Server memory to a regular computer. Microscope on nails: should you put server hardware in your home PC? What Determines Demand for Memory: Supplier Views
Hello Giktime! Popular belief says that a neighbor's grass is always greener, and computers that meticulous entrepreneurs buy for their needs are more reliable and more productive than models flavored with marketing in retail. A whole caste of enthusiasts hunt for server hardware and worship enterprise-grade hardware performance. We are figuring out whether large organizations are really splashing around in "IT paradise", or have geeks created an idol for themselves out of nothing?
There are no barriers to enthusiasts, especially if these barriers are erected by cunning marketers who divided all electronic devices into corporate and consumer! Because even in the media advertising the mysterious "user experience", software and hardware developers blur out, they say, "the camera of this smartphone provides professional quality pictures!", And in another way the stamp about professionals who do not use nonsense has been exploiting for a long time. And if we are looking for the notorious "professional equipment" and quality of service, then it is better to ask the hardware and methods of service of the corporate class, right?
The motives that are guided by restless enthusiasts lie on the surface - even if consumer technology develops more vigorously due to the appetites of buyers, "battle-hardened" corporate-class components will clearly be more reliable, and even cheaper in the secondary market. Somehow geeks play on video cards for workstations, they collect powerful and "eternal" home PCs with server stuffing! So it makes sense to try your luck?
And of course, there is a fraction of this very meaning in such an undertaking, but with the acquisition of corporate "attributes" for home conditions, you can "get stuck" and, at best, overpay for unclaimed functionality, and at worst - go into a minus in comparison with the options, available to the retail buyer. Let's figure out what the catch is in using hardware designed for corporations.
The server is also a game one. Intel Xeon in home PCs
The first thing tech enthusiasts love to use from the corporate segment is server processors. Not exotic, but the most "understandable", that is, based on the x86 architecture. This pleasure is not cheap, therefore, "zeonovody", relatively speaking, include two camps with slightly different landmarks in the construction of a PC:
Xeon - initially not for games and "racing" in benchmarks, but sometimes useful
Enthusiasts targeting high-end components... This is the level when large-scale versions of Intel Core i7 are already lacking, and when looking at the LGA-2011 platform (of any generation), the thought comes to mind that the "supercharged" Core i7 offer "the same eggs", only in smaller quantities and without overclocking.
Because, since we are talking about the price, there have been moments in history when eight-core Xeons turned out to be one third cheaper and much "colder" than 6-core Core i7 Extreme Edition. For example, this was the case after the debut of Intel Haswell-E chips in 2014 - firstly, the difference in price between the six-core Core i7-5960X and the "civilian" quad-core i7-4790K was a measly 15%. And secondly, the junior server eight-core Xeon E5-2609 v4 cost about 30% cheaper than the candidate from the Haswell-E camp. At the same time, in contrast to the “just” Core i7, the Xeon has a lower TDP level and there is no graphics integrated into the processor useless for enthusiasts.
At the same time, the L3 cache in all three models is piled up in tons, and the frequency is even lower in the Xeon, but the beliefs "in the work of the core are not superfluous" and "very soon the games will be optimized so that they work quickly on 8 or more cores" do not give thrifty fans of the speed of rest, after which the hot guys send the younger versions of the Xeon to the Intel X99 chipset and ... do not admit to anyone how things are in games.
Therefore, four cores diluted with Hyper-Threading almost always turn out to be more efficient in games than eight low-frequency "pots" in the Xeon, which cannot even be overclocked in any way (locked multiplier, near-zero bus overclocking).
"Kulibins" who wanted to modernize the old platform at minimal cost... For example, to replace the old Core 2 Duo processor, not the old Quad, but a much cooler and higher-frequency quad-core Xeon X5460, which using a simple adapter can be installed not into a server motherboard with Socket 771, but into a “civilian” motherboard for Socket 775.
The main thing in such a scenario is to take care of high-quality cooling (server "stones" flaunt a TDP of about 120 W instead of 95 W for standard quad-core processors), but in the end this option of upgrading from a very old platform to a "tolerant old" justifies itself, especially since on some motherboards, the processor can be overclocked up to 4 GHz.
And after all, "Zions" have advantages with which they compensate for their multi-core sluggishness in games! For example, the ability to fence multiprocessor configurations, with which video / music / photo encoding and CAD modeling are much faster than in the top-end Core i7 Extreme. Support for register memory with ECC, for example, allows you to correct errors on the fly, which comes in handy when the uptime is large (the server is the same!). Support for "horse" amounts of RAM and a huge number of cores will also come in handy when the server needs to process incoming connections as quickly as possible. But all this is almost useless on a home PC.
And it is useful for him - a lot of cores at a high frequency. If these conditions are met, the processor itself is compatible with LGA 2011 or LGA 2011-3 platforms and is cheaper than “just” Core i7 - there is a point in purchasing it. Otherwise, it is better to either get by with massive quad-cores with eight threads, or design a workstation for specific use cases (rendering, encoding).
High-frequency Intel Xeons (if they are cheaper than mainstream CPUs) can be a good help not only in work, but also in games (source: ferra.ru)
Mow frags on a workstation with hacked NVIDIA drivers
If using a server processor can be played against rather than with the hardware installed, the graphics that should be used for video modeling or design have historically been cool in gaming disciplines. In the confrontation between AMD and NVIDIA, even scenarios of "misuse" of video accelerators have always been different: "red" gaming video cards have recently been snapped up by miners, and NVIDIA Quadro, so historically, has been persuaded to retrain into a gaming video card.
