ICODE-ADC/lammps
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #2692 from arghdos/master

Browse Source 
Porting to new default platform for AMD/HIP in ROCm 4.1
This commit is contained in:
Axel Kohlmeyer
2021-04-06 14:39:30 -04:00
committed by GitHub
parent 0b4d3588b5 160f2cc630
commit 00e4bdb2b7
7 changed files with 33 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
cmake/Modules/Packages/GPU.cmake
View File

@ -218,7 +218,7 @@ elseif(GPU_API STREQUAL "HIP")
if(NOT DEFINED HIP_PLATFORM)
if(NOT DEFINED ENV{HIP_PLATFORM})
set(HIP_PLATFORM "hcc" CACHE PATH "HIP Platform to be used during compilation")
set(HIP_PLATFORM "amd" CACHE PATH "HIP Platform to be used during compilation")
else()
set(HIP_PLATFORM $ENV{HIP_PLATFORM} CACHE PATH "HIP Platform used during compilation")
endif()
@ -226,7 +226,7 @@ elseif(GPU_API STREQUAL "HIP")
set(ENV{HIP_PLATFORM} ${HIP_PLATFORM})
if(HIP_PLATFORM STREQUAL "hcc")
if(HIP_PLATFORM STREQUAL "hcc" OR HIP_PLATFORM STREQUAL "amd")
set(HIP_ARCH "gfx906" CACHE STRING "HIP target architecture")
elseif(HIP_PLATFORM STREQUAL "nvcc")
find_package(CUDA REQUIRED)
@ -284,7 +284,7 @@ elseif(GPU_API STREQUAL "HIP")
set(CUBIN_FILE "${LAMMPS_LIB_BINARY_DIR}/gpu/${CU_NAME}.cubin")
set(CUBIN_H_FILE "${LAMMPS_LIB_BINARY_DIR}/gpu/${CU_NAME}_cubin.h")
if(HIP_PLATFORM STREQUAL "hcc")
if(HIP_PLATFORM STREQUAL "hcc" OR HIP_PLATFORM STREQUAL "amd")
configure_file(${CU_FILE} ${CU_CPP_FILE} COPYONLY)
if(HIP_COMPILER STREQUAL "clang")
@ -386,6 +386,12 @@ elseif(GPU_API STREQUAL "HIP")
target_compile_definitions(hip_get_devices PRIVATE -D__HIP_PLATFORM_HCC__)
target_include_directories(hip_get_devices PRIVATE ${HIP_ROOT_DIR}/../include)
elseif(HIP_PLATFORM STREQUAL "amd")
target_compile_definitions(gpu PRIVATE -D__HIP_PLATFORM_AMD__)
target_include_directories(gpu PRIVATE ${HIP_ROOT_DIR}/../include)
target_compile_definitions(hip_get_devices PRIVATE -D__HIP_PLATFORM_AMD__)
target_include_directories(hip_get_devices PRIVATE ${HIP_ROOT_DIR}/../include)
endif()
target_link_libraries(lammps PRIVATE gpu)

13
doc/src/Build_extras.rst
View File

@ -125,7 +125,7 @@ CMake build
# default is sm_50
-D HIP_ARCH=value # primary GPU hardware choice for GPU_API=hip
# value depends on selected HIP_PLATFORM
# default is 'gfx906' for HIP_PLATFORM=hcc and 'sm_50' for HIP_PLATFORM=nvcc
# default is 'gfx906' for HIP_PLATFORM=amd and 'sm_50' for HIP_PLATFORM=nvcc
-D HIP_USE_DEVICE_SORT=value # enables GPU sorting
# value = yes (default) or no
-D CUDPP_OPT=value # use GPU binning on with CUDA (should be off for modern GPUs)
@ -169,17 +169,24 @@ desired, you can set :code:`USE_STATIC_OPENCL_LOADER` to :code:`no`.
If you are compiling with HIP, note that before running CMake you will have to
set appropriate environment variables. Some variables such as
:code:`HCC_AMDGPU_TARGET` or :code:`CUDA_PATH` are necessary for :code:`hipcc`
:code:`HCC_AMDGPU_TARGET` (for ROCm <= 4.0) or :code:`CUDA_PATH` are necessary for :code:`hipcc`
and the linker to work correctly.
.. code:: bash
# AMDGPU target
# AMDGPU target (ROCm <= 4.0)
export HIP_PLATFORM=hcc
export HCC_AMDGPU_TARGET=gfx906
cmake -D PKG_GPU=on -D GPU_API=HIP -D HIP_ARCH=gfx906 -D CMAKE_CXX_COMPILER=hipcc ..
make -j 4
.. code:: bash
# AMDGPU target (ROCm >= 4.1)
export HIP_PLATFORM=amd
cmake -D PKG_GPU=on -D GPU_API=HIP -D HIP_ARCH=gfx906 -D CMAKE_CXX_COMPILER=hipcc ..
make -j 4
.. code:: bash
# CUDA target (not recommended, use GPU_ARCH=cuda)

6
lib/gpu/Makefile.hip
View File

@ -1,6 +1,6 @@
# /* ----------------------------------------------------------------------
# Generic Linux Makefile for HIP
# - export HIP_PLATFORM=hcc (or nvcc) before execution
# - export HIP_PLATFORM=amd (or nvcc) before execution
# - change HIP_ARCH for your GPU
# ------------------------------------------------------------------------- */
@ -42,6 +42,10 @@ ifeq (hcc,$(HIP_PLATFORM))
HIP_OPTS += -ffast-math
# possible values: gfx803,gfx900,gfx906
HIP_ARCH = gfx906
else ifeq (amd,$(HIP_PLATFORM))
HIP_OPTS += -ffast-math
# possible values: gfx803,gfx900,gfx906
HIP_ARCH = gfx906
else ifeq (nvcc,$(HIP_PLATFORM))
HIP_OPTS += --use_fast_math
HIP_ARCH = -gencode arch=compute_30,code=[sm_30,compute_30] -gencode arch=compute_32,code=[sm_32,compute_32] -gencode arch=compute_35,code=[sm_35,compute_35] \

5
lib/gpu/README
View File

@ -212,8 +212,8 @@ additionally requires cub (https://nvlabs.github.io/cub). Download and
extract the cub directory to lammps/lib/gpu/ or specify an appropriate
path in lammps/lib/gpu/Makefile.hip.
2. In Makefile.hip it is possible to specify the target platform via
export HIP_PLATFORM=hcc or HIP_PLATFORM=nvcc as well as the target
architecture (gfx803, gfx900, gfx906 etc.)
export HIP_PLATFORM=amd (ROCm >= 4.1), HIP_PLATFORM=hcc (ROCm <= 4.0)
or HIP_PLATFORM=nvcc as well as the target architecture (gfx803, gfx900, gfx906 etc.)
3. If your MPI implementation does not support `mpicxx --showme` command,
it is required to specify the corresponding MPI compiler and linker flags
in lammps/lib/gpu/Makefile.hip and in lammps/src/MAKE/OPTIONS/Makefile.hip.
@ -278,4 +278,3 @@ and
Brown, W.M., Masako, Y. Implementing Molecular Dynamics on Hybrid High
Performance Computers - Three-Body Potentials. Computer Physics Communications.
2013. 184: p. 2785–2793.

8
lib/gpu/lal_pre_cuda_hip.h
View File

@ -30,7 +30,7 @@
// -------------------------------------------------------------------------
#ifdef __HIP_PLATFORM_HCC__
#if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)
#define CONFIG_ID 303
#define SIMD_SIZE 64
#else
@ -161,7 +161,7 @@
// KERNEL MACROS - TEXTURES
// -------------------------------------------------------------------------
#ifdef __HIP_PLATFORM_HCC__
#if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)
#define _texture(name, type) __device__ type* name
#define _texture_2d(name, type) __device__ type* name
#else
@ -201,7 +201,7 @@
#define mu_tex mu_
#endif
#ifdef __HIP_PLATFORM_HCC__
#if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)
#undef fetch4
#undef fetch
@ -266,7 +266,7 @@ typedef struct _double4 double4;
#endif
#endif
#if defined(CUDA_PRE_NINE) || defined(__HIP_PLATFORM_HCC__)
#if defined(CUDA_PRE_NINE) || defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)
#ifdef _SINGLE_SINGLE
#define shfl_down __shfl_down

2
tools/singularity/ubuntu18.04_amd_rocm.def
View File

@ -94,7 +94,7 @@ From: ubuntu:18.04
###########################################################################
export PATH=$PATH:/opt/rocm/bin:/opt/rocm/profiler/bin:/opt/rocm/opencl/bin/x86_64
git clone -b rocm-3.7.x https://github.com/ROCmSoftwarePlatform/hipCUB.git
git clone -b rocm-4.1.x https://github.com/ROCmSoftwarePlatform/hipCUB.git
mkdir hipCUB/build
cd hipCUB/build
CXX=hipcc cmake -D BUILD_TEST=off ..

4
tools/singularity/ubuntu20.04_amd_rocm.def
View File

@ -2,7 +2,7 @@ BootStrap: docker
From: ubuntu:20.04
%environment
export PATH=/usr/lib/ccache:/usr/local/cuda-11.0/bin:${PATH}:/opt/rocm/bin:/opt/rocm/profiler/bin:/opt/rocm/opencl/bin/x86_64
export PATH=/usr/lib/ccache:${PATH}:/opt/rocm/bin:/opt/rocm/profiler/bin:/opt/rocm/opencl/bin/x86_64
%post
export DEBIAN_FRONTEND=noninteractive
apt-get update
@ -90,7 +90,7 @@ From: ubuntu:20.04
###########################################################################
export PATH=$PATH:/opt/rocm/bin:/opt/rocm/profiler/bin:/opt/rocm/opencl/bin/x86_64
git clone -b rocm-3.7.x https://github.com/ROCmSoftwarePlatform/hipCUB.git
git clone -b rocm-4.1.x https://github.com/ROCmSoftwarePlatform/hipCUB.git
mkdir hipCUB/build
cd hipCUB/build
CXX=hipcc cmake -D BUILD_TEST=off ..