Sample programs

Sample programs for eGrabber are provided in a dedicated package named egrabber-<OS>-sample-programs-<YY.MM.RE.BU>.<EXT> where <OS> is the operating system (linux, macos, or win) and <YY.MM.RE.BU> is the version number of the package.

These sample programs will use Coaxlink by default. To use them with Grablink:

• either define EURESYS_DEFAULT_GENTL_PRODUCER=grablink in your environment;
• or pass Grablink() to the EGenTL constructor.

Similarly, to use Gigelink:

• either define EURESYS_DEFAULT_GENTL_PRODUCER=gigelink in your environment;
• or pass Gigelink() to the EGenTL constructor.

And for Playlink (which requires a license¹):

• either define EURESYS_DEFAULT_GENTL_PRODUCER=playlink in your environment;
• or pass Playlink() to the EGenTL constructor.

eGrabber C++ code snippets

cpp/egrabber contains the following code snippets:

eGrabber Python sample programs

eGrabber Recorder Python sample programs

eGrabber Recorder C# sample programs

NVIDIA RDMA

NVIDIA RDMA is only suppported on Linux.

NVIDIA RDMA samples require an NVIDIA GPU that suppports RDMA.

The NVIDIA RDMA samples allocate memory on the GPU and announce this memory using NvidiaRdmaMemory.

See the following files in the eGrabber sample programs:

• cpp/egrabber-snippets/samples/503-grabn-cuda-rdma-process.*
• cpp/nvidia-cuda with the command line argument cudaRDMA
• python/cuda-rdma.py

Preliminary note:

Please read the Supported Systems section and its IOMMUs subsection in https://docs.nvidia.com/cuda/gpudirect-rdma/index.html#supported-systems if the system on which the NVIDIA GPU is installed has an IOMMU.

The IOMMU may be disabled from the BIOS settings or via kernel command-line parameters (https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html).

NVIDIA RDMA support is only available with the Proprietary flavor of the NVIDIA drivers.

When the Proprietary version cannot be used, RDMA might be achived alternatively with Euresys DmaBufMemory.

An example of this alternative solution with the Open flavor is provided in the cpp/egrabber-snippets/samples/504-grabn-cuda-dma-buf-fd-process.cpp file from the eGrabber sample programs.

Installation instructions:

• NVIDIA CUDA drivers:
  • Follow the installation instructions from:
    https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_network
  • Make sure to explicitly select the Proprietary flavor with the extra argument --kernel-module-type=proprietary
• NVIDIA driver sources:
  • These are needed to produce the Module.symvers file associated with the installed nvidia driver. This file will be required to install the eGrabber package
  • Select the appropriate driver from https://www.nvidia.com/download/index.aspx?lang=en-us
  • Make sure to download the version that matches the nvidia-<version> already installed in /usr/src/
  • Extract the archive with -x option
  • Change to kernel directory in the extracted archive (not kernel-open directory)
  • Run make module
  • The file Module.symvers should have been generated
• eGrabber package
  • Extract the egrabber-linux-x86_64 archive
  • Install the package with the following command:
    • sudo NVIDIA_KERNEL_PATH=<dir path containing Module.symvers> ./install.sh

The line Enabling NVIDIA RDMA build should appear during the installation of the eGrabber package.

A successful build can be confirmed if the command lsmod | grep coaxlink (or lsmod | grep grablink) indicates that coaxlink (or grablink) module depends on nvidia.

DMA-BUF

DMA-BUF is only suppported on Linux.

The DMA-BUF sample requires an NVIDIA GPU that suppports DMA-BUF.

The NVIDIA DMA-BUF samples allocate memory on the GPU and announce this memory using DmaBufMemory.

See the following file in the eGrabber sample programs:

• cpp/egrabber-snippets/samples/504-grabn-cuda-dma-buf-fd-process.cpp

Preliminary note:

Please read the Supported Systems section and its IOMMUs subsection in https://docs.nvidia.com/cuda/gpudirect-rdma/index.html#supported-systems if the system on which the NVIDIA GPU is installed has an IOMMU.

The IOMMU may be disabled from the BIOS settings or via kernel command-line parameters (https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html).

When the Proprietary version of the NVIDIA driver cannot be used, RDMA might be achived alternatively with Euresys DmaBufMemory as long as the linux kernel supports DMA-BUF and the NVIDIA device supports the CU_DEVICE_ATTRIBUTE_DMA_BUF_SUPPORTED attribute, please check the NVIDIA documentation for more information.

Installation instructions:

• NVIDIA CUDA drivers:
  • Follow the installation instructions from:
    https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_network
• eGrabber package
  • Please follow the default linux eGrabber package installation instructions.