NVIDIA Quadro professional graphics cards are significantly more productive than their gaming counterparts
Moreover, Quadro is quite suitable for these purposes - the fact is that gaming GeForce is most often a professional video card with partially disabled GPU pipelines (from marketing considerations to chip rejection) at a more affordable price. For example, the new professional video card Quadro P6000 contains the most "complete" version of the GP102 graphics chip and for this reason outperforms the cool gaming GeForce 1080 by almost 20%, and the mighty Titan X based on the same Pascal architecture invariably leaves behind.
In general, among fans of NVIDIA video cards, a proprietary sport has long been formed - with the help of a hardware modification of the GeForce to the Quadro (for example, the GTX 680 into an analogue of the Quadro K5000 in terms of performance), and game lovers, on the contrary, cross the hedgehog with make professional video cards work faster in shooters / pokatushki / rushing. Such activity does not allow "to play as intended", but the persistence of enthusiasts can only be envied.
In mobile workstations, almost every NVIDIA Quadro video card has a funny pattern: every NVIDIA Quadro mobile video accelerator is equal to a gaming GeForce class lower in gaming tasks and a couple of levels more cool gaming GeForce in CAD disciplines.
NVIDIA Quadro mobile performance versus GeForce peers (source: msi.com)
For example, the Quadro M2000M in games shows itself at the level of the GeForce GTX 960M, but as soon as it comes to simulation, it "jumps" in the results to the GeForce GTX 980M. Roughly the same ratio holds true for other Quadro models: the M5000M competes with the GTX 980M in games, and the M1000M competes with the 950M in games.
NVIDIA Quadro M6000 vs. Fastest Gaming Graphics Cards
(source: techgage.com
Ice cream for children, flowers for a lady: priorities in corporate memory and storage
Server RAM is not compatible with motherboards in home PCs, not because someone decided to "spite" the end customers. It's just that the server RAM is arranged a little differently - it contains a register between the microcircuits and the system memory controller in order to reduce the electrical load on the controller and be able to install more modules in one memory channel.In other words, additional microcircuits and the ability to automatically recognize and correct errors greatly increases the fault tolerance of this type of memory, but also increases its cost. In a word, do not be surprised if you find that even low-frequency (by the standards of DDR4) modules will turn out to be 50% or more expensive than their “everyday” counterparts - the inhuman requirements for endurance in systems turned on around the clock have noticeably changed the server RAM. In everyday use, it will be neither faster nor more efficient than its "civilian" counterparts, so for high performance you should turn to gaming kits - for example, HyperX Savage, if you need convenient memory for overclocking for gamers, and HyperX Predator, if you want to squeeze out of the subsystem. RAM maximum. For the nominal frequencies, the budget Kingston ValueRAM is great - reliable, installed and forgotten once.
A server processor in a home PC can come in handy, but instead of register memory, it is better to purchase a standard DDR3 / DDR4 kit
Enterprise-class SSDs have also undergone a "tuning" towards reliability - they, for example, have the ability to flexibly manage the reserve volume for the needs of the controller. The larger the volume, the lower the wear of the cells and the higher the durability of the drive. And a huge number of algorithms that are effective in difficult operating conditions, especially in terms of data safety in case the drive is turned off in emergency mode. Reconfigured for minimal latency in multi-user access mode and struggle for stable performance even with freelance writes and reads. A home computer does not experience such a load, even if you "torture" the SSD with torrents. On the other hand, industrial SSDs are not champions in typical operations either - typical SATA drives will become outdated faster in terms of memory capacity than they will completely exhaust the number of rewrite cycles available for cells - a proven long comparative test with HyperX models. And records of speed with the same level of reliability have long been passed to drives based on the NVMe interface, which are implemented in one of the newfangled form factors "over" PCI-Express. In the Kingston / HyperX lineup, the Predator SSD PCI-E was and remains the "king of the hill".
The longevity gains of an enterprise-class SSD don't match the joy of a gaming PCI-e drive
If you can't, but really want to - then you can
Enterprise-class hardware is not so different from its "civilian" counterparts to be deemed unsuitable for work as a home PC, you just always need to proceed from whether the game is worth the candle. Because the situation is as follows:Buying a platform that uses Error Correcting Register Memory (ECC) for the home is a bad idea. The excess durability does not compensate for expensive components and the average (in comparison with gaming counterparts) level of performance will not please, especially since the prices for server memory are noticeably higher than for the average DDR3 / DDR4 module.
Enterprise drives in your home computer are needed if you are paranoid, extremely worried about the safety of data in the event of a power outage, and worried about the reliability of modern SSDs in general. Organizational drives let you maximize reliability for peace of mind.
A server processor for games ... an interesting and quite effective idea, but only when it comes to a cheaper (in comparison with mainstream analogs) and, most importantly, high-frequency model. Or about upgrading an old computer to a server CPU "with little blood", that is, almost for a song. And yes, ideally the platform should be borrowed from the "usual" Extreme series of mass processors.
Professional graphics cards do an excellent job not only with modeling, but also with games. But it should be remembered that in mobile workstations (with a "strangled" TDP), a professional video accelerator of the middle class will be able to compete in gaming disciplines only with gaming video cards of the budget class. And desktop professional video cards, in turn, although fast in all scenarios of work, are prohibitively expensive, and certainly not suitable for the role of an economy option for "work and play."
