Get Started on Linux#

The Audio Effects SDK requires specific GPUs, specific distros of Linux and other software dependencies.

The Audio Effects SDK is designed and optimized for server-side (datacenter and cloud) deployment. Using this SDK for testing, experimentation, or production deployment of these SDKs outside of this use case is not officially supported.

Hardware Requirements#

The Linux SDK is supported on systems that have a minimum of 10 GB RAM and NVIDIA GPUs with Tensor Cores.

Hardware

Required Version

NVIDIA GPU

NVIDIA GPUs with Tensor Cores (NVIDIA Turing™/NVIDIA Ampere/Ada/Hopper/Blackwell)

  • Turing: NVIDIA Tesla® T4

  • Ampere: A2, A10, A16, A30, A40, A100

  • ADA: L4, L40

  • Hopper: H100

  • Blackwell: B100, B200, RTX PRO 6000

Note

The SDK supports Multi-Instance GPU (MIG) only on Tesla A30 and Ampere A100. When MIG is enabled, the GPU instance and corresponding compute instance must be defined, regardless of whether the SDK is executed on a specific GPU instance or on the entire GPU.

Note

For best performance with NVIDIA T4 and other server GPUs, be sure to use a server that meets the thermal and airflow requirements for these types of products. For the latest list of qualified servers, refer to the Qualified System Catalog.

Software Requirements#

Software

Required Version

Linux distribution

64-bit Linux distribution.

Supported distros:

  • Ubuntu (18.04)

  • RHEL8

  • CentOS8

  • Debian 10+

NVIDIA Graphics Driver for Linux

  • 570.26 or later.

  • 470/535/550 can be used with CUDA Forward-Compatible Upgrade.

For more information, refer to Using Older Drivers with CUDA with CUDA Forward-Compatible Upgrade.

CUDA/TensorRT/CuDNN

Note: All libraries that are required to use the SDK are in the package, under external/cuda, and do not need to be separately installed.

  • CUDA: 12.8.1

  • TensorRT: 10.9.0.34

  • CuDNN: 9.7.1

Sample program compilation

  • CMake: 3.10 or later

  • g++