[Inductor][quant]Enable decomposed.quant/dequant lowering and vec code gen (#99131)
**Summary**
Since current quantization flow has not decomposed quant/dequant into prim ops, in this PR
- We enable the quant/dequant decomposition as lowering inside inductor.
- For the `decomposed.quant/dequant.tensor` overload, there are loading of scalar tensor of `zero point` and `scale`, we need to enable the vec code gen for these op overloads.
- Minor change as adding `is_load_uint8_as_float` and `is_store_float_as_uint8` default value `False` into `OptimizationContext`.
**TestPlan**
```
cd test/inductor && python -m pytest test_cpu_repro.py -k test_dequant_quant_lowering
```
co-author with @Xia-Weiwen
Pull Request resolved: https://github.com/pytorch/pytorch/pull/99131
Approved by: https://github.com/jgong5, https://github.com/EikanWang, https://github.com/jansel