support shape broadcast in UOp.alu

[chenyuxyz]

i think it can integrate tighter, but now Tensor also does ufix from UOp and implicit dtype upcast

[github-actions]

Changes

Name                          Lines    Diff    Tokens/Line    Diff
--------------------------  -------  ------  -------------  ------
tinygrad/tensor.py             1443      -2           21.2    -0.0
tinygrad/uop/ops.py            1332      +9           18.3    -0.0
tinygrad/mixin/movement.py      152      -3           19.6    -0.0


total lines changes: +4