📜 ⬆️ ⬇️

Will servers on ARM be able to make Intel nervous?

My acquaintance with the ARM version, close to the server one, began with the installation of a Ubuntu server version on CubieTruck and the deployment of a test site on Drupal. I wondered if the work of the engine on Cubie would differ greatly from the work of the same engine on the server based on the Intel Dual-Core G850, which I use for my projects. I will tell you about the results of the experiment, as well as about the differences between the ARM and x86 platforms as server hardware, in my post.


A set of processor instructions ARM began to emerge in the 80s of the twentieth century. By 1985, the first version of the ARM-based computer appeared, and a year later the second version saw the light. In those years, the market was just forming, companies conquered the market and chose their ways. As a result, Intel has become a world leader in the production of processors for user PCs and servers, with its serious practical developments and production base. ARM also focused mainly on the development of architecture, cores and third-party licensing.

Another difference is in the direction of development. Intel started with more powerful user solutions and took the path of reducing power and power consumption. ARM Holdings started with weak and cost-effective solutions, such as processors for mobile phones and routers, then for smartphones, tablets and other devices, mostly battery-powered and often without space for active cooling.
')
So, let's consider what advantages and disadvantages are attributed to ARM platforms in comparison with other existing ones (mainly with the industry leader - Intel).



Production policy


Intel is the sole developer and manufacturer of its chips and processors. Accordingly, he and only he sets the direction of development of chips based on their considerations. ARM itself does not produce anything, it only sells the copyright to the architecture and design of the chips. The chips are produced by all and sundry who received a license from the copyright holder, the company ARM Holdings. This allows the technology to be more flexible and adapt to the needs of both the specific consumer and the market as a whole. Manufacturers, on the other hand, have the opportunity to create customized SoCs (System on a Chip - systems containing not only the processor itself, but also any additional hardware components needed to solve the problem) on a chip. The license fee is charged for each released chip and, according to public data, amounts to several dozen cents per unit.



In the case of Intel, the company provides a full cycle, from production to marketing. That is why the company is rather rigidly focused on sales volumes. How could it be otherwise if the new chip factory, which will become obsolete in a few years, costs about $ 5 billion? Naturally, the cost of such plants falls on the shoulders of future buyers of processors. In addition, the company must produce universal mass solutions and can not afford to concentrate on niche highly specialized tasks.

Interesting fact: since October 2013, Intel also has a license to manufacture ARM chips. The companies agreed that Intel will build a 4-core 64-bit ARM Cortex-A53. It should also be mentioned that AMD announced its plans to release an ARM-based server back in 2012.

In January 2014, AMD presented the first line of AMD Opteron A1100 Series server processors, codenamed “Seattle”. These are 4- or 8-core ARM Cortex-A57 processors, containing up to 4 MB of cache 2 levels and up to 8 MB of cache 3 levels, supporting DDR3 and DDR4 ECC memory with a capacity of up to 1866 MT / s, 8 SATA ports, 2 10Gbps Ethernet ports and other buns . Together with the processor, a platform for developers was presented.

I could not find information about the price and ways to get it, because if someone in the know, please share. Based on the press release, we can conclude that the company plans to divert the server market from mass solutions, where approximately the same configurations are used for different tasks, to more specialized options.

Simpler architecture


ARM supports fewer microprocessor commands and functions than x86. On the one hand, x86 supports more embedded protections, buns and enhancers, on the other hand, ARM has a simpler topology, simpler assembler and requires fewer transistors on a chip. And this, in turn, gives another bonus.

Smaller size and power consumption


Fewer transistors consume less power and take up less space with the same technology. This has become a key factor in choosing processors for mobile platforms in the early stages of their development. The batteries were not as good as they are now, and most processors consumed so much energy that they needed mains power.

Currently, technologies are not changing in favor of ARM. Intel is one of the few companies capable of financing fundamental development. The size of the transistor also affects the power consumption. Intel scientists and engineers have already achieved 22nm and even 14nm technology, while most other manufacturers are forced to settle for more affordable 28nm.

In favor of less power consumption and space saving, the X-Gene family released by AppliedMicro testifies. Their solution based on the ARM architecture absorbed a 64-bit processor with a frequency up to 2.4 GHz, a memory controller, PCI, SATA, 2x10Gbps Ethernet and other useful things in one chip. According to the developer, the solution, adapted specifically for web applications, cloud and blade systems, is close in performance to modern processors of the E3 and E5 series, while consuming 50% less electricity and allowing equipment to be placed up to 4 times denser due to the fact that Almost everything you need to work is placed on a single chip. Due to this, owners will be able to significantly reduce TCO (Total Cost of Ownership; equipment maintenance costs). The product is currently available as a developer kit.



