Project
Gapless Real-Time RF Analysis with NI PXI, LabVIEW and G²CPU
Summary
Organisation: G²CPU – developed by Natan Biesmans
Challenge: Process up to 1.25 GS/s of wideband IQ data in real time without dropping samples or relying on complex custom FPGA programming.
Solution: Combining the NI PXIe-5840, NI PXI hardware, LabVIEW and G²CPU enables continuous, gapless RF analysis with flexible FFT sizes and GPU acceleration, without requiring custom FPGA logic.
How do you analyze a full gigahertz-wide RF data stream in real time, without losing samples or relying on complex FPGA programming? G²CPU, developed by Natan Biesmans, demonstrates how NI PXI hardware, LabVIEW and GPU acceleration can work together to make this possible. Using a real-time spectrum analyzer (RTSA), the solution processes the complete data stream from an NI PXIe-5840 vector signal transceiver at 1.25 GS/s, continuously and without dropping samples. The system is designed to run indefinitely at full data rate, without requiring customers to discard samples or compromise on continuous analysis. By combining GPU computing performance with the flexibility and development productivity of LabVIEW, G²CPU makes high-performance, gapless RF analysis more accessible.
Get G²CPU for LabVIEW:
The challenge
Processing Wideband RF Data Without Losing a Single Sample
In wideband RF measurements, capturing large amounts of data is only half the challenge. That data must also be analyzed continuously, in real time. An NI PXIe-5840 can capture up to 1 GHz of instantaneous bandwidth at a sample rate of 1.25 GS/s, generating approximately 5 GB of data per second.
A traditional CPU cannot process this data stream fast enough. FPGAs offer the required performance, but they require specialized HDL expertise, involve lengthy compile cycles and are constrained by limited on-chip memory. Large FFT operations and changes to the analysis while it is running therefore become difficult to implement.
Recording data to disk and analyzing it afterwards is no complete alternative either. Short-lived signals, radar pulses, interference and frequency-hopping transmissions can easily be missed when processing is not continuous.
G²CPU’s challenge was to analyze the entire wideband data stream in real time, without dropping samples and without developing custom FPGA logic. At the same time, the solution needed to remain flexible, allowing engineers to change FFT settings and add their own analysis without rebuilding the entire implementation.
The solution
Combining NI PXI, LabVIEW and GPU Acceleration
G²CPU combines NI PXI-based RF acquisition with the LabVIEW development environment and the computing power of GPUs. The result is a real-time spectrum analyzer developed entirely in LabVIEW, using GPU acceleration for computationally intensive signal processing.
The system consists of several components:
- NI PXIe-1092 chassis: provides the modular hardware platform and high-bandwidth PXI Express backplane.
- NI PXIe-8881 controller: serves as the host system and orchestrates the application.
- NI PXIe-5840 vector signal transceiver: captures wideband IQ data with up to 1 GHz of instantaneous bandwidth at 1.25 GS/s.
- LabVIEW 2026 Q3 and the NI-RFSA driver: provide the application development environment and RF data streaming.
- RADX Technologies RTX Pro 4000 SFF (Blackwell) PXIe GPU: performs the intensive analysis in TCC mode.
- NVIDIA RTX 2000 Ada GPU: handles visualization in WDDM mode.
- G²CPU 1.7.0: makes GPU computing directly accessible within the LabVIEW dataflow environment.
A key advantage is that engineers can develop and run their analysis in LabVIEW without writing CUDA code or HDL themselves. Even data residing on the GPU can be inspected using LabVIEW probes.
This simplifies not only development, but also maintenance and further expansion. FFT sizes can be changed while the analyzer is running, and the architecture allows customers to add their own analysis methods and integrate additional instruments. Depending on the application, custom analysis can be developed in LabVIEW, Python or MATLAB.
The modular NI PXI architecture also provides an integrated measurement platform in which hardware components work together within a unified ecosystem. This reduces the complexity associated with combining separate instruments, managing cabling, synchronizing clocks and configuring triggers.
Video Wi-Fi 7: Real-time generation and transmission of an artificial Wi-Fi 7 signal, combined with a live 2.4 GHz signal received via an antenna. G²CPU uses GPU acceleration in LabVIEW to process and visualize the signals in real time.
The result
Continuous Real-Time Analysis at Full Data Rate
The real-time spectrum analyzer processes the full data stream from the NI PXIe-5840 at 1.25 GS/s, without dropping samples. By combining GPU acceleration with LabVIEW, G²CPU enables continuous, gapless analysis of wideband RF signals at full data rate.
The solution supports FFT sizes ranging from 2 points to hundreds of millions of points, giving engineers the flexibility to choose between extremely high processing throughput and fine frequency resolution. The FFT size can be changed to suit the measurement requirements without redesigning the hardware or recompiling FPGA code.
Beyond raw performance, G²CPU simplifies development and maintenance by keeping the analysis within LabVIEW. Engineers can adapt their processing workflows and integrate custom analysis in LabVIEW, Python or MATLAB without modifying the core architecture.
The result is a flexible, high-performance RF measurement platform designed to run continuously without requiring customers to discard samples to keep up with the data stream.
Want to know more?
Ask one of our experts
Dennis Motshagen
Account Manager & Application Engineer
Contact
