NVIDIA GPUDirect allows higher bandwidth and lower latency between NVIDIA GPUs and third party PCI-Express cards such as network adapters, frame grabbers and FPGA modules. We’ve implemented GPUDirect for 4DSP’s FPGA cards. These FPGA cards can be equipped with a wide range of analogue-to-digital and digital-to-analogue converters (ADCs and DACs) as well as frame grabbers and other I/O modules. When using the 4DSP PCI-Express IP core, drivers and API, the Hybrid DSP GPUDirect library can be used on top of this to improve and simplify data transfer between FPGA card and GPU. Please send an email for more information.


CUDAfy is a set of libraries and tools that permit general purpose programming of NVIDIA CUDA Graphics Processing Units (GPUs), AMD GPUs and Intel CPUs (OpenCL) from the Microsoft .NET framework. The aim of this project is to create the leading open source tools for this task; combining flexibility, performance and ease of use.

European Space Agency Fast Fourier Transform Chip SDK

The new space qualified FFTC processor from ESA provides very high performance floating point FFTs at low power. Hybrid DSP have produced an SDK to simplify the development of applications for this chip and the two current FFTC development boards from NLR and Astrium. More information here.

Data Acquisition and Processing – FPGA – ADC – DAC – Video

Hybrid DSP work closely with 4DSP to offer a range of high end data acquisition and processing modules. 4DSP offer the most extensive range of FMC modules in the industry.

  • Range of video input, digitizers and waveform generators – FMC and Front Panel
  • cPCI, PMC, XMC, PCIe and VPX Xilinx FPGA modules (Virtex-4, -5, 6, -7, Kintex-7)
  • IP cores – Fourier Transform, Wideband DDC, Polyphase filter bank
  • Reference designs for Xilinx ML605, VC707, Zynq, KC705