Be that as it may, you cannot save on high-quality and fast RAM ... But today you can! We remind you that from 2 to 20 February for all memory kits
By virtue of its specificity, the server theme is a relatively rare guest on the covers of IT periodicals. If you take an interest in the quantitative ratio of the "server" and "desktop" information existing on the Internet, you can get approximately 1:20. As for even more subtle matters, such as organizing a data storage system on enterprise-scale servers, it is very difficult to obtain such knowledge. Today we will touch upon, perhaps, one of the "quietest", but still present areas of the market for professional components, namely server memory.
Since the very phrase "server memory" is rather an abstract slang expression containing too many different meanings, we decided to combine the "general educational" part with the definition of the economic feasibility of the existence of this direction as such.
What is Server Memory? To conduct our conversation more substantively, for a start it would be nice to understand the terminology and explain what is meant by this phrase in this publication. So, server memory (hereinafter referred to as SP) represents memory modules with parity and error correction, as well as additional functionality to ensure greater stability (register buffered memory), created according to standards different from those used in desktop products and certified for use in servers from A-brands... We will not pretend to have this definition included in the textbook, but it seems to us that it reflects the essence.
The second question that I would like to discuss is how does server memory differ from desktop memory, in addition to the above additional bits, registers and buffers?
Differences between server and desktop memory
Manufacturers
The number of brands on this market is much more modest than on the “tabletop” field. In order not to get confused about who is doing what and for whom, we will divide the companies that issue the joint venture into several subcategories.
A-brands - memory created for specific server manufacturers with their unique markings on chips and modules. All top-tier developers (HP, Dell, etc.) typically take this approach to maximize overall profitability. In this case, it is probably not worth focusing on who specifically produces this or that module and / or microcircuit (it can be any company that won a tender or has direct contracts). The most important thing to know: modules sold, for example, under the Dell trademark, are exactly compatible with Dell servers, and are 100% likely to work in them.
Chip A-brands (manufacturers of memory chips and modules on them) - companies engaged in the production of both microcircuits and memory modules under their own trademark. This category includes Micron, Samsung, Hynix, Quimonda (formerly Infineon). By and large, it is they who regulate the memory market as a whole, since in total they produce about 70% of DRAM chips. All of the above companies have product lines of server modules of any standards. Of course, a tight connection to production allows us to have more competitive prices compared to companies focusing exclusively on server brands, but in this case there is another side of the coin - difficulties with certification. For example, a change in the series or generations of microcircuits leads to a change in marking (sometimes only chips, and sometimes modules), which requires new tests from server manufacturers to issue a conclusion on whether the new memory is suitable for use in their systems or not. There are situations when under the identical labeling on the market there are essentially two products with different characteristics (and both are original), which gives a lot of headaches to server builders.
Modular A-brands (manufacturers of modules on third-party chips) is the most common category. Among them are such famous names as Kingston, Corsair, Transcend, Apacer. Actually, such companies are often called "test" companies in relation to the joint venture, because their engineers spend most of their time testing modules for operability with commercially available server platforms. As a result, a situation arises that is in many ways similar to JVs from server A-brands, in addition, such firms have much less problems associated with labeling. Therefore, the end consumer or server assembler can easily get information that a module with "such" marking is suitable for use in "such and such" server, and it does not matter on the basis of microchips of which manufacturer it is made.
All three approaches have their positive and negative sides, but there is no reason to expect any changes in the presence or arrangement of brands in the JV market in the near future.
fault tolerance
In relation to the joint venture, such a rather desktop term as "reliability" is usually replaced by "fault tolerance", which more accurately reflects the meaning. Since such equipment has to work non-stop 99.9% of the time from the moment it is put into operation, much more rigorous approaches are used in production and testing than in the manufacture of desktop products.
For example, the technology of "artificial aging" provides the identification of manufacturing flaws within two days - during testing, the server modules are heated to 100 ° C, which allows them to quickly bring them into a state corresponding to two months of operation. Next, in a set of tests that heavily load the memory subsystem, there is a check for compatibility of modules with various server platforms, which takes about another day. As a result, a joint venture with a rejection rate of about 0.02% (one module per five thousand) gets into the channel.
In the same section, it is worth saying a few words about such a concept as a "successful model", which again came from the desktop world. It is a well-known fact that there are “successful” video cards that work equally well on any platform, “successful” hard drives that are compatible with most controllers, and “successful” memory modules that provide stable operation with almost all motherboards. As for the joint venture, everything is exactly the opposite. The main criterion by which the manufacturer is guided is the absence of "unsuccessful" models, because a well-created joint venture bar should work everywhere and always and without any reservations. So, if a product from brand X did not work with the server of brand Y, then such a joint venture will most likely not be allowed on the conveyor until the reason is determined. Of course, no one can afford a similar approach with regard to components for mainstream computers.
Criterias of choice
The end customer, choosing memory when assembling or upgrading his computer, is usually guided by the following criteria: brand (this also includes a guarantee), price, test results. In other words, it is important for him who the manufacturer is, what timings and frequencies can be obtained, and how much they will have to pay for all this. If you go up a notch and see what principles the assembly company is guided by, then the price immediately comes to the fore. She, in view of very tough competition in the market, usually chooses the cheapest option from the problem-free.
