NVIDIA's DAQIRI Pipes 500 Gbps of Microscope Data Straight Into GPUs
NVIDIA rolled out upgrades to DAQIRI, its GPU-native data acquisition library, which feeds high-bandwidth sensor data straight into GPU memory instead of writing it to disk first. A Lawrence Berkeley National Laboratory camera built on it pushes over 500 Gbps — more than 500 times the bandwidth of standard electron energy-loss spectroscopy setups.
- Original source: blockchain.news.
- The source published it on Nov 20, 2026 at 2:57 AM.
- No second outlet has reported it yet. Treat it as a single-source claim.
NVIDIA has detailed new capabilities for DAQIRI, its GPU-native data acquisition library, aimed at letting scientific instruments analyze data while they are still collecting it. The software-defined library runs inside NVIDIA's Holoscan platform and moves high-bandwidth sensor data directly into GPU workflows, skipping intermediate disk storage. The newest release adds hardware packet reordering, improved throughput and low-latency NIC polling for time-sensitive processing.
The proof point comes from a collaboration between NVIDIA and Lawrence Berkeley National Laboratory's Molecular Foundry. A newly developed high-dynamic-range electron microscopy camera captures strong and faint signals at the same time, producing over 500 gigabits per second of data — more than 500 times the bandwidth of conventional electron energy-loss spectroscopy systems. FPGA readout electronics push that stream over Ethernet, and DAQIRI drops it straight into GPU memory for real-time handling.
From there, the GPU pipeline assembles and corrects detector frames, reduces them to energy-loss spectra and sharpens spatial resolution down to submicron levels. Because the results land during acquisition rather than after, researchers can inspect what they are getting and adjust acquisition parameters on the fly, reshaping how experiments are run.
NVIDIA says the same approach reaches past electron microscopy. By pairing sensors with GPU processing for real-time analysis and adaptive control, DAQIRI gives instrument makers room to reconsider their acquisition constraints and design more flexible, efficient hardware.
TopicsNVIDIA · DAQIRI · Holoscan · Lawrence Berkeley National Laboratory · Molecular Foundry
Source: blockchain.news
Written by AI from verified reporting.
More stories
- Tech & AI hindustantimes.com Caught by @pz126716 1 viewUS Closes Nvidia Loophole, Cutting China's AI Labs Off From Top Chips
- Tech & AI fool.com 2 viewsNvidia Drops 3.3% as Huang Exit and China Chip Curbs Outweigh Record Quarter
- Tech & AI cryptopolitan.com 2 viewsAlibaba beats revenue estimates as cloud jumps 53%, but warns growth will cool
- Tech & AI bbc.com Caught by @bailey 3 viewsCanada rolls out the welcome mat for AI data centres as US voters cool

