In early 2016, machine learning was still regarded as a scientific experiment, but it has now been widely used in data exploration, computer vision, natural language processing, biometrics, search engines, medical diagnosis, credit card fraud detection, securities market analysis, Applications in voice and handwriting recognition, strategy games and robotics. In this short period of one year, machine learning has grown faster than expected.
DeloitteGlobal's latest forecast report pointed out that in 2018, large and medium-sized enterprises will pay more attention to the application of machine learning in the industry. Projects deployed and implemented with machine learning will double in comparison to 2017 and will double again in 2020.
At present, more and more types begin to enrich the new term "AI chip", including GPU, CPU, FPGA , ASIC , TPU, optical flow chip and so on. According to Deloitte, in 2018, GPUs and CPUs are still the mainstream chips in machine learning. The market demand for GPUs is about 500,000, and the demand for FPGAs in machine learning tasks exceeds 200,000, while the demand for ASIC chips is around 100,000.
It is worth noting that Deloitte said that by the end of 2018, more than 25% of the data centers used to accelerate machine learning in the data center will be FPGAs and ASIC chips. It can be seen that FPGA and ASIC are expected to achieve a rise in the field of machine learning.
In fact, some users who started using FPGA and ASIC chip acceleration earlier mainly use machine learning inference tasks, but soon, FPGA and ASIC chips will also have some training in module training. Play.
In 2016, global FPGA chip sales have exceeded $4 billion. In the early 2017 report "CanFPGAsBeatGPUsinAcceleraTIngNext-GeneraTIonDeepNeuralNetworks", the researchers said that in some cases, the speed and computing power of the FPGA may be stronger than the GPU.
At present, Amazon's AWS and Microsoft's Azure cloud services have introduced FPGA technology; domestic Alibaba also announced cooperation with Intel to accelerate cloud applications with Xeon-FPGA platform; Intel recently It is emphasized that the data center can adjust the cloud platform through FPGA to improve the execution efficiency of machine learning, video and audio data encryption and so on.
In addition, although ASIC is a chip that performs only a single task, there are many manufacturers of ASIC chips. In 2017, the total revenue of the entire industry is about $15 billion. It is reported that Google and other manufacturers began to use ASIC in machine learning, and chips based on TensorFlow machine learning software have also been released.
Deloitte believes that the combination of CPU and GPU has greatly contributed to the development of machine learning. If future FPGA and ASIC solutions can also exert sufficient impact on improving processing speed, efficiency and cost reduction, then machine learning applications will be able to make explosive progress again.
French Power Strip,Surge Protector Power Strip,Power Strip With Flat Plug,Overload Protector Power Strip
CIXI KYFEN ELECTRONICS CO.,LTD, , https://www.kyfengroup.com