[AMD][Docker] Create Dockerfile for ROCm environment setup (#355)

* [Add] Create Dockerfile for ROCm environment setup

* Introduced a new Dockerfile for setting up a ROCm environment with PyTorch.
* Configured the working directory and installed necessary packages including Miniconda, Python, and development tools.
* Cloned the tilelang repository and executed the ROCm installation script.
* Set environment variables for compatibility and performance optimization.

* [Remove] Delete Dockerfile for ROCm environment setup

* Removed the Dockerfile used for setting up a ROCm environment with PyTorch.
* Updated README to reflect changes in Docker image naming conventions for AMD GPU support.

docker/Dockerfile.rocm

+FROM rocm/pytorch:rocm6.3.2_ubuntu22.04_py3.10_pytorch_release_2.4.0
+
+WORKDIR /root
+
+RUN echo "LC_ALL=en_US.UTF-8" >> /etc/environment
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  build-essential git wget \
+  libgtest-dev libprotobuf-dev protobuf-compiler libgflags-dev libsqlite3-dev llvm-dev \
+  && apt-get clean autoclean && rm -rf /var/lib/apt/lists/{cache,log} /tmp/* /var/tmp/*
+
+RUN wget https://repo.anaconda.com/miniconda/Miniconda3-py310_23.5.2-0-Linux-x86_64.sh -O install_miniconda.sh && \
+  bash install_miniconda.sh -b -p /opt/conda && rm install_miniconda.sh
+
+ENV PATH="/opt/conda/bin:${PATH}"
+
+ENV LIBGL_ALWAYS_INDIRECT=1
+
+RUN conda install pip cmake && conda install -c conda-forge libstdcxx-ng=12 && conda clean --all
+
+RUN apt-get install -y python3 python3-dev python3-setuptools gcc libtinfo-dev zlib1g-dev build-essential cmake libedit-dev libxml2-dev
+
+RUN git clone https://github.com/tile-ai/tilelang.git --recursive -b main tilelang \
+  && cd tilelang && ./install_rocm.sh
+
+CMD bash

docker/README.md

@@ -5,7 +5,8 @@ git clone --recursive https://github.com/tile-ai/tilelang TileLang
 cd TileLang/docker
 # build the image, this may take a while (around 10+ minutes on our test machine)
 # replace the version number cu124 with the one you want to use
-docker build -t tilelang_cuda -f Dockerfile.cu124 .
+# replace .cu** with .rocm for AMD GPU
+docker build -t tilelang_workspace -f Dockerfile.cu124 .
 # run the container
-docker run -it --cap-add=SYS_ADMIN --network=host --gpus all --cap-add=SYS_PTRACE --shm-size=4G --security-opt seccomp=unconfined --security-opt apparmor=unconfined --name tilelang_test tilelang_cuda bash
+docker run -it --cap-add=SYS_ADMIN --network=host --gpus all --cap-add=SYS_PTRACE --shm-size=4G --security-opt seccomp=unconfined --security-opt apparmor=unconfined --name tilelang_test tilelang_workspace bash
 ```