[DOCS] Various improvements (#224)

- Added docstr for autotune, Config, heuristics
- Added docstr for atomics
- Hiding internal _builder argument used for built-in language primitives
- Re-factor docstr to use common templates between similar functions.

python/tutorials/03-matrix-multiplication.py

@@ -322,7 +322,7 @@ else:
     triton.testing.Benchmark(
         x_names=['M', 'N', 'K'],  # argument names to use as an x-axis for the plot
         x_vals=[
-            128 * i for i in range(1, 33)
+            128 * i for i in range(2, 33)
         ],  # different possible values for `x_name`
         line_arg='provider',  # argument name whose value corresponds to a different line in the plot
         # possible values for `line_arg``