Intel NVIDIA processor audibles tear force war who have more

Intel NVIDIA processor: audibles tear force war who have more power? In recent years, Qualcomm, ARM and other companies in the mobile Internet, the strong chip chip, let Intel in the mobile Internet era, facing the attack, and NVIDIA (Ying Weida) to join, so Intel has been proud of backyard on fire. As GPU began to be widely used in the field of general computing, NVIDIA (Ying Weida) advantage is gradually emerging. In August 12th this year, NVIDIA announced the latest earnings, total revenue of $1 billion 428 million, an increase of 24%, while profits reached $253 million, an increase of 873%. This brings a small threat to Intel. And recently, the two chip manufacturers in the face of fierce confrontation staged. Previously, Intel said in the test report, the new Intel Xeon Phi processor family, its computing power is higher than the current market GPU processor. However, in response to this report, NVIDIA disagree, and its new design for the depth of learning to create the operating system DGX-1, for example, write a blog to fight back Intel. (for Intel statement) (Figure DGX-1) NVIDIA stressed that the operation system of DGX-1 deep learning is in response to the arrival of the era of artificial intelligence and launched by Tesla P100, NVIDIA Pascal technology innovation accelerator support building, and has pre installed multiple application software, including accelerated all major deep learning architecture library, NVIDIA SDK, DeepLearning DIGITS GPU, and CUDA driver training system. The system includes cloud management service access for container establishment and deployment, system update and application storage mechanism. Due to the advantages of implementing these applications on Tesla GPU and other advantages, compared with the speed of the application of various types of old version of GPU accelerated solutions, the speed can be up to 12 times faster. Intel in order to emphasize the performance advantages of the Xeon Phi processor, the report emphasizes the difference between the product and the performance of today’s GPU processor. The report pointed out that the Xeon Phi processor training speed 2.3 times faster than GPU, Xeon chip Phi in the extended way of multiple nodes is 38%, and up to one hundred and twenty-eight nodes, which is currently on the market can not do GPU. At the same time, the system consists of one hundred and twenty-eight Xeon Phi processor is 50 times faster than a single Xeon Phi processor, Xeon Phi processor will expand the advantage obviously. However, according to Intel’s argument, NVIDIA made a strong rebuttal, and pointed out that Intel is using 18 months ago data, compared to the four Maxwell GPU and four Xeon Phi processor. If you use the updated Caffe AlexNet data.相关的主题文章: