GitLab

* Init apex.ppu (pipeline model parallel utility)

Reference commit:

```
commit 5ab646376d67831601d5552c193241d017f1b35c (HEAD -> main, internal/main)
Merge: 14f2c684 7b293d9b
Author: Mohammad Shoeybi <mshoeybi@nvidia.com>
Date:   Wed Sep 22 22:57:54 2021 -0700

    Merge branch 'add_BOS' into 'main'

    Add Beginning of Sentence token option and adding semaphore while multi-threading to prevent crashes and hangs due to connection keep-alives

    See merge request ADLR/megatron-lm!328
```

* removing get_args and replace import - phase 1

* removing get_args and replace import - phase 2

* move ppu to apex.transformer.pipeline_parallel

* update two __init__.py

* update READMEs

* mpu -> parallel_state & tensor_parallel

* fix

* remove not pipeline files

* separate schedules.py - phase 1

* dissect schedules.py

* data_iterators -> batch

* remove optimizer from forward_backward_step funcs

* init test

* Apply 2 suggestion(s...

* switch from clone to out-of-place subtract

* Update apex/mpu/cross_entropy.py

* Apply 1 suggestion(s) to 1 file(s)
Co-authored-by: Eddie Yan <eddiey@nvidia.com>

Changes include
- THC headers removal
- TH macros replacement
- fix some typo in comment

change chunking scheme for full-allreduce case, add parameter order argument, both to enable contiguous chunking of allgather (#1190)

1. remove the weight broadcast in the constructor
2. disable unnecessary allreduces for clip-after-ar

* run backward

* remove custom_fwd/custom_bwd

* [ColumnParallelLinear] Test behavior in autocast

* fix test

* casts manually to autocast dtype

Co-authored-by: Piotr Bialecki <pbialecki@nvidia.com>
Co-authored-by: Eddie Yan <eddiey@nvidia.com>
Co-authored-by: Rishi Puri <riship@nvidia.com>
Co-authored-by: Sangkug Lym <slym@nvidia.com>

- passing include directories to `CUDAExtension`'s `include_dirs` argument
- removing `-I/path/to/dir` arguments from `extra_compile_args`

* support for fused dense layer with cublasLt, fusion in both fprop and bprop

* fix typo causing syntax error

* add fused GEMM+gelu+GEMM modue

* fix typo for workspace size

* update cublas check for 11600

* add tests for fused dense layer

* fix CUDA 10.x path

* safer guard around CUBLAS constants, remove unreferenced variable

* more guard changes

* guard against cublas version instead of cuda
Co-authored-by: Sukru Eryilmaz <seryilmaz@computelab-dgx1v-32.nvidia.com>

Optional NCCL communicator argument to init method

Bug fix in wgrad

Bug fixes

Various bug fixes in fused spatial parallel bottleneck block

* option to set param views to flat buffer

* remove redundant variables in init_stage1
Co-authored-by: Sukru Eryilmaz <seryilmaz@computelab-dgx1v-32.nvidia.com>
Co-authored-by: ptrblck <ptrblck@users.noreply.github.com>

* add full all-reduce code path

* debug

* debug
Co-authored-by: ptrblck <ptrblck@users.noreply.github.com>

Add functions to compute grad_out1, grad_out1_halo

* fuse norm into scale

* add fused norm into dlamb
Co-authored-by: Sukru Eryilmaz <seryilmaz@computelab-dgx1v-32.nvidia.com>

* support for fused dense layer with cublasLt, fusion in both fprop and bprop

* fix typo causing syntax error

* add fused GEMM+gelu+GEMM modue

* fix typo for workspace size

* update cublas check for 11600

* add tests for fused dense layer

* fix CUDA 10.x path
Co-authored-by: Sukru Eryilmaz <seryilmaz@computelab-dgx1v-32.nvidia.com>

wrapper function for flat view creation in _lazy_init_stage2

Spatially Distributed Fast Bottleneck block

Wrote a small wrapper function for flat view creation in _lazy_init_stage2 to support channels last data formats