High Processing on-the-go training seamless scale-up Compact
NEXUS HIGH-PERFORMANCE ai SERVERs
AICRAFT is offering the NEXUS series of high-performance computing servers.
The NEXUS servers turn raw data into actionable signatures that can be pre-loaded and augmented over time as new data is acquired, effectively training the Artificial Intelligence (AI) engine on the go. With a massive 600 TOPS-equivalent AI performance by default (max 1200 TOPS), it will satisfy ultra-fast inferencing demands of transformers, convolutional neural networks and many other advanced machine learning models. This is supported by a number of power-efficient, hard cores Intel i7, Intel i9, ARMv8 or Intel Xeon family processors to handle any pre/post-processing and interface tasks seamlessly.
The NEXUS servers turn raw data into actionable signatures that can be pre-loaded and augmented over time as new data is acquired, effectively training the Artificial Intelligence (AI) engine on the go. With a massive 600 TOPS-equivalent AI performance by default (max 1200 TOPS), it will satisfy ultra-fast inferencing demands of transformers, convolutional neural networks and many other advanced machine learning models. This is supported by a number of power-efficient, hard cores Intel i7, Intel i9, ARMv8 or Intel Xeon family processors to handle any pre/post-processing and interface tasks seamlessly.
key benefits
benefits
- Up to 40% saving in TCO (multi million dollar p.a.)
- Significant reduction in energy consumption and form factor: pack more servers in the same workspace
- Significant reduction of raw data turned into actionable data: actionable data is readily available for MLOps
- Uniquely search millions of anomalies in billions of records
- Uniquely learn new knowledge and recognise same in real time: no need to retrain the ML models on entire data set.
AICRAFT’s GPU vs standard GPU
PARAMETER | STANDARD GPU | AICRAFT NEXUS |
Qualification standard | Industrial grade only | Industrial grade & radiation tolerant |
Number of classes | Limited (thousands) | Unlimited (millions) |
New sample data | Retrain model on the whole dataset (weeks)Reduced accuracy with incremental learning | Add instance to live database (milliseconds)Enhanced accuracy |
Accuracy | Supports fixed quantisation(Prone to loss of accuracy) | Supports variable quantisation(Optimised for desired latency) |
Infrastructure | Large system, lots of RAM & storage capacity(Bulky and costly spacecrafts) | Significantly smaller and more efficient system(Ideal for space) |
Programming skills | Specialised languages (CUDA, VHDL, Verilog)Transcoding customer code is prone to errors | Generic languages (C, C++, Python)Simple interface program to customer code |
Development time | Long | Short |
use cases
DATA CENTRE
Enhance your data warehouse, accelerate data mining; combined or set up independently from cloud infrastructure
GROUND STATIONS
Big Data edge processing of downlinked satellite data to forward informative data to the cloud or results directly to customers, reduce irrelevant data traffic, storage and cost
ON PREMISE
Tailored server racks and workstations for smaller enterprises
ON PLATFORM
In space, at sea, underwater, on aircraft, autonomous robots, mobile trucks or medical equipment, we provide solutions from industrial grade to space radiation tolerant for any orbit
core applications
Cybersecurity
Medical Imaging
Computer Vision
LiDAR Processing
Signal/Radar Processing
3D Modelling & Ray Tracing
Earth Observation
Satellite Telecommunication
performANCE EXAMPLES
Synthetic Aperture Radar
- Ultra-fast beamforming and back-propagation
- 20,000 I/Q data turned into 10,000 x 10,000 image in 50 seconds! That is 100 million reconstructed pixel intensities with histogram equalisation in 50 seconds.
- Training on-the-fly in milliseconds
- Enhanced accuracy using Fast Fourier Transforms
Video processing – CCTV
Single 1U server identifies up to 2 years of 10 events recorded by CCTV cameras per second.
our offer
High-performance operating system
Highly efficient GPU
Reliable software backend
Ongoing service & maintenance
WE PROVIDE
- Choice between different server sizes: single and dual-card computers, quad-card rack, 8-card rack, 16-card rack
- AI compiler, firmware, host management software
- Software licence for specialised libraries on demand
- Single software suite from edge to cloud and anything in-between
- One year maintenance with first purchase (renewable).
HOW WE SUPPORT customers
- Deliver the requisite performance for customer applications
- Provide hardware acceleration and interface programs to reuse existing customer code
- Scale the solution to any size dataset or performance requests
- Support seamless deployment and integration on premise and/or on platform
- Provide tailored customer training to maximise their business operations
- Offer tailored design services for software, firmware, web/mobile applications.