Robert Schmidt
f8d9c54eff
Integrate GPU-Accelerated Channel Simulation + CI Test Pipeline This MR builds on top of MR !3588 (feat: GPU Accelerated Channel Simulation Pipeline). The main portion of the changes is directly taken from MR !3588, and focuses on integrating the GPU-accelerated channel simulation. On top of that, this MR adds CI-related configurations to introduce a new channel simulation testing pipeline into the CI --> RAN-Channel-Simulation This MR introduces a CUDA-based pipeline to offload the multipath_channel and add_noise channel simulation functions to the GPU. This enables large-scale, real-time physical layer simulations not possible on the CPU. For full documentation, please see the file at openair1/SIMULATION/TOOLS/DOC/gpu_acceleration.md. System & Compiler Requirements: An NVIDIA GPU with a compatible driver and the CUDA Toolkit installed. This feature was developed and tested on the GH200 server using the GCC 12 toolchain. To ensure reproducibility, building with GCC 12 is highly recommended. Build Instructions The primary flag to enable this feature is -DCUDA_ENABLE=ON. The following command shows a sample configuration used during development on the GH200 server: ./build_oai --phy_simulators --cmake-opt "\ -DCUDA_ENABLE=ON \ -DUSE_UNIFIED_MEMORY=ON \ -DUSE_ATS_MEMORY=OFF \ -DCMAKE_C_COMPILER=/usr/bin/gcc-12 \ -DCMAKE_CXX_COMPILER=/usr/bin/g++-12 \ -DENABLE_TESTS=ON" Memory Model Configuration The pipeline supports multiple CUDA memory management modes, which can be configured at compile time. ATS Memory (-DUSE_ATS_MEMORY=ON) is enabled by default. To use another memory model, you must first disable ATS. For a detailed explanation of these models, please refer to Section 3 of the gpu_acceleration.md report.
OpenAirInterface documentation overview
This is the general overview page of the OpenAirInterface documentation. This page groups links to general information, tutorials, design documents, radio integration, and special-purpose libraries.
IMPORTANT NOTE:
Before reading this documentation, we strongly advise you to keep your own repository rebased on develop
or at least to checkout the documentation on the version of the repository you are using.
Then the documentation will better reflect the features available in your repository so that you may avoid some errors.
Beware if you previously pulled the develop branch that your repository may be now behind develop.
[[TOC]]
General
- FEATURE_SET.md: lists supported features
- GET_SOURCES.md: how to download the sources
- BUILD.md: how to build the sources
- code-style-contrib.md: overall working practices, code style, and review process
- cross-compile.md: how to cross-compile OAI for ARM
- clang-format.md: how to format the code
- sanitizers.md: how to run with ASan/UBSan/MemSAN/TSan
- environment-variables.md: the environment variables used by OAI
- tuning_and_security.md: performance and security considerations
There is some general information in the OpenAirInterface Gitlab Wiki
Tutorials
- Step-by-step tutorials to set up 5G:
- RUNMODEM.md: Generic information on how to
- Run simulators
- Run with hardware
- Specific OAI modes (phy-test, do-ra, noS1)
- (5G) Using SDAP and custom DRBs
- IF setups and arbitrary frequencies
- MIMO
- How to run OAI with O-RAN 7.2 FHI
- How to run a 5G-NSA setup
- How to run a 4G setup using L1 simulator Note: we recommend the RFsimulator
- How to use the L2 simulator
- How to use the OAI channel simulator
- How to use GPU-accelerated channel simulation
- How to use multiple BWPs
- How to run OAI-VNF and OAI-PNF: how to run the FAPI/nFAPI split, including some general remarks on FAPI/nFAPI.
- How to use the positioning reference signal (PRS)
- How to use device-to-device communication (D2D, 4G)
- How to run with E2 agent
- How to run the physical simulators
- How to setup OAI with Nvidia Aerial and Foxconn
- How to setup OAI with LDPC accelerators (Xilinx T2/Intel ACCs)
- How to setup OAI with the XDMA FPGA LDPC accelerator
- How to do a handover
- How to setup gNB frequency
- How to use the RT data recording app
- How to use packages
Legacy unmaintained files:
L2NFAPI_NOS1.md,L2NFAPI_S1.md: old L2simulator, not valid anymore
Designs
- General software architecture notes: SW_archi.md
- Information on E1
- Information on F1
- Information on how NR nFAPI works
- Flow graph of the L1 in gNB
- L1 threads in NR-UE
- Information on gNB MAC
- Information on gNB RRC
- Information on analog beamforming implementation
- Information on the UE 5G NAS implementation
- Information on UL-MIMO: UL-MIMO specific notes
Building and running from images
- How to build images
- How to run 5G with the RFsimulator from images
- How to run 4G with the RFsimulator from images
- How to run physical simulators in OpenShift
Libraries
General
- The T tracer: a generic tracing tool (VCD, Wireshark, GUI, to save for later, ...)
- OPT: how to trace to wireshark
- The configuration module
- The logging module
- The shared object loader
- The threadpool used in L1
- The LDPC implementation is a shared library
- The time management module
Radios
Some directories under radio contain READMEs:
The other SDRs (AW2S, LimeSDR, ...) have no READMEs.
Special-purpose libraries
- OAI has two scopes: one based on Xforms and one based on imgui, described in this README
- OAI comes with an integrated telnet server to monitor and control
- OAI comes with an integrated web server
Testing
- UnitTests.md explains the unit testing setup
- Component tests are under
tests/. Currently, there is a simple CU-UP tester, see the corresponding README.md. - TESTBenches.md lists the CI setup and links to pipelines
- The CI setup uses a custom framework to run end-to-end tests.
Developer tools
- formatting is a clang-format error detection tool
- iwyu is a tool to detect
#includeerrors - docker-dev-env is a ubuntu24 docker development environment
- doc_best_practices.md: overall best practices for writing documentations