With the joint venture, the situation is as follows - integrators, as a rule, are primarily guided by the experience of operating certain modules in certain platforms and are guided by one (maximum of two) suppliers who are able to strictly comply with conditions such as stability of supply and quick resolution of technical issues with the manufacturer if they occur. Although the price is, of course, an important component, it is far from the determining factor. If a joint venture from one brand turns out to be 20% more expensive than another, but at the same time provides better compatibility, then the choice will most likely fall on her.
What Determines Demand for Memory: Server Vendor Opinions
Evgeny Bobruiko
everest server hardware product manager
In accordance with the company's policy, we try to offer the optimal solution for specific applications, always with a small margin for the future. If the client has "outgrown" his configuration, then in many cases it is more expedient for him to purchase a more modern and powerful server, and transfer the old one to other, less responsible and laborious tasks.
Server modernization services are in very low demand in our company (about 2% of systems are exposed to it). After all, a prematurely arising problem of improving the memory subsystem indicates either a significant increase in streaming tasks for the client (for example, due to an increase in the staff), or an error of the integrator who proposed an ineffective solution.
With regard to memory types, the FB-DIMM has excellent streaming performance, which is very much in demand when working with databases (especially OLAP), and also allows you to install a significant amount of RAM in the server. On the other hand, it has high latency, which is not always good. For DDR2 this figure is lower, but it also has lower bandwidth. It is to combat high latency that Intel has a good "weapon" - a large and shared cache for two cores in Woodcrest.
On the whole, it seems to me that FB-DIMM has very good prospects, since DDR2, due to the parallel method of data transfer, approaches the "ceiling" as the frequency increases. However, I do not exclude that at first Intel may make a budget version of the system logic based on DDR2.
Andrey Tishchenko
entry company manager
In my opinion, modernization of servers is a thankless task, since the entire component base is outdated and many problems cannot be solved simply by adding memory. Most of our customers are growing companies, constantly updating their server park, they adapt old models for less resource-intensive applications without modernization. Therefore, requests for installing additional memory are rare for us, such sales do not exceed 2-3% of the total supply.
I must say that there is no competition between DDR2 and FB-DIMM as such - there is a planned transition of the two main chipmakers to a new architecture. The popularity of this or that memory depends on who this period will be easier for. Intel has taken a big step forward by pushing the limits of the shared processor bus and lowering the power consumption of new CPUs. Nevertheless, the bet on FB-DIMM in the near future may work against the company. Compared to DDR2, its performance is worse: higher latency, power consumption and heat dissipation, and the cost of modules is about 10%.
With the transition to Socket 1207 (Rev F) and the adaptation of the memory controller on the core for DDR2, AMD retains the ability to work with memory at the core frequency, use the most popular standard, and scalability of multiprocessor multi-core platforms. In the future, the company plans to switch to FB-DIMM memory, but sees its trump cards in something else. Figuratively speaking, while Intel constantly stimulates developers of new types of memory and maintains high rates of DDR-DDR2-FB-DIMM evolution, AMD is more focused on meeting the price / performance / power consumption criterion.
Igor Przhegarlinsky
commercial Director of Onyx
If earlier on most servers a 32-bit operating system was installed, which imposed restrictions on the amount of memory up to 4 GB, then with the transition to 64-bit operating systems, the limit will increase to 32 GB (for 1-4-processor servers based on Windows Server 2003 Standard Edition ).
With the release of new Intel (Dempsey and Woodcrest) and AMD (Socket F) processors, the FB-DIMM and DDR2-667 ECC Reg. Both chip makers have long since abandoned support for a single standard, and the competition between memory types, as a result, will be reduced to competition between chip manufacturers.
The bulk of server memory is currently used in new systems - the share of modules sold for upgrades does not exceed 5%.
Performance
For a very long time, the joint venture lagged significantly behind the desktop in its speed characteristics. Suffice it to recall that the DDR400 standard, which managed to become widespread for desktop systems and laptops, came to server systems only with the advent of the Opteron. Even modern chipsets for powerful workstations continued to use register buffered DDR266 for a long time.
The next leap in professional systems based on Intel was the use of DDR2-400, despite the fact that in desktops, the start of DDR2 started at 533 MHz. For reference, note: for Itanium 2-based servers, only DDR200 with 128-bit access was generally used to provide the necessary bandwidth. The reasons why the creators of the platforms chose lower frequencies are clear: increasing the reliability and reducing the load on the memory controller integrated with the main set of logic.
Today, server technology is also required to perform as quickly as possible, which forces us to use the most modern standards that are not inferior to, and sometimes even surpass, those for desktop systems. Pay attention to the specifications of the latest server platforms - the total performance of the northbridge buses for Intel chipsets and integrated controllers for Opteron can exceed the fantastic mark of 30 GBps. And most importantly, at such speeds, you need to provide a 24/7 operation when running very resource-intensive applications.
Perhaps the situation is even more complicated with the memory subsystem. Today's standards are FB-DIMM 667 MHz for Intel and register buffered DDR2-667 with double parity for Opteron. The amount of memory to ensure comfortable work is constantly increasing, and, as you know, the probability of an error in RAM grows exponentially with increasing volume. As a result, manufacturers of modules and, first of all, JV microcircuits have to look for ways to ensure reliability no less than with the previously used DDR200 / 266 standards, as well as with increased volumes, now reaching 32 GB, huge frequencies of 667 MHz and maintaining the difference in cost is no more than 40% compared to desktop modules.
Price
Although the joint venture market is not subject to strong fluctuations, which are often observed in the desktop segment, at the same time it is quite dynamic, and such methods of competitive struggle as dumping, marketing promotions, OEM deliveries at reduced prices, etc. are not excluded. e. Today, the cost of JV modules differs on average by 20-50% from similar desktop modules. On the one hand, this is not small, on the other, you can recall a more complex technical implementation (additional microcircuits for parity and buffering control), the need to conduct a set of tests after the product leaves the assembly line, certification from the server manufacturer and, of course, a lifetime warranty. As a result, it turns out that the vendor's earnings on JV modules are hardly much more than on desktop "slats". And the modern market requires constant price reductions: just remember that FB-DIMMs began to fall in price literally from the day they appeared on the free market. So, if at first the cost of FB-DIMM was approximately twice as high as the joint venture of DDR2 and DDR standards, now this difference is much more modest and amounts to about 30-40%.
What Determines Demand for Memory: Supplier Opinions
Elena Krivoshienko
head of the sales department of memory modules and flash products of the Kiev-TEK company
According to our data, the volume of the Ukrainian segment of server memory accounts for 10-13% of the memory market as a whole, which in financial terms equals approximately 5-7 million dollars per year.
Over the past months of 2006, the sales structure of server modules has changed somewhat. DDR and DDR2 are sold in roughly the same quantities, with quite a few DDR modules being ordered recently to upgrade servers. As for FB-DIMM, today their share in the total sales structure is no more than 2%.
In quantitative terms, the server memory market is expected to grow. It can be assumed that in the next six months the sales structure will shift towards DDR2. Regarding FB-DIMMs, I believe that their share will gradually increase, although in the next six months it is unlikely to be significant. Solutions using this type of memory are still very expensive, primarily due to the cost of processors.
Dmitry Borovsky
general Manager of TNG
Due to the difficult political situation this year, the corporate segment of the IT market has slowed down. This has led to a decrease in the supply of complex solutions, which are based on servers. It is hoped that the situation will change at the end of the year, but it is difficult to guarantee that the demand for components, including server memory, will be fully satisfied during the hot season. Based on this, the annual volume of the server memory market in Ukraine can be estimated only roughly - according to our forecasts, it will be from 40 to 60 thousand modules, and in financial terms - from 4 to 5 million dollars.
I would like to note that in 2006 the sales of DDR333 server memory sharply decreased, practically the same happens with DDR400 - their share is now less than 5%. The main product today is DDR2-400 modules, and the demand for FB-DIMMs is due to the availability of motherboards for them on the market. FB-DIMM deliveries began only in the III quarter, and so far they occupy less than 5% of our sales (however, the upward trend of this indicator is evident).
There is no particular difference between trends in the domestic and international markets in general. For systems in which fault tolerance is a critical factor, manufacturers prefer memory from well-established vendors. Perhaps the only difference is that large world brands mainly use first-hand server memory in their products - from Samsung, Hynix, Micron. On the local market, there is a significant demand for modules from Corsair, Kingstone, etc.
Andrey Semenovsky
cEO of Nebesa
The server memory segment is just a part of the server market, however, due to the high cost of modules, it plays a more prominent role than memory in the desktop market. While desktop PCs are concerned with the release of successful memory models around which marketing is built, for servers it is the absence of unsuccessful ones. Problems with the availability and quality of server memory can lead to costs that are many times higher than the price of modules, which means that the responsibility for delivery times, the declared compatibility of models and quality in this case is much higher. And given the low ROI of server memory, you can imagine how difficult this segment is from a business point of view.
Of the peculiarities of the domestic market, I would like to note the rather high potential of Ukrainian server collectors. The volume of memory consumption is about 6-8 thousand modules per quarter, and the average price of the bar is close to $ 150.
Changes in the memory sales structure depend on those in the lineup of platform manufacturers (primarily Intel and AMD). By the time FB-DIMM was released, the ratio of DDR2 to DDR sales was 70:30. With the advent of fully buffered memory, a transition process began: the demand for platforms with DDR2 support is noticeably decreasing, and with FB-DIMMs, they are not being purchased very actively due to the rather high price and underdevelopment. It is noteworthy that during this period DDR sales increased - their share in the supply structure approached 50%.
Victor Shcherbyak
head of Sales Department, ASBIS Ukraine
The release of new server operating systems made it possible to significantly increase the maximum amount of memory in the system (earlier versions of the OS had significant restrictions on this parameter). Therefore, sales of modules both in piece and in monetary terms increased. According to our data, the server memory segment today accounts for about 5% of the total market for this product.
The demand for DDR has significantly decreased with the advent of new server platforms. The largest sales now fall on DDR2 modules, sales of FB-DIMMs are also growing, although the latter's share is still insignificant. According to my estimates, the sales volumes of both types of memory are likely to be equal by the end of the year.
In general, servers are required every year more and more scalability, flexibility, capacity, speed of operation, as well as reliability and safety of investments. Based on new technologies, server memory manufacturers are releasing modules to achieve these improvements.
Ukrainian realities
According to the information available from integrators, about 10,000 servers are sold in Ukraine per year (let's specify: not computers acting as a server, but systems using joint ventures). Even if this figure is too optimistic, the bottom line is that the number of JV modules sold is approximately 20,000 "strips", of which about a third is used to modernize previously assembled systems. For averaging, let's assume that the cost of one module is $ 100, therefore, the total volume of the joint venture segment in our country is approximately $ 2 million. Moreover, the number of firms offering such products does not exceed five, and they mainly work with vendors such as Kingston. Samsung and Hynix, of which the first two share about 70% of the Ukrainian market.
Probably, the information given in this article will seem too "basic" for some, but we think that for the first topic of the issue, devoted to server memory, it clearly will not hurt. And in more detail with the technical aspects of this type of product, we invite you to get acquainted in the next material.
Hello, friends! In this article, we have tried to answer many of your questions regarding RAM. ? How do I know what kind of RAM I have installed and how much? How to choose the right RAM for your computer. How do you know if your RAM is working in dual channel mode or not? Is it better to buy one 8GB DDR3 memory stick or two 4GB memory sticks each? And finally.
- If you are interested or, also read our articles.
- Hello admin, a friend of mine asks to install more RAM for him. Computer properties show 2 GB. They turned off the computer, opened the system unit, there is one RAM bar, took it out, but there are no marks on it. Interestingly, it was not possible to determine the model of the motherboard. The computer was bought a long time ago, so the question arose - how to find out the type of RAM that it needs? After all, RAM differs in type, frequency and timings.
- Hello! I wanted to buy additional RAM, took off the cover of the system unit, took out the RAM bar and I cannot decipher the information printed on it, the serial number is just written there and that's it. It is completely unclear at what frequency it works and what type it has, DDR3 or DDR2. How to distinguish DDR3 from DDR2, how do they differ in appearance?
- I have one 4 GB DDR3-1600 memory stick in the system unit, I want to put another 4 GB stick, but working at a higher DDR3-1866 frequency. Will my computer work properly, and most importantly, in dual channel mode?
My friend installed three RAM strips of different size and frequency in the system unit. Is this allowed? But strangely, his computer works fine! - Tell me how to check if my RAM is working in dual channel mode or not? And what conditions are needed for my memory to work in dual-channel mode. Same volume? Same frequency or same timings? How much faster the computer runs in dual-channel mode than in single-channel mode. They say that there is also a three-channel mode.
- What will work better, two 4 GB sticks of RAM in dual-channel mode, or one stick, but with a volume of 8 GB, respectively, will the memory operation mode be single-channel?
To find out all the information about a RAM module, you need to carefully consider it, usually the manufacturer marks the RAM with proper information about the frequency, volume and type of RAM. If there is no such information on the module, then you need to find out everything about the motherboard and the installed processor, sometimes this action turns into a whole investigation.
- Important notes: Friends, do not forget that all new processors have Intel Core i3, Intel Core i5, Intel Core i7 the memory controller is located in the processor itself (it used to fill the north bridge of the motherboard) and the memory modules are now directly controlled by the processor itself, the same applies to the latest AMD processors.
- This means that it doesn't matter what frequency of RAM your motherboard supports. It is important what frequency of RAM your processor supports. If your computer has a processorIntel Core i3, Intel Core i5, Intel Core i7, then the officially supported memory standards of these processors: PC3-8500 (DDR3-1066 MHz), PC3-10600 (DDR3-1333 MHz), PC3-12800 (DDR3-1600 MHz), it is at these frequencies that your RAM will work, even if the passport of the motherboard indicates that the motherboard can work with high frequency RAM PC3-19200 (DDR3-2400 MHz).
- It's another matter if your processor with unlocked multiplier, that is, with the letter "K" at the end, for example CPU Intel Core i7-4770K, 3.5 GHz.An unlocked multiplier means that in a computer with such a processor, you can install memory sticks of the highest frequency, for example, DDR3-1866 MHz or DDR3-2400 MHz, such a processor can be overclocked and during overclocking the RAM will operate at its own frequency of 2400 MHz ... If you install the RAM barDDR3-1866 MHz or DDR3-2400 MHz into a computer with a conventional processor, that is, with locked multiplier without letter" K ”at the end, for exampleIntel Core i7-3770, 3.9GHz, then such a bar will work at best at a frequencyDDR3-1600 MHz, and at worst the computer will not boot. Therefore, buy RAM suitable for your processor.
- Concerning processorsAMD recent years, they work with memoryPC3-10600 (DDR3-1333 MHz).
Firstly, all the information you are interested in should be on the RAM bar itself, only you need to read it correctly. I do not argue, there are memory bars on which there is practically nothing, but we can also cope with them.
For example, let's take a Hynix RAM bar, it contains the following information: 4 GB PC3 - 12800.
Which means the following:
firstly, the volume of 4 GB,
secondly, 1Rx8 - Rank is a memory area created by several or all chips of a memory module, 1Rx8 are one-sided memory ranks, and 2Rx8 are two-sided memory ranks.
As you can see, this bar does not say that it is DDR2 or DDR3, but the bandwidth of PC3-12800 is indicated. PC3 - the designation of the peak bandwidth belonging to the DDR3 type only (for DDR2 RAM, the designation will be PC2, for example PC2-6400).
This means that our Hynix ram is of the DDR3 type and has a bandwidth of PC3-12800. If the bandwidth of 12800 is divided by eight, we get 1600. That is, this DDR3 memory stick operates at 1600 MHz.
Read everything about DDR2 and DDR3 RAM on the website
http://ru.wikipedia.org/wiki/DDR3 and everything will become clear to you.
Let's take another RAM module - Crucial 4GB DDR3 1333 (PC3 - 10600). This means the following: 4 GB, DDR3 memory type, 1333 MHz frequency, the bandwidth of PC3-10600 is also indicated.
