GPU Toolkit with GPU Direct

For researchers who need to perform signal processing or image reconstruction using GPUs, Verasonics offers the GPU Toolkit with GPU Direct – only available in a Linux OS environment – which accelerates data transfer by:

1. Increasing memory copy speed, which reduces latency and improves frame rate.

2. This option also provides programming tools and capabilities that enhance the overall development process.

Accelerate Data Transfer

Available as a licensable option, the GPU Toolkit with GPU Direct Package includes an NVIDIA GPU card, software license(s), and a Verasonics Host Controller. This option includes the following features:

User selected memory allocation of the Vantage and Vantage NXT buffers utilized in the image reconstruction process. The current default for the system is the CPU. New choices include:

CUDA allocated host memory – enabling up to 2X faster asynchronous memory copies between the host computer and the GPU device

CUDA allocated unified memory – simplifying CUDA programming by utilizing NVIDIA’s unified memory architecture
An Event Profiling Tool that allows users to see timing information for each operation in the event sequence, which can help to identify and eliminate bottlenecks.
A 1-step compilation of CUDA source code into .mex files that does not require the MATLAB® Parallel Computing Toolbox.
Example scripts illustrating the implementation of RF or IQ data processing. These examples can be tested to run synchronously or asynchronously with different memory allocation methods for comparison.

Verasonics understands that many researchers have already invested time and money in their own GPU hardware. A GPU diagnostic tool (gpuDiag) can be used by existing Verasonics users to determine if their GPU card(s) are compatible with GPU Toolkit option features. Support for GPU Toolkit options varies; the Verasonics sales team can assist with customer questions.

Direct Data Transfer

For users performing their own reconstructions or processing on the GPU, the GPU Toolkit with GPU Direct allows researchers to move data directly from Verasonics Systems to the graphics card.

Representation of the Standard data flow from the Verasonics system using the DMA to the RAM, then to the CPU, and finally to the GPU memory
Representation of the Data flow with GPU direct the data goes from the Verasonics system using the DMA directly to the GPU memory

This will benefit applications by:

Reducing latency

Increasing throughput

Freeing up the CPU for other tasks

Download the GPU Toolkit Options datasheet