Skip to main content

GPUs In The Era of Artificial Intelligence


Graphical Processing Units was first developed in 1999 by NVidia and was referred to as the GeForce 256. This GPU model could process 10 million polygons per second and had more than 22 million transistors. The GeForce 256 was a single-chip processor with integrated transform, drawing and BitBLT support, lighting effects, triangle setup/clipping and rendering engines.

Conventionally, computing power is related to the quantity of CPUs and therefore the cores per processing unit. WinTel started to breach the enterprise data center, application performance and information throughput were directly proportional to the number of CPUs and available RAM. Whereas these factors are important to achieving the desired performance of enterprise applications, a new processor began to gain attention – Graphics Processing Unit or GPU.

But within the time of Machine Learning and AI, GPUs found a new place that makes them as applicable as CPUs. Deep learning, and advanced machine learning technique that is heavily utilized in AI and Cognitive Computing. Deep learning powers many synopses including autonomous cars, computer vision, cancer diagnosis, speech recognition, and many of alternative intelligent use cases.

GPUs became more popular as the requirement for graphics applications expanded. Eventually, they became not just an enhancement but a necessity for optimum performance of a computer. The designed logic chip enables the process of quick graphics and video exertion. Generally, the GPU is connected to the CPU and is completely distinctive from the motherboard. The random access memory (RAM) is connected through the accelerated graphics port (AGP) or the peripheral component interconnect express (PCI-Express) bus. Some GPUs are integrated into the northbridge on the motherboard and use the main memory as a digital storage area, but these GPUs are slower and have poorer performance.

Most GPUs use their transistors for 3-D computer graphics. However, some have increased memory for mapping vertices, equivalent to geographic information system (GIS) applications. Some of the more modern GPU technology supports programmable shaders implementing textures, mathematical vertices and accurate color formats. Applications like computer-aided design (CAD) can course over 200 billion operations per second and deliver up to seventeen million polygons per second. Many scientists and engineers use GPUs for additional in-depth calculated studies utilizing vector and matrix features.

Comments

Popular posts from this blog

Machine Learning for Automated Data Science Tools

For the past five years, a team of research scientist at MIT’s Laboratory for Information and Decision Systems has been trying to create automation tools that enable the subject matter experts to use ML. The team first divided the process into a discrete set of steps. For example, the first step involved searching for buried patterns with predictive power, known as “feature selection engineering.” Second is called "model selection," in which the best modeling technique is chosen from the many available options. These steps are automated; releasing open-source tools to help domain experts efficiently complete them. Then these automated tools are grouped together, turning raw data into a trustworthy, conveyable model over the chain of seven steps. This chain of automation makes it possible for subject matter experts even those without data science experience to use machine learning to solve business problems. Through machine learning for automated data scie...

Neural Networks and Deep Learning

Neural Networks and Deep Learning have grown widely over the last few years. By using neural network architecture, softwares of AI can go through and check millions of images to find the right tone to fit any image. This method could be used to colorize still frames of white and black movies, surveillance footage or any number of images. Because neural networks can derive data from any number of resources with access to millions of sounds and videos, it can make predictive judgments. Neural network architecture can now synthesize audio to fill in the blank spots of a silent video. Neural network architecture can perform translations of text without preprocessing the sequence so that the algorithm can learn word relationships. The network then processes these relationships through its image mapping technology to create a contextual solution to a translation issue. By getting access to a wide variety of images and learning the context of each one, neural network architecture can...

Does Neural Networks help in inventing Anti-Cancer Medications?

Scientists have for the first time applied a generative neural network to create new pharmaceutical medicines with the desired characteristics. By using Generative Adversarial Networks (GANs) developed and trained to "invent" new molecular structures, there may soon be a reduction in the time and cost of searching for substances with potential medicinal properties. The researchers intend to use these technologies in the search for new medications within various areas from oncology and even anti-infective. Currently, the inorganic molecule base contains hundreds of millions of substances, and only a small fraction of them are used in medicinal drugs. The pharmacological methods of making drugs generally have a hereditary nature. For example, pharmacologists might continue to research aspirin that has already been in use for many years, perhaps adding something into the compound to reduce side effects or increase efficiency, yet the substance still remains the sam...