Cao E

[Inductor][CPP] Select tiling factor for lower precision data types

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #133830 * #126502 * #126501 * #126500 cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @yf225 @chenyang78 @kadeng...

Add oneDNN support for Half LSTM on CPU

6

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #132607 cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @gujinghui @PenghuiCheng @jianyuh @min-jean-cho @yanbing-j @Guobing-Chen @Xia-Weiwen @snadampal

Use float data type for Half var_sum in batchnorm stats updating on CPU

5

Using float data type for Half `var_sum` in batchnorm stats updating on CPU to avoid `var_sum` overflow since the representation range of Half is small.

[Inductor][CPP] Leverage full bits for bf16/fp16 vectorization

2

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #133830 * __->__ #126502 * #126501 * #126500 cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx...

Add specializations for VecMaskLoad and VecMaskCast

2

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #133830 * #126502 * __->__ #126501 * #126500 cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx...