Skip to content

Use amrex::GpuArray instead of std::array in deposition (#1183) #1381

Use amrex::GpuArray instead of std::array in deposition (#1183)

Use amrex::GpuArray instead of std::array in deposition (#1183) #1381

Workflow file for this run

# Copyright 2022-2023
#
# This file is part of HiPACE++.
#
# Authors: Axel Huebl, Maxence Thevenet
# License: BSD-3-Clause-LBNL
name: 🐧 local CUDA
on:
push:
branches:
- development
pull_request:
branches:
- development
concurrency:
group: ${{ github.ref }}-${{ github.head_ref }}-cudaLocal
cancel-in-progress: true
jobs:
build_nvcc:
name: local V100 NVCC 11.1 DP
runs-on: self-hosted
if: github.event.pull_request.draft == false
env:
CXXFLAGS: "-Werror"
steps:
- uses: actions/checkout@v4
- name: Build & Install
run: |
source $HOME/profile.hipace
which nvcc || echo "nvcc not in PATH!"
cmake -S . -B build \
-DCMAKE_VERBOSE_MAKEFILE=ON \
-DAMReX_CUDA_ARCH=7.0 \
-DHiPACE_COMPUTE=CUDA \
-DHiPACE_PRECISION=DOUBLE \
-DAMReX_CUDA_ERROR_CROSS_EXECUTION_SPACE_CALL=ON \
-DAMReX_CUDA_ERROR_CAPTURE_THIS=ON
cmake --build build -j 8
- name: Run Tests
run: |
source $HOME/profile.hipace
conda activate openpmd
ctest --test-dir build --output-on-failure