The manufacturer is Patriot, the volume is 1 GB, the bandwidth of PC2 is 6400. PC2 is the designation of the peak throughput belonging to the DDR2 type only (the designation of DDR3 RAM will be PC3, for example PC3-12800). Divide the 6400 bandwidth by eight and get 800. That is, this DDR2 memory stick operates at 800 MHz.
Another plank - Kingston KHX6400D2 LL / 1G
Manufacturer Kingston, bandwidth 6400, type DDR2, volume 1 GB. Divide the bandwidth by 8, we get a frequency of 800 MHz.
But on this bar of RAM there is more important information, it has a non-standard supply voltage for the microcircuits: 2.0 V - is set manually in the BIOS.
The RAM modules differ in the size of the contact pads and in the location of the cutouts. A notch will prevent you from installing a RAM module in a slot that is not intended for it. For example, you won't be able to install a DDR3 memory bar into the DDR2 slot.
Everything is clearly visible in this scheme.
Sometimes there will be no understandable information on the RAM module, except for the name of the module itself. And the module cannot be removed, since it is under warranty. But by the name you can understand what kind of memory it is. for instance
Kingston KHX1600C9D3 X2K2 / 8G X, it all means:
KHX 1600 -\u003e Operative operates at 1600 MHz
C9 -\u003e Timings (Delays) 9-9-9
D3 -\u003e DDR3 RAM type
8G X -\u003e 4 GB volume.
You can just type the name of the module in the search engines and you will find out all the information about it.
For example, the information of the AIDA64 program about my RAM. RAM modules Kingston HyperX installed in RAM slots 2 and 4, memory type DDR3, frequency 1600 MHz
DIMM2: Kingston HyperX KHX1600C9D3 / 4GX DDR3-1600 DDR3 SDRAM
DIMM4: Kingston HyperX KHX1600C9D3 / 4GX DDR3-1600 DDR3 SDRAM
Can I install RAM sticks with different frequencies in my computer?
The frequency of the RAM does not have to be the same. The motherboard will set the frequency for all installed RAM sticks to the slowest module. But I want to say that often a computer with bars of different frequencies is unstable.
Let's do a simple experiment. For example, take my computer, it has two identical Kingston HyperX RAM modules, DDR3 memory type, 1600 MHz frequency.
If I run the AIDA64 program in my Windows 8, it will show such information (see the next screenshot). That is, the program AIDA64 shows the simple technical characteristics of each of the RAM strips, in our case both strips have a frequency1600 MHz. But the programAIDA64 does not show at what frequency the ramps are currently working, you need to look at this in another program calledCPU-Z.
If you run the free CPU-Z program and go to the Memory tab, it will show at what frequency your RAM sticks work. My memory works in Dual mode, 800 MHz, since DDR3 memory, its effective (doubled) speed is 1600 MHz. This means that my RAM sticks work exactly at the frequency for which they are intended 1600 MHz. But what will happen if next to their RAM sticks operating at a frequency1600 MHz i will set another bar with frequency1333 MHz !?
Let's install an additional DDR3 memory stick in my system unit, operating at a lower frequency of 1333 MHz.
We look at what AIDA64 shows, the program shows that an additional 4 GB bar is installed, the frequency is 1333 MHz.
Now let's run the CPU-Z program and see at what frequency all three bars work. As you can see, the frequency is 668.7 MHz, since the memory is DDR3, its effective (doubled) speed is 1333 MHz.
That is, the motherboard automatically set the operating frequency of all RAM strips to the slowest 1333MHz module.
It is not advisable to install RAM strips into the computer with a frequency higher than the motherboard supports. For example, if your motherboard supports a maximum RAM frequency of 1600 MHz, and you have installed a RAM module operating at a frequency of 1866 on your computer, then in the best case this module will operate at a lower frequency of 1600 MHz, and in the worst case the module will work at its own frequency 1866 MHz, but the computer will periodically restart itself oryou will receive a blue screen when you boot your computer, in this case you will have to enter BIOS and manually set the frequency of the RAM to 1600 MHz.
Timings (signal delay) determine how often the processor can access the RAM, if you have a quad-core processor and it has a large L2 cache, then too long timings are not terrible, since the processor is less likely to access RAM. Can I install RAM sticks with different timings in my computer? Timings don't have to match either. The motherboard will automatically set the timings for all strips for the slowest module.
What conditions are needed for my memory to work in dual-channel mode Before buying RAM, you need to study as much information about the motherboard as possible. All information about your motherboard can be found in the manual that came with it upon purchase. If the manual is lost, you need to go to the official website of your motherboard. Also, the article "How to find out the model and all information about your motherboard" will be useful to youMost often nowadays there are motherboards that support the modes of operation of RAM described below. Dual Mode (most common) - upon closer examination of the motherboard, you can see that the RAM slots are colored in different colors. This is done on purpose and means that the motherboard supports a two-channel RAM mode. That is, two RAM modules with the same characteristics (frequency, timings) and the same volume are specially selected and installed in the RAM slots of the same color.
If your computer has one RAM bar, but the motherboard supports dual-channel mode, you can buy a RAM bar with the same frequency and volume and install both strips in the same color DIMM slots.
Is there an advantage of the two-channel mode over the single-channel
During normal work on a computer, you will not notice the difference, but when working in applications that actively use RAM, for example, Adobe Premiere Pro (video editing), (Canopus) ProCoder (video encoding), Photoshop (working with images), games, the difference can be feel.
