Yu, Guangye

generalize custom_fwd&custom_bwd to be device-agnostic

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #126531 * #126527 cc @mcarilli @ptrblck @leslie-fang-intel @jgong5 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang

Add XPU memory-related APIs

16

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #129919 * #130923 # Motivation According to https://github.com/pytorch/pytorch/issues/116322, we will help unify the device allocator. So we introduce a simple xpu device...

Zero check has an inconsistent behavior in different versions

21

I found a difference between sympy 1.12 and 1.13. ```python # for 1.12 >>> import sympy >>> a = sympy.Number(0.0) >>> a == 0 True ``` ```python # for 1.13...

Introduce a device-agnostic runtime API design

10

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #132371 * #137493 * #133572 * __->__ #132204 # Motivation According to [[RFC]A device-agnostic Python runtime API design for stream-based accelerators](https://github.com/pytorch/pytorch/issues/128403), this PR...

[WIP] Refactor distributed code via device-agnostic API

5

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #132371 * #137493 * #133572 * #132204 cc @XilunWu @H-Huang @awgu @kwen2501 @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @c-p-i-o

Clean up accelerator APIs

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #132371 * __->__ #137493 * #133572 * #132204 cc @XilunWu @H-Huang @awgu @kwen2501 @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @c-p-i-o

Add UTs for accelerator device-agnostic runtime APIs

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #132371 * #137493 * __->__ #133572 * #132204 cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10

[WIP] Support torch.Event elapsed_time method on XPU

4

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * #140873 * __->__ #140865 # Motivation This PR aims to support c10::Event/torch.Event elapsed_time method on XPU. We create a profiling tag Event when...

[WIP] [Reland] Enable XPUEvent elapsed_time function

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #140873 * #140865 cc @gujinghui @EikanWang @fengyuan14

[WIP] Support ExternalStream for XPU backend

2

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #141123 * #141119 * #141114