Aerial CUDA-Accelerated RAN
Aerial CUDA-Accelerated RAN 24-2.1

L2 integration notes

Refer to <aerial_sdk>/cuPHY-CP/gt_common_libs/README.md for L2 to copy NVIPC source and build libnvipc.so for L2.

Recommend L2 to create nvipc_config.yaml and copy the “transport” configure from <aerial_sdk>/cuMAC-CP/config/cumac_cp.yaml. NVIPC configuration should be same between L2 and cuMAC-CP.

Refer to Aerial NVIPC section for NVIPC technical details.

Public header file for cuMAC message definitions is <aerial_sdk>/cuMAC-CP/lib/cumac_msg.h. Copy it to L2 and include it in L2 source.

Note for 24-2 release: Supports 4 cuMAC modules in functionality. But only the first module can be finished in 1 slot window yet.

  • multiCellUeSelection

  • multiCellScheduler

  • multiCellLayerSel

  • mcsSelectionLUT

Previous cuMAC-CP API Messages
Next Tests
© Copyright 2024, NVIDIA. Last updated on Sep 27, 2024.