Revise performance guidelines and backend references

Updated performance guidelines and backend list in instructions.

Author: ricardoV94

.github/copilot-instructions.md

@@ -21,6 +21,12 @@ Emulate NumPy user-facing API as much as possible.
 
 **Uses pre-commit with ruff**
 
+**Performance** 
+* Could should be performant
+* Avoid expensive work in hot loops
+* Avoid redundant checks. Let errors raise naturally
+* In contrast, silent errors should be prevented
+
 **Comments**: Should be used sparingly, only for complex logic
 
 **Testing**: Should be succinct
@@ -44,7 +50,7 @@ Emulate NumPy user-facing API as much as possible.
 - `gradient.py`: Auto-differentiation
 - `compile/`: Function compilation
 - `graph/`: IR and optimization (`graph/rewriting/`)
-- `link/`: Backends (`c/`, `jax/`, `numba/`, `pytorch/`)
+- `link/`: Backends (`c/`, `jax/`, `numba/`, `mlx/`, `pytorch/`)
 - `tensor/`: Tensor ops (largest module, subdirs: `random/`, `rewriting/`, `conv/`)
 - `scalar/`: Scalar ops
 - `scan/`: Loop operations (`scan_perform.pyx` Cython)
@@ -74,7 +80,7 @@ micromamba run -n pytensor-test python -m pytest tests/ --runslow    # Include s
 
 Tests are run with `config.floatX == "float32"` and `config.mode = "FAST_COMPILE"`. If needed:
 - Cast numerical values `test_value.astype(symbolic_var.type.dtype)` 
-- Use custom function mode `get_default_mode().excluding("fusion")` or skip tests in `FAST_COMPILE`
+- Use custom function mode `get_default_mode().excluding("fusion")` or skip tests in `FAST_COMPILE` if they are not directly relevant to the mode.
 
 Alternative backends (JAX, NUMBA, ...) are optional. Use `pytest.importorskip` to fail gracefully.