操作系统版本：ubuntu20.04
DTK版本：DTK24.04.1
驱动版本：rock-5.7.1-6.2.18 或者提供如下指令结果rpm -qa | grep rock

ubuntu@ubuntu:~$ dpkg --list|grep rock
ii  rock-5.7.1                                     6.2.18                                         amd64        hydcu built using CMake

DCU类型：Z100L

Name:                    ZIFANG                       
  Uuid:                    GPU-6f096431180e5141         
  Marketing Name:          C878186                      
  Vendor Name:             HYGON                        
  Feature:                 KERNEL_DISPATCH              
  Profile:                 BASE_PROFILE                 
  Float Round Mode:        NEAR                         
  Max Queue Number:        128(0x80)                    
  Queue Min Size:          64(0x40)                     
  Queue Max Size:          131072(0x20000)              
  Queue Type:              MULTI                        
  Node:                    4                            
  Device Type:             DCU                          
  Cache Info:        
    L1:                      16(0x10) KB                  
    L2:                      8192(0x2000) KB              
  Chip ID:                 21431(0x53b7)                
  Cacheline Size:          64(0x40)                     
  Max Clock Freq. (MHz):   700                          
  BDFID:                   1536                         
  Internal Node ID:        4                            
  Compute Unit:            60                           
  SIMDs per CU:            4                            
  Shader Engines:          4                            
  Shader Arrs. per Eng.:   1                            

      Allocatable:             TRUE                         
      Alloc Granule:           4KB                          
      Alloc Alignment:         4KB                          
      Accessible by all:       FALSE                        
    Pool 3             
      Segment:                 GLOBAL; FLAGS: FINE GRAINED  
      Size:                    33538048(0x1ffc000) KB       
      Allocatable:             TRUE                         
      Alloc Granule:           4KB                          
      Alloc Alignment:         4KB                          
      Accessible by all:       FALSE                        
    Pool 4             
      Segment:                 GROUP                        
      Size:                    64(0x40) KB                  
      Allocatable:             FALSE                        
      Alloc Granule:           0KB                          
      Alloc Alignment:         0KB                          
      Accessible by all:       FALSE                        
  ISA Info:          
    ISA 1              
      Name:                    amdgcn-amd-amdhsa--gfx906:sramecc+:xnack-
      Machine Models:          HSA_MACHINE_MODEL_LARGE      
      Profiles:                HSA_PROFILE_BASE             
      Default Rounding Mode:   NEAR                         
      Default Rounding Mode:   NEAR                         
      Fast f16:                TRUE                         
      Workgroup Max Size:      1024(0x400)                  
      Workgroup Max Size per Dimension:
        x                        1024(0x400)                  
        y                        1024(0x400)                  
        z                        1024(0x400)                  
      Grid Max Size:           4294967295(0xffffffff)       
      Grid Max Size per Dimension:
        x                        4294967295(0xffffffff)       
        y                        4294967295(0xffffffff)       
        z                        4294967295(0xffffffff)       
      FBarrier Max Size:       32           

编译过程：

• 运行容器

docker run -it --device=/dev/kfd --device=/dev/dri --device=/dev/mkfd --shm-size=1G -v ./:/data --rm 
image.sourcefind.cn:5000/dcu/admin/base/pytorch:2.1.0-ubuntu20.04-dtk24.04.1-py3.8 bash

• 执行命令

apt-get update && apt upgrade -y && \
	DEBIAN_FRONTEND=noninteractive apt install --no-install-recommends -y \
        gnupg \
        ca-certificates \
        wget  \
        git \
        curl \
        cmake \
        python3.8-dev \
        build-essential \
        software-properties-common && \
    wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add - &&\
    echo "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-17 main" | sudo tee /etc/apt/sources.list.d/llvm-17.list &&\
    echo -e '[global]\nindex-url = https://pypi.tuna.tsinghua.edu.cn/simple\nbreak-system-packages = true' > /etc/pip.conf  &&\
    apt update &&\
    apt install -y clang-17 &&\
    update-alternatives --install /usr/bin/clang clang /usr/bin/clang-17 100 &&\
    update-alternatives --install /usr/bin/clang++ clang++ /usr/bin/clang++-17 100  &&\
    echo -e '[global]\nindex-url = https://pypi.tuna.tsinghua.edu.cn/simple\nbreak-system-packages = true' > /etc/pip.conf  &&\
    git clone -b 2.1.0-release http://developer.hpccube.com/codes/aicomponent/torch-scatter.git &&\
    cd torch-scatter &&\
    git submodule sync &&\
    git submodule update --init --recursive &&\
    pip install numpy  &&\
    pip install 'urllib3==1.26.14'  &&\
    pip install wheel  &&\
    source /opt/dtk/env.sh &&\
    source /opt/dtk/cuda/env.sh &&\
    source /opt/dtk-24.04.1/.hyhal/hydm/hydmi_env.sh &&\
    export PATH=/usr/bin:$PATH &&\
    python setup.py bdist_wheel

部分报错日志：

错误1
In file included from /opt/dtk/cuda/include/driver_types.h:14:
/opt/dtk/cuda/include/vector_types.h:119:30: error: definition of type 'ushort4' conflicts with type alias of the same name
__cuda_builtin_vector_align8(ushort4, unsigned short x; unsigned short y; unsigned short z; unsigned short w;);
                             ^
/opt/dtk/hip/include/hip/amd_detail/amd_hip_vector_types.h:1137:1: note: 'ushort4' declared here
__MAKE_VECTOR_TYPE__(ushort, unsigned short);
^
/opt/dtk/hip/include/hip/amd_detail/amd_hip_vector_types.h:1112:15: note: expanded from macro '__MAKE_VECTOR_TYPE__'
        using CUDA_name##4 = HIP_vector_type<T, 4>;
              ^
<scratch space>:59:1: note: expanded from here
ushort4

错误2
In file included from /opt/dtk/cuda/include/cuda_runtime.h:36:
In file included from /opt/dtk/cuda/include/builtin_types.h:8:
In file included from /opt/dtk/cuda/include/driver_types.h:14:
/opt/dtk/cuda/include/vector_types.h:113:27: error: definition of type 'ushort3' conflicts with type alias of the same name
struct __device_builtin__ ushort3
错误3
In file included from /opt/dtk/cuda/include/thrust/iterator/iterator_traits.h:
__MAKE_VECTOR_TYPE__(short, short);
^
/opt/dtk/hip/include/hip/amd_detail/amd_hip_vector_types.h:1111:15: note: expanded from macro '__MAKE_VECTOR_TYPE__'
        using CUDA_name##3 = HIP_vector_type<T, 3>;\
              ^
<scratch space>:62:1: note: expanded from here
short3
^
In file included from csrc/hip/segment_coo_cuda.hip:4:
还有其余基本都是类似错误