and add benchmark (#1543)

* Use cfg to seperate benchmark results based on features

* Add bfloat affine and benchmarks

* Fix flops calculation

* Remove allow pragma

* Avoid some unnecessary returns.

* Improve benchmarks layout

---------

Co-authored-by: Laurent <laurent.mazare@gmail.com>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

candle-core/benches/benchmarks/mod.rs

@@ -1,3 +1,4 @@
+pub(crate) mod affine;
 pub(crate) mod matmul;
 pub(crate) mod where_cond;