Chinese car companies do not choose Nvidia nextThe DRIVE THOR chip will connect the intelligent driving chip and the entertainment chip, and one chip can solve the needs of intelligent driving and in-car entertainment application. Simply put, an Nvidia DRIVE Thor chip can do what the previous 8295 Orin chip can do, and even more powerful in terms of computing power.
NVIDIA's famous CUDA (a common parallel computing platform and programming model launched by Nvidia), also provides a variety of deep learning frameworks, machine learning libraries and computing libraries and other tools to simplify the development of models. This is the core competence of Nvidia products, it borrowsThe industry does not use X86 processors in the development of smart cars. In the field of autonomous driving computing, Nvidia's own GPU chips have leading computing advantages. In addition, the majority of low-power chips used in automobiles are based on the ARM architecture, and most of them have ARM patents. ARM has long pointed its finger at the automotive sector. They know that with the development of intelligent cars, more computing power is needed in cars.
There is still a long way to go before Chinese chip companies enter the automotive supply chain on a large scale. However, from a dynamic point of view, the participation of Chinese enterprises is increasing, and the proportion of Chinese enterprises' automotive chip business is increasingIt is only a matter of time to level with or even surpass the vehicle business. In the consumer market, car chips are not as concerned as other auto parts, and the former is usually not within the scope of after-sales maintenance that needs to be replaced.Architecture and design: Different AI chips may adopt different architecture and design concepts. Nvidia's GPU architectures excel in AI computing, such as its Volta, Turing, and Ampere architectures, which are widely used for deep learning and machine learning tasks. Performance and power consumption: The performance and power consumption of AI chips vary by manufacturer and product model. Nvidia's Gpus typically have high performance in AI computing and are capable of handling large-scale deep learning tasks.
At the same time, Nvidia's AI chips also have good energy efficiency, which can help devices achieve longer battery life.
Secondly, AI chips have advantages in terms of power consumptionMedium and high-end product models) and other advantages, and A card this is cost-effective, strong computing power, but there are shortcomings such as energy consumption and the lack of high-end game products. In general, both AMD and NVIDIA have their own strengths and weaknesses when it comes to performance and price.
NVIDIA's graphics cards have obvious performance advantages in the high-end market, performing well on compute-intensive tasks such as artificial intelligence and deep learning, and NVIDIA's graphics cards have stronger competitiveness in the high-end market.
This highlights Nvidia's dominant position in the high-end standalone GPU market. It is worth noting that withWith Intel's foray into the independent GPU space, the battle for market share has been particularly fierce in the past year. A year ago, Intel's white space gave Nvidia a commanding lead with 81 percent, followed by AMD with 19 percent. However, in the fourth quarter, Intel joined the fray with a 5% share, AMD's share dropped to 18%, and Nvidia remained the leader at 78%.
1, the RTX 4060 graphics card is expected to be available in early 2023.Nvidia has officially announced that the RTX 40 series graphics card will be released at the developer conference in September 2023. The first RTX 40 series graphics cards to be released will include the RTX 4080 and RTX 4090 flagship products. Based on Nvidia's historical release pace, the 60 series mid-range graphics cards in the RTX 40 series are typically released four to five months later than the flagship.
2, the market time of the system graphics card is October 12, 2022.
3, yes. The z390 is capable of 40 series graphics cards. Nvidia on September 20, 2022The new RTX 40 series graphics card was released. Nvidia's RTX 40 series graphics card released this time uses the new Ada Lovelace architecture core, which is manufactured by the 4nm process from TSMC, with 76 billion transistors and more than 18,000 CUDA cores. 70% more than the previous generation Ampere architecture core.
1. The Dojo D1 computing chip uses 5,760 computing powerThe 321TFLOPS NVIDIA A100 graphics card, a supercomputer built with 720 nodes, managed to achieve 18EFLOPSEFLOPS, 10 petabytes of storage, and a read/write speed of 16TBps. Note that this is the computing power of a single Dojo D1, the future Tesla.
2, good graphics card brands are: Asus, Yingchi, Nvidia, Gigabit, rainbow and MSI and so on. Asustek is a well-known Taiwanese computer brand and one of the world's largest manufacturers of motherboards and graphics cards. Asus graphics card with high performance,Known for its high stability, high heat dissipation and high appearance level, it has multiple series such as ROG, TUF and DUAL, covering different prices and needs.
3, Nvidia 1660super graphics card equivalent to AMD's RX5600 graphics card, are mid-end game graphics card, game performance is good, the price is more appropriate, are very good graphics card. In terms of heat dissipation performance is also relatively excellent, the highest temperature can only reach 58 ° C, and in terms of appearance is also relatively high-end, its design is relatively small and exquisite, the overall is particularly simple and fashionable, more suitable for the formation of some itx host. Replenishment.
4, NVIDIA A100 graphics card and 4090 graphics card compared as follows: Architecture. A100 adopts Ampere architecture; The 4090 uses the Volta architecture. Video memory. A100 video memory is 40/80GB; 4090 video memory is 24GB. Core performance. A100 has strong core performance; The 4090 has weak core performance. Cost performance. A100 has high cost performance; The 4090 has low cost performance. Applicable scenarios. A100 is suitable for deep learning training; 4090 is suitable for deep learning reasoning.
5, NVIDIA T239 belongs to the mid-range graphics cardLevel. NVIDIA's T239 graphics card, based on the Turing architecture, is a professional graphics processing unit (GPU) designed for data centers and high-performance computing environments. The graphics card is not aimed at ordinary consumers or gamers, but for professional users who need to handle large-scale data sets, perform complex simulations, or perform deep learning tasks.
6. Nvidia launches a new series of graphics cards every two years. In 2018 it will be 20 series, in 2020 it will be 30 series, and in 2022 it will be 40 series. Second of all, every time they release a graphics cardAt this point in time in September-October. In addition, the graphics card price in the graphics card market has begun to show a downward trend, which is likely because the 40 series graphics card is coming.