Skip to main content

The most all-domestic server CPU chip in history has a big bottom

In August 2018, China National Procurement Center (CNCNET) confirmed the agreement suppliers of server products for 2018-2019 through open bidding. The shortlist included seven domestic CPU servers: Loongson CPU server, Feiteng CPU server, Shenwei CPU server, Huawei CPU server, Haigang CPU server, Hongxin CPU server and Megabite CPU server.

These CPUs are divided into three main genres:
CPU with independent intellectual property rights: Loongson, Shenwei
CPU with partial intellectual property rights: Feiteng, Huawei
CPU without independent intellectual property rights: Haiguang, Grand core, mega core
Let's take a closer look at the Campus in each of these three genres.
I. Comparison of technical parameters
The technical specs for these latest CPUs are as follows, with Xeon Platinum 8180 added for Intel CPUs:
Among them, the instruction set and microstructure are the two most important parameters. The instruction set is equivalent to the DESIGN program of CPU, while the microstructure is equivalent to the implementation of the CPU design program. The instruction set is relatively simple, while the microstructure is the most technical thing.
Both Loongson 3B4000 and Shenwei SW1621 have adopted instruction sets and microstructures with their own intellectual property rights, but their manufacturing processes are the most backward, and the main frequency and core number of loongson are also the lowest. STMicroelectronics, the OEM of their long-term cooperation, is located in Europe and the United States. Strictly speaking, the loongson is designed in China and manufactured in Europe and the United States. Shenwei has a military background and its CPU must be made by a Chinese company.
The FT-2000 and Huawei Kunpeng 920 both use ARMv8 instruction set permanently licensed by ARM and design their own microstructures. Huawei kunpeng 920 USES the most advanced 7nm manufacturing process and claims to be the fastest ARM server.
Megabyte Core Kh-37800D USES VIA licensed x86 instruction set and microstructure, Hygon 7185 USES AMD licensed x86 instruction set and microstructure, and Macro core CP1 USES IBM licensed Power instruction set and microstructure. None of these has proprietary intellectual property rights, which could easily be blocked by the US government.
Ii. Comparison of performance parameters
The performance parameters of these latest CPUs are compared as follows. It is said that Hagwon Hygon 7185 is AMD EPYC 7551, but its data is not available. Here we compare AMD EPYC 7551's data, and what we compare is SPECint_rate_2006, the hourly computing capacity of multi-processor (complete machine) :
As can be seen from the above table, in addition to HAkuang's Zen architecture authorization from AMD, which is actually a comprendo CPU, Huawei Haiti's Kunpeng 920 has the best performance, which is also related to its adoption of the most advanced 7nm manufacturing process at present. It is worth mentioning that the product specification of loongson 3B4000 has been released for a long time, but the product has not been released, perhaps due to no small difficulties.
I should mention that none of the above domestic Cpus can find the official test data on SPEC official website. The data I have collected is directly released by the manufacturer and may contain a lot of moisture. Even with the moisture, it's a long way from the Xeon Platinum 8180 released by the UK in 2017.
3. Comparison of manufacturers
Let's take a look at the three major schools of CPU manufacturers, many of which have prominent backgrounds:
Let's describe the development prospects of the three schools respectively:
1. Independent intellectual Property Group
The loongson is supported by the Chinese Academy of Sciences and the Beijing municipal government. Shen Wei is even more formidable, behind is the fifty-sixth Research institute of the General Staff Of the national Army, namely this time by the United States government to join the entity list "Wuxi Jiangnan Computing Technology Research Institute." Due to the use of a self-developed instruction set and microstructure, with independent intellectual property rights, it is impossible to be strangled by foreign countries, but it also brings about a difficult problem of marketization. Although this problem cannot be solved for a while, and it will take a long time and process to cultivate users and ecological circle, for national information security, enterprises with independent and controllable development capabilities such as Loongson and Shenwei are required to develop continuously with the support of the state.
2. Part of intellectual Property school
Behind Feiteng is the National University of Defense Science and Technology, which has a military background. The development of Feiteng has received strong support from the country. Behind Hess is Huawei, China's most high-tech private enterprise and the only Chinese private enterprise listed in the Fortune global 500, spending tens of billions of yuan on research and development every year. Hess should be the Chinese enterprise with the best prospect of domestic CPU marketization. Although it is a purchase of the ARMv8 instruction set and has no independent knowledge products, there is nothing wrong with it in order to accelerate the marketization process. After all, Hess does not have strong government background like Feiteng. Due to the obstruction of the US government, Huawei may not be able to obtain the subsequent new VERSION of ARM license, and may be forced to develop its own instruction set based on the ARMv8 instruction set, which may be a good thing for Huawei.
3. No independent intellectual property rights group
Signs in the core, sea ray and macro core is completely based on the Intel, AMD, or IBM's instruction set and microstructure of chip design, no independent knowledge industry, although has the backing of state assets and the government, can only be used for domestic products, and because of the limitation of information security of the country is unlikely to enter is more sensitive to safety requirements of the unit. With a limited market and limited technology, the Zen architecture taking acquired is unlikely to be upgraded to the Zen2 architecture. Mega core and macro core are also facing the same problem, the outlook is not very optimistic. Macro chip is now facing the issue of staff unpaid wages and shareholders change repeatedly three times, the enterprise turbulence, the outlook is more pessimistic.
In short, the development of domestic server CPU, more resources will be inclined to the government to support the representatives: Shenwei and Feiteng; Huawei Haiti, which relies on market performance, wants to do well in both the national information security market and the enterprise market, perhaps doing neither well.


Popular posts from this blog

AMD's GPU technology enters the mobile phone chip market for the first time

In addition to the release of the Exynos2100 processor, Samsung also confirmed a major event at this Exynos event, that is, the custom GPU that they have worked with AMD for many years will soon appear and will be used on the next flagship machine. The current Exynos2100 processor uses ARM’s Mali-G78GPU core with a total of 14 cores, so the GPU architecture developed by Samsung will be the next Exynos processor, and the GPU will be the focus. This is probably the meaning of Exynos2100’s GPU stacking. The key reason. Dr. InyupKang, president of Samsung’s LSI business, confirmed that the next-generation mobile GPU in cooperation with AMD will be used in the next flagship product, but he did not specify which product. Samsung is not talking about the next-generation flagship but the next one, so it is very likely that a new Exynos processor will be available this year, either for the GalaxyNote21 series or the new generation of folding screen GalaxyZFold3. In 2019, AMD and Samsung reached

Apple and Intel want to join the game, what happened to the GPU market?

Intel recently announced that it will launch Xe-LP GPU at the end of this year, officially entering the independent GPU market, and will hand over to TSMC for foundry. At the 2020 WWDC held not long ago, Apple also revealed that it is possible to abandon AMD's GPU and use a self-developed solution based on the ARM architecture. It will launch a self-developed GPU next year. What happened to the GPU market? Why are the giants entering the game?    Massive data calls for high-performance GPU    Why has the demand for GPUs increased so rapidly in recent years? Because we are entering an era where everything needs to be visualized. Dai Shuyu, a partner of Aiwa (Beijing) Technology Co., Ltd., told a reporter from China Electronics News that visualization requires a large amount of graphics and image computing capabilities, and a large amount of high-performance image processing capabilities are required for both the cloud and the edge.    Aiwa (Beijing) Technology Co., Ltd. is an enterp

NVIDIA officially launches RTX 30 series mobile graphics cards

In the early morning of January 13, NVIDIA officially launched the RTX30 series of mobile graphics cards at the CES2021 exhibition. Ampere-based GPUs have also reached the mobile terminal, mainly including RTX3080, RTX3070 and RTX3060 models. In addition to improving game performance, the RTX30 series of mobile graphics cards have twice the energy efficiency of the previous generation, and support the third-generation Max-Q technology, mainly supporting DynamicBoost2.0 dynamic acceleration technology, WisperMode2.0 noise control, ResizableBAR (similar to AMD’s SAM technology) and DLSS. The third-generation Max-Q technology uses AI and new system optimization to make high-performance gaming laptops faster and more powerful than ever. These technologies include: ·DynamicBoost2.0: The CPU and GPU powers of traditional gaming notebooks are fixed, while games and creative applications are dynamic, and the requirements for the system will vary with the number of frames. With DynamicBoost2.0,