However, a blade version prototype was also created. The size of the entire board can be estimated by imagining the size of the memory bar.



AppliedMicro, together with its partners, also prepared ARM assemblies of popular Linux branches and server software for them, thanks to which anyone can now experiment with a set for developers and experience all the advantages of the system.

Less heat dissipation


This feature arises as a consequence of lower power consumption and the laws of physics. The principle of operation of modern chips is about the same, therefore, the efficiency is also about the same. It is logical that chips that consume less energy, with equal efficiency will dissipate less heat. This reduces cooling costs and saves space.

Higher performance


Actually, the net performance of most ARM below. Only some recent developments are catching up with Intel processors. However, if we consider the performance per watt, here, according to many data, ARM still wins. Also, according to some reports, ARM wins in the performance / $ ratio. The only caveat is that most of this data is presented in tests of the ARM platform developers themselves, therefore some consider them dubious. Allowing an independent testing of platforms for developers, an ordinary man in the street is unlikely to be able to because of their cost.

Personal practice


The only device I had the opportunity to practice with was CubieTruck. This small but proud device carries a 2-core ARM Cortex-A7 processor, 2 GB of DDR3 480MHz RAM (960MT / s), SATA port, micro-SD slot, Ethernet, Wifi, Bluetooth and flash defiantly bright diodes in progress. Unfortunately, I did not have an extra SSD disk, so I had to install an OS on a micro-SD class 10. I assume that with SSD the results of the experiment would have been better. Apache, PHP + APC, MySQL were also installed.

I used Drupal 7 as a test site with automatically generated 10,000 pages of content, of which 100 were randomly selected using the Views module. Testing was done using the ab utility with a total of 1000 requests and 10 competing requests. Actually, the number of requests 30 testing failed because ab flew out on timeout. Similar tests were conducted on a server based on Intel Dual-Core G850 with 4 GB of RAM and SATA drives. Some of the test values ​​are listed in the table below.



The average time of page generation by an inactive server (not during tests) in both cases is 3-4 seconds, but on Cubie there were spontaneous manifestations of instability, expressed in this absolutely unexpected generation time:



Thus, although CubieTruck loses the Intel G850, but it shows itself quite decently as for hardware that was not designed as a specialized server. Small applications can work on the basis of it, while the cost of equipment (charge plus SSD disk) will be about $ 150-200. It is also worth mentioning the fact that the board consumes only about 10-15W. The only trouble with the practical use of Cubie is the placement problem. Data centers refuse to put such non-standard equipment on collocation, and it is not economically viable to occupy even a full-fledged 1U. If suddenly one of the Kiev data centers reads these lines, who is ready to take part in the experiment and put Cubie on for a month or two, I will be glad to cooperate.

I look forward to the release of Cubieboard 8 , which, unfortunately, was somewhat delayed. The device will contain 8 cores with an operating frequency of about 2GHz and support more than 2 GB of RAM, which brings it closer to the entry-level servers in performance. I hope that in the next few months, developers will complete the final debugging and present their brainchild to the world.

In conclusion


Summing up, we can say that although ARM-based platforms are still far from mass production, they already show some competitive advantages in comparison with key market players. The only question that remains is how the players will act: will they oppose or will they still join the new technology? It seems that the second scenario is becoming more and more likely. According to some experts, by 2019, the ARM architecture will cover up to 25% of the server market, and AMD is already claiming to claim leadership in this segment.

Now the ARM market is in an embryonic experimental state. Someone is developing test platforms, someone is assembling blade systems, someone is launching computing clusters based on Cubieboard , which processes files in dozens of gigabytes.



However, now you can seriously think about the commercial application of solutions based on ARM in the near future, as did a small French startup NanoXion, which offered colocation service for Cubieboard . The range of commercial ideas can be huge - from a micro-dedicated service based on the same CubieTruck, which even in 1U can be pushed a lot, to actively participate in promoting solutions like X-Gene and AMD Opteron A1100 to the market, while the market is practically free. It remains to turn on the fantasy and enter it correctly.

Source: https://habr.com/ru/post/229689/


All Articles