Carding 4 Carders
Professional
The processor will usher in a new era of AI with significantly lower power consumption.
Kaspersky Lab has made a new breakthrough in the field of artificial intelligence-neuromorphic chips that can simulate the work of the human brain with significantly lower energy consumption and high data processing speed. This was reported by the company to RIA Novosti.
The development is a response to the current state of affairs in the field of AI, which traditionally relies on the power of graphics processors with tensor cores. The new chips offer an alternative that surpasses classical approaches in a number of parameters.
One of the key advantages of the new neuromorphic chips is energy efficiency. An example is the fact that a chip containing 8 thousand neurons consumes only 4 milliwatts, while the human brain with its 90 billion neurons uses about 20 watts. In comparison with GPU-based data centers, where kilowatts and megawatts are involved, the difference in power consumption becomes even more noticeable.
The second important advantage is the speed of information processing. Neuromorphic chips paired with the corresponding camera can process up to 1250 frames per second. This allows the chips to analyze fast-moving processes, such as the movement of grains of sand in an hourglass or tracking raindrops on glass during rain.
The potential applications of the new technology range from industrial thin film deposition to traffic control systems. However, the high cost of neuromorphic cameras currently limits their mass distribution among consumers.
Kaspersky Lab said that new chips have already entered the international market, and the company is actively looking for partners to create pilot projects. Kaspersky Lab also needs cooperation in the field of microelectronics to produce chips with a process of 28 nanometers or less. The LC expressed the hope that a similar production facility will be established in Russia.
Kaspersky Lab has made a new breakthrough in the field of artificial intelligence-neuromorphic chips that can simulate the work of the human brain with significantly lower energy consumption and high data processing speed. This was reported by the company to RIA Novosti.
The development is a response to the current state of affairs in the field of AI, which traditionally relies on the power of graphics processors with tensor cores. The new chips offer an alternative that surpasses classical approaches in a number of parameters.
One of the key advantages of the new neuromorphic chips is energy efficiency. An example is the fact that a chip containing 8 thousand neurons consumes only 4 milliwatts, while the human brain with its 90 billion neurons uses about 20 watts. In comparison with GPU-based data centers, where kilowatts and megawatts are involved, the difference in power consumption becomes even more noticeable.
The second important advantage is the speed of information processing. Neuromorphic chips paired with the corresponding camera can process up to 1250 frames per second. This allows the chips to analyze fast-moving processes, such as the movement of grains of sand in an hourglass or tracking raindrops on glass during rain.
The potential applications of the new technology range from industrial thin film deposition to traffic control systems. However, the high cost of neuromorphic cameras currently limits their mass distribution among consumers.
Kaspersky Lab said that new chips have already entered the international market, and the company is actively looking for partners to create pilot projects. Kaspersky Lab also needs cooperation in the field of microelectronics to produce chips with a process of 28 nanometers or less. The LC expressed the hope that a similar production facility will be established in Russia.
