GPU Computing

cancel
Showing results for 
Search instead for 
Did you mean: 

What is needed besides LabVIEW RTE to run GPU-accelerated LabVIEW applications?

For LabVIEW applications that don’t utilize a GPU to accelerate the application, we develop applications on development workstations equipped with a LabVIEW license, and then execute these applications on test workstations equipped with LabVIEW RTE. Our plan is to develop some GPU-accelerated applications on development workstations equipped with both a LabVIEW license and a LabVIEW GPU Analysis Toolkit license, and then execute these applications on test workstations equipped with LabVIEW RTE.

My question is what needs to be installed on the test workstations besides the same LabVIEW RTEs we currently have installed on the test workstations. I understand one required workaround is to add two files to our standard test workstation software image - "For a 32-bit app, copy lvcuda32.dll to <windows_root>\system\lvgpu. For a 64-bit app, copy lvcuda64.dll to <windows_root>\System32\lvgpu."  Are there additional files that will need to be added to the test workstation software image? For example, I assume the Nvidia CUDA 6 downloads must be installed on our development workstations along with LabVIEW and the LabVIEW GPU Analysis Toolkit. Correct? Are there elements of the Nvidia CUDA 6 downloads that must also be installed on the test workstations along with LabVIEW RTE, lvcuda32.dll, and lvcuda64.dll? What I'm hoping for is to end up with a new standard test workstation software image on all of our 80 test workstations, allowing the test workstations to accommodate any LabVIEW GPU-accelerated application a user brings to an individual test workstation.

0 Kudos
Message 1 of 3
(8,343 Views)

The lvcuda*.dll are the resource files that ship w/ the LabVIEW GPU Analsysis Toolkit. They do not include functions provided by NVIDIA. You will need to install NVIDIA's CUDA Toolkit (v4.0 or later) on each test system.

The function wrappers in the LabVIEW toolkit are configured to work with a CUDA toolkit installed in the default location(s). If you need to install the NVIDIA libraries to an alternate path, refer to documentation on how to find them from a LabVIEW application at runtime.

NOTE: It is recommended that you install the CUDA toolkit for your OS (e.g. N-bit CUDA for N-bit Windows). In the case of 64-bit Windows, both 32- and 64-bit versions of the libraries for the CUDA toolkit are installed allowing for the building and execution of both 32- and 64-bit applications.

0 Kudos
Message 2 of 3
(5,436 Views)

Thanks. That's exactly what I needed to know. I wasn't able to figure that out by looking through the various LabVIEW GPU Analysis Toolkit material I found on the NI web site.

0 Kudos
Message 3 of 3
(5,436 Views)