mirror of
https://github.com/kvcache-ai/ktransformers.git
synced 2026-04-20 14:29:22 +00:00
57d14d22bc
* refactor: move legacy code to archive/ directory - Moved ktransformers, csrc, third_party, merge_tensors to archive/ - Moved build scripts and configurations to archive/ - Kept kt-kernel, KT-SFT, doc, and README files in root - Preserved complete git history for all moved files * refactor: restructure repository to focus on kt-kernel and KT-SFT modules * fix README * fix README * fix README * fix README * docs: add performance benchmarks to kt-kernel section Add comprehensive performance data for kt-kernel to match KT-SFT's presentation: - AMX kernel optimization: 21.3 TFLOPS (3.9× faster than PyTorch) - Prefill phase: up to 20× speedup vs baseline - Decode phase: up to 4× speedup - NUMA optimization: up to 63% throughput improvement - Multi-GPU (8×L20): 227.85 tokens/s total throughput with DeepSeek-R1 FP8 Source: https://lmsys.org/blog/2025-10-22-KTransformers/ This provides users with concrete performance metrics for both core modules, making it easier to understand the capabilities of each component. * refactor: improve kt-kernel performance data with specific hardware and models Replace generic performance descriptions with concrete benchmarks: - Specify exact hardware: 8×L20 GPU + Xeon Gold 6454S, Single/Dual-socket Xeon + AMX - Include specific models: DeepSeek-R1-0528 (FP8), DeepSeek-V3 (671B) - Show detailed metrics: total throughput, output throughput, concurrency details - Match KT-SFT presentation style for consistency This provides users with actionable performance data they can use to evaluate hardware requirements and expected performance for their use cases. * fix README * docs: clean up performance table and improve formatting * add pic for README * refactor: simplify .gitmodules and backup legacy submodules - Remove 7 legacy submodules from root .gitmodules (archive/third_party/*) - Keep only 2 active submodules for kt-kernel (llama.cpp, pybind11) - Backup complete .gitmodules to archive/.gitmodules - Add documentation in archive/README.md for researchers who need legacy submodules This reduces initial clone size by ~500MB and avoids downloading unused dependencies. * refactor: move doc/ back to root directory Keep documentation in root for easier access and maintenance. * refactor: consolidate all images to doc/assets/ - Move kt-kernel/assets/heterogeneous_computing.png to doc/assets/ - Remove KT-SFT/assets/ (images already in doc/assets/) - Update KT-SFT/README.md image references to ../doc/assets/ - Eliminates ~7.9MB image duplication - Centralizes all documentation assets in one location * fix pic path for README
69 lines
2.0 KiB
Docker
69 lines
2.0 KiB
Docker
# Base image
|
|
FROM intel/oneapi-basekit:2025.0.1-0-devel-ubuntu22.04
|
|
|
|
ARG http_proxy
|
|
ARG https_proxy
|
|
|
|
ENV DEBIAN_FRONTEND=noninteractive
|
|
ENV CONDA_DIR=/opt/conda
|
|
|
|
# Install dependencies
|
|
RUN apt-get update && apt-get install -y \
|
|
wget \
|
|
curl \
|
|
bash \
|
|
git \
|
|
vim \
|
|
ca-certificates \
|
|
binutils \
|
|
cmake \
|
|
g++ \
|
|
&& rm -rf /var/lib/apt/lists/*
|
|
|
|
# Install Miniforge
|
|
RUN wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh -O /tmp/miniforge.sh && \
|
|
bash /tmp/miniforge.sh -b -p $CONDA_DIR && \
|
|
rm /tmp/miniforge.sh && \
|
|
$CONDA_DIR/bin/conda clean -afy
|
|
|
|
# Add conda to PATH
|
|
ENV PATH=$CONDA_DIR/bin:$PATH
|
|
|
|
RUN bash -c "\
|
|
source /opt/conda/etc/profile.d/conda.sh && \
|
|
conda create --name ktransformers python=3.11 -y && \
|
|
conda activate ktransformers && \
|
|
conda env list && \
|
|
conda install -c conda-forge libstdcxx-ng -y && \
|
|
strings \$(find /opt/conda/envs/ktransformers/lib -name 'libstdc++.so.6') | grep GLIBCXX | grep 3.4.32 \
|
|
"
|
|
|
|
RUN bash -c "\
|
|
source /opt/conda/etc/profile.d/conda.sh && \
|
|
conda activate ktransformers && \
|
|
pip install ipex-llm[xpu_2.6]==2.3.0b20250518 --extra-index-url https://download.pytorch.org/whl/xpu && \
|
|
pip uninstall -y torch torchvision torchaudio && \
|
|
pip install torch==2.7+xpu torchvision torchaudio --index-url https://download.pytorch.org/whl/test/xpu && \
|
|
pip uninstall -y intel-opencl-rt dpcpp-cpp-rt && \
|
|
pip list \
|
|
"
|
|
|
|
# Clone and set up ktransformers repo
|
|
RUN bash -c "\
|
|
source $CONDA_DIR/etc/profile.d/conda.sh && \
|
|
conda activate ktransformers && \
|
|
git clone https://github.com/kvcache-ai/ktransformers.git && \
|
|
cd ktransformers && \
|
|
git submodule update --init && \
|
|
sed -i 's/torch\.xpu\.is_available()/True/g' setup.py && \
|
|
bash install.sh --dev xpu \
|
|
"
|
|
|
|
# Init conda and prepare bashrc
|
|
RUN conda init bash && \
|
|
echo "source $CONDA_DIR/etc/profile.d/conda.sh" >> ~/.bashrc && \
|
|
echo "conda activate ktransformers" >> ~/.bashrc
|
|
|
|
WORKDIR /ktransformers/
|
|
CMD ["bash"]
|