Yanming W
269150b6f4
|
2 years ago | |
---|---|---|
.. | ||
README.md | 2 years ago | |
__init__.py | ||
builder.py | ||
cpu_adam.py | ||
fused_optim.py | ||
layernorm.py | ||
moe.py | ||
multi_head_attn.py | ||
scaled_masked_softmax.py | ||
scaled_upper_triangle_masked_softmax.py | ||
utils.py | 2 years ago |
README.md
Build PyTorch Extensions
Overview
Building PyTorch extensions can be a difficult task for users not from the system background. It is definitely frustrating if the users encounter many strange technical jargons when install Colossal-AI. Therefore, we will provide two methods of building the PyTorch extensions for the users.
- Build CUDA extensions when running
pip install
ifCUDA_EXT=1
- Build the extension during runtime
The first method is more suitable for users who are familiar with CUDA environment configurations. The second method is for those who are not as they only need to build the kernel which is required by their program.
These two methods have different advantages and disadvantages. Method 1 is good because it allows the user to build all kernels during installation and directly import the kernel. They don't need to care about kernel building when running their program. However, installation may fail if they don't know how to configure their environments and this leads to much frustration. Method 2 is good because it allows the user to only build the kernel they actually need, such that there is a lower probability that they encounter environment issue. However, it may slow down their program due to the first build and subsequence load.
PyTorch Extensions in Colossal-AI
The project DeepSpeed has proposed a solution to support kernel-build during either installation or runtime. We have adapted from DeepSpeed's solution to build extensions. The extension build requires two main functions from PyTorch:
torch.utils.cpp_extension.CUDAExtension
: used to build extensions insetup.py
duringpip install
.torch.utils.cpp_extension.load
: used to build and load extension during runtime
Please note that the extension build by CUDAExtension
cannot be loaded by the load
function and load
will run its own build again (correct me if I am wrong).
Based on the DeepSpeed's work, we have make several modifications and improvements:
- All pre-built kernels (those installed with
setup.py
) will be found incolossalai._C
- All runtime-built kernels will be found in the default torch extension path, i.e. ~/.cache/colossalai/torch_extensions. (If we put the built kernels in the installed site-package directory, this will make pip uninstall incomplete)
- Once a kernel is loaded, we will cache it in the builder to avoid repeated kernel loading.
When loading the built kernel, we will first check if the pre-built one exists. If not, the runtime build will be triggered.