Note: Some motherboards will work in dual channel mode even if you install different memory modules in the same color DIMM slots. For example, you will install a 512MB module in the first DIMM slot, and a 1GB bracket in the third slot. The motherboard activates the dual-channel mode for the entire volume of the first strip 512MB, and for the second strip (which is interesting) also 512MB, and the remaining 512MB of the second strip will work in single-channel mode.
How do I know if my RAM is working in dual channel mode or not? Download the free CPU-Z program and go to the Memory tab, look at the Channel parameter in our case - Dual, which means the RAM works in dual-channel mode. If the Channels parameter is Single, then the RAM works in single-channel mode.
In my opinion, during normal work on a computer they will work the same way, personally I did not notice much difference. I worked for a long time on a computer with one large stick of RAM and the performance was the same as on exactly the same computer with two sticks of RAM working in dual-channel mode. A survey of friends and acquaintances of system administrators strengthened me in this opinion. But when working with programs that actively use RAM, for example, Adobe Premiere Pro, Canopus ProCoder, Photoshop, games, a computer with two sticks of RAM will work faster.
Of course you can, but not desirable. The computer will work more stable if it implements the mode of operation of the RAM, which is recommended in the passport of the motherboard. For example, two-channel mode.
Server RAM has its own special properties, which are not possessed by an ordinary RAM strip. Many users have tried to use server RAM for gaming purposes, and why they did not succeed, it will be clear later.
DDR3 RAM for regular computers
At the moment, most users use DDR3 RAM in their computers. Due to its frequencies, it is able to support work with both simple programs and applications, and with game tasks.
Motherboards working with DDR3 memory strips have no more than four RAM slots on their motherboards. Many motherboards support dual-channel streaming, that is, they are used in pairs. To improve the performance of your computer, you can add more strips, but their total amount of memory should not exceed the 64 gigabyte mark.
You can find strips of RAM in the form of a simple board and in the form of a strip with a protective case. There is a significant difference between the two types of planks. For example, a simple bar without all the additions does not have adequate cooling, therefore it has no overclocking potential. Another thing is strips protected from all sides by radiators. This is done so that when overclocking and increasing the timing, the bar is not heated, which has a positive effect on performance.
The ability to overclock RAM still depends on the capabilities of the motherboard. If the BIOS has an XMP profile, then the frequencies of the RAM can easily be overclocked from base to maximum frequencies.
DDR3 RAM for server hardware
During the creation of a modern server platform, you will need not only a large amount of memory for storing information, but also the possibility of fault tolerance and efficiency of all equipment. To increase the processing speed, you can simply upgrade the server RAM, while reducing material costs for equipment.
There are several for servers:
- Parity RAM.
- Registered random access memory.
- Low voltage RAM.
- LR DIMM - RAM that is capable of storing data from large-scale servers.
However, these types of server RAM are not compatible with each other.
The criteria for choosing equipment for creating a server must meet three points - high performance, reliability and the ability to purchase on the market. As for the material costs of maintaining server equipment, you need to take into account the total size of the RAM memory, because the more energy it consumes, the more expensive it is to maintain the server.
An interesting fact, but server memory is tested under high loads, while the temperature of the bar is exceeded by up to 100 degrees Celsius. If 2 bars out of 10 thousand have not passed the test, then the entire series is rejected.
The difference between ordinary RAM and server memory
The first and striking difference between ordinary RAM and server RAM is the presence of a buffering chip. This microcircuit performs a protective role when the memory controller accesses several RAM strips at the same time, since the intensive load of the processor leads to an increase in the current load. This is only hardware support.
As for software protection, it happens that any powerful electromagnetic radiation, or rather, its neurons, can damage the bitness of the server DDR3 RAM. To avoid the consequences of such phenomena, manufacturers have added an error code correction (ECC) function, which can be found on some standard-purpose rails. The algorithm of this function processes the code digitally, calculating errors on its own and correcting them.
In order not to confuse server memory with regular memory when purchasing, just look at the barcode sticker. If the literal value "REG" or "R" is present, then this is an explicit labeling of the server RAM. But the presence of ECC marking does not have to be server-side, since error correction control can also be in ordinary RAM. These are the main criteria for how the server room differs from the usual one.
Motherboard specifications for working with servers
The first thing that catches your eye when considering a server motherboard for DDR3 Server Memory is the sockets for multiple processors. Also on the board there are many connectors for RAM. There are many connectors for connecting the rest of the server equipment, although many modern cards have the ability to wirelessly connect server components.
Server motherboards have a special chipset tuned for computation and calculation. It is its presence that determines whether the motherboard will be able to support work with registers and ECC memory bars. The motherboard does most of the work with the three inherent parts of the server — RAM, hard drive, and processors.
The popular Kingston company presents 17 types of DDR4 server ramps. The memory sizes of each strip vary from 4 to 32 gigabytes. These strips can be purchased separately, unlike the "Whales" - kits, which are produced in four pieces and have a total amount of RAM from 16 to 128 gigabytes. The last option is the most advantageous, since server creators do not have to worry about compatibility and overall power consumption.
Can regular RAM be used in servers and vice versa
Many users, seeing what kind of fault tolerance the server RAM strips have, want to install such RAM on their game board. This misconception drives gamers to a standstill, so here are several arguments to refute rumors that server RAM may exist in a gaming computer.
Despite all its power and reliability, server RAM has completely different parameters for working with numbers and calculations.
It is not for nothing that two or even four processors are installed on the server motherboard, since data transfer requires additional cycles from the processor to work. Also, when exchanging data, RAM uses ECC protocols, which significantly slows down the RAM.
