Introduction to Turing Deep Learning Workstation

        Artificial Intelligence, or artificial intelligence, like immortality and interstellar travel, is one of the best dreams of mankind. Although computer technology has made great strides, so far, no computer has produced a sense of "self". Yes, a computer can be very powerful with the help of humans and lots of readily available data, but without both, it can't even tell a cat from a dog.

        Turing (Turing, everyone knows it. The originator of computer and artificial intelligence, corresponding to his famous "Turing machine" and "Turing test" respectively) In his 1950 paper, he proposed the idea of ​​Turing's test, namely , dialogue through the wall, you will not know whether you are talking to a person or a computer. This undoubtedly sets a high expectation for computers, especially artificial intelligence. But half a century later, the progress of artificial intelligence is far from the standard of the Turing test. This not only makes people who have been looking forward to it for many years, disheartened, thinking that artificial intelligence is a sham and related fields are "pseudoscience".

       But since 2006, breakthroughs have been made in the field of machine learning. The Turing experiment, at least not so elusive. As for technical means, it not only depends on the parallel processing ability of cloud computing for big data, but also on algorithms. This algorithm is, Deep Learning. With the help of Deep Learning algorithms, humans have finally found a way to deal with the age-old problem of "abstract concepts".
        In June 2012, the New York Times disclosed the Google Brain project, which attracted widespread public attention. This project is jointly led by Andrew Ng, a famous machine learning professor at Stanford University, and Jeff Dean, the world's leading expert in large-scale computer systems, using a parallel computing platform of 16,000 CPU Cores to train a kind of "deep neural network" (DNN) , Deep Neural Networks) machine learning model (there are a total of 1 billion nodes inside. This network is naturally not comparable to the human neural network. You know, there are more than 15 billion neurons in the human brain, interconnected nodes. That is to say, the number of synapses is more like the number of sand in the Milky Way. Someone once estimated that if the axons and dendrites of all nerve cells in a person's brain are connected in turn and pulled into a straight line, they can be connected from the earth to the moon, from the moon back to the earth), has achieved great success in fields such as speech recognition and image recognition.

        Andrew, one of the project leaders, said: "We did not frame the boundaries ourselves as we usually do, but directly put massive data into the algorithm, let the data speak for itself, and the system will automatically learn from the data." Another person in charge Jeff said: "We never tell the machine during training: 'This is a cat.' The system actually invented or understood the concept of "cat" by itself.

      UltraLAB Turing Supercomputing Workstation GX490M is a CPU+GPU supercomputer system integrating ultra-high frequency + multi- GPU + mass storage, based on office quiet environment, applied to graphics, images, and all-round computing .

Compared with the deep learning computer systems on the market, significant advantages:

n   Completely in the office environment (silent level), not bothered by noise

n   Equipped with massive high-speed parallel storage based on PCIe bus (maximum capacity 180TB), low latency, supports up to 15 parallel reads, greatly improved hard disk io performance, and performance and management far surpassing traditional DAS/NAS storage systems

n   Equipped with super powerful computing power, up to 7 GPU cards, 38,200 computing cores, single-precision floating point 120Tflops

n   No special computer room is required, it does not take up too much space, and the maintenance cost is extremely low

n   No job scheduling system is required, and the management difficulty is greatly reduced

Configuration Specifications

 

Typical application areas

 Tailor-made high-performance, high-reliability computing hardware architecture for industry software and computing scale, coupled with unique professional-level hardware system optimization, compared with other brands of machines with the same configuration, the software computing speed is increased by 10%~10000 %, a perfect, high-performance professional graphics workstation.

    computing class

l Scientific computing, numerical simulation, mathematical programming  Matlab, Mathematica, Octave, Gurobi…

l Structure, fluid, thermal analysis, multiphysics coupling  Ansys, Abaqus, Nastran, LS-Dyna, Fluent, STAR CCM+, COMSOL Multiphysics...

l Electromagnetic field simulation  HFSS, CST Studio Suit, Feko, Agilent

l Electronic system design and simulation ( EDA)  HSPICE,

l Power system simulation  PSASP, Siemens PSS/E, ETAP, EMTP-RV, RSCAD, PSCAD

l Reservoir simulation, seismic data interpretation  Eclips, Nexus, Paradigm GeoDepth

l Quantum Chemistry, Molecular Dynamics Simulation  Gaussian, Lammps, Amber, NAMD, Docks,

l Material simulation   Material Studio

l Optical design and simulation   CodeV, Zemax, Lightools, OptiWave

l Financial calculation  KX, Wombat

l Artificial intelligence, deep learning, machine learning  Tensorflow, Torch, Café, Nvidia DIGITS

l Big data mining analysis  Hadoop

    Design class

l Pre- and post-processing  ICEM CFD, HyperMesh, ANSA, Femap

l 3D mechanical CAD/CAM Catia,  UG, Creo, Solidworks

l 3D architecture, infrastructure visual design  Revit, ArchiCAD, Microstion,

l 3D measurement, point cloud processing, reverse engineering

l Electrical design AutoCAD Electrical, Microstation V8, PDMS, PKPM, Protel, Aveva E3D

    Image class

l Digital photogrammetry ErDAS, PCI Geomatica

l Oblique photography 3D modeling  Smart3D, Pix4D

l geographic information system ArcGIS, MAPinfo

l Medical imaging  ViTAL, Mimics

l Industrial CT real-time image reconstruction VGStudio Max

l Meteorological data processing  WRF, MM5

l Video capture and structured analysis

    Virtual display, visual simulation, 3D visualization

l Virtual reality, visual simulation VegaPrime, 3DVIA, OpenGVS

l 3D architectural visualization  Lumion, SketchUp, Navisworks

l Large screen splicing and synthesis

l 3D data visualization (petroleum, medical, marine ...) Geoprobe, OpenInventor, VTK, Avizo

l Multi-channel real-time visual simulation

digital content creation

l 3D modeling /animation design  3DS Max, Maya, Photoshop, Sketchup, Lumion

l Video clips Premiere Pro, Mdeia Composer, FCP, Edius , Vegas

l Special effects compositing After Effects, Nuke X, Fustion Studio, Flame

l Rendering 3DSmax, Maya

l Grade  DaVinci Resolve, Speed ​​Grade, Lustre

    Private cloud, virtual, shared graphics server

l Shared graphic design server

l Shared CPU computing server

l Shared CPU+GPU supercomputing server

l Shared high io parallel storage server

Guess you like

Origin http://43.154.161.224:23101/article/api/json?id=325516379&siteId=291194637