canopy

platform-independent sparse matrix library for quantifying large PRA models

Setup & Install

This project uses git submodules. Begin by cloning recursively, or initializing and updating the repo.

git clone --recursive git@github.com:openpra-org/canopy.git

git submodule update --init --recursive

Development

Develop Using Docker/Linux

This is the preferred method. You can directly use the [Dockerfile](Dockerfile) to set up your build environment. If you're building directly in Linux, please follow the steps in the Dockerfile for your build environment. Development in other environments should be possible but has not been documented yet. Please open a pull request if you'd like to contribute to this effort.

docker build

docker context create ssh-box --docker "host=ssh://user@my-box"

docker build -t canopy:ssh-debugger --target=ssh-debugger .

docker run

In daemon mode:

docker run -d \
  --name=canopy-dev \
  -p 2222:2222 \
  --device=/dev/dri \
  --device=/dev/kfd \
  --cap-add=ALL \
  --security-opt seccomp=unconfined \
  --group-add video \
  canopy:ssh-debugger
  canopy:ssh-debugger

oder

Develop on macOS

Since macOS does not support OpenCL, CUDA, HIP, Level Zero, ...ugh anything but Metal and OpenMP, support is currently limited. It should still be possible to develop for testing purposes. Start by installing brew, and then AdaptiveCpp. Please open a PR if you make any progress here...

brew install seanfarley/adaptivecpp/adaptivecpp

Tested Versions

• CUDA: 12.4.1
• LLVM/Clang: 18.1.8
• AMD ROCm: 6.2.3
• OpenCL:
  • Intel OneAPI OpenCL 3.0
  • OpenCL 2.1 AMD-APP (3625.0)
  • OpenCL 3.0 CUDA

AdaptiveCpp Detected Backends

Note: No HIP device since this host does not have a AMD GPU.

coder@6954020af52d:~/projects/canopy$ acpp-info -l
=================Backend information===================
Loaded backend 0: CUDA
  Found device: NVIDIA GeForce GTX 1660 SUPER
Loaded backend 1: HIP
  (no devices found)
Loaded backend 2: Level Zero
  Found device: Intel(R) UHD Graphics 630
Loaded backend 3: OpenCL
  Found device: Intel(R) UHD Graphics 630
  Found device: Intel(R) Core(TM) i7-10700 CPU @ 2.90GHz
Loaded backend 4: OpenMP
  Found device: AdaptiveCpp OpenMP host device

OpenCL Backend support

coder@6954020af52d:~/projects/canopy$ clinfo -l
Platform #0: AMD Accelerated Parallel Processing
Platform #1: Intel(R) OpenCL Graphics
 `-- Device #0: Intel(R) UHD Graphics 630
Platform #2: Intel(R) OpenCL
 `-- Device #0: Intel(R) Core(TM) i7-10700 CPU @ 2.90GHz
Platform #3: Intel(R) FPGA Emulation Platform for OpenCL(TM)
 `-- Device #0: Intel(R) FPGA Emulation Device
Platform #4: NVIDIA CUDA
 `-- Device #0: NVIDIA GeForce GTX 1660 SUPER