What happens if the block. Computer power supply: normal operation and malfunctions. ⇡ What power is needed for modern gaming PCs

16.04.2020

A reward is given for finding a Bitcoin block

In May 2017, the Bitcoin network faced a major challenge. The number of unconfirmed transactions reached 200 thousand, and the total volume of raw data exceeded 120 MB. Considering that 1 block in the Bitcoin network is equal to 1 MB, and the average time of its creation is about 10 minutes, the queue of 120 blocks lasted for several days, as new and new unconfirmed transactions constantly arrived.

By increasing transfer fees, it was possible to temporarily reduce the number of unprocessed transactions in the queue, but this measure, of course, could not be considered sustainable. And it is even more surprising that miners from time to time find and close empty blocks, that is, instead of completely filling them to 1 MB, or 4-5 thousand transactions, the block does not contain any transaction-related information.

At some point, the number of empty blocks reached a quarter of all blocks generated by the system, and they continued to be created even when the mempool was overloaded with tens of thousands of unconfirmed transactions.

According to statistics provided by Bitfury, at the end of 2015, more than two hundred empty blocks were generated every month; by the end of 2016, their number dropped to several dozen. The improvements are due to improvements in the architecture, which made it possible to increase the speed of transaction processing, but empty blocks still continue to be created.

Bitcoin empty block statistics

What's the matter here? Let's try to figure it out.

How is a Bitcoin block created?

Each new block is a chain element, which contains a set of records of completed operations on the network that are new from the point of view of the previous chain. A new block is added to the end of the blockchain, it also contains information about previous state chain, and any further changes to its structure are impossible.

That is, a continuous chain of blocks is a kind of accounting book where all operations that have ever been performed in the system are recorded. Any user must be sure that the accounting system is not tampered with. How is such confidence formed?

The block structure includes a header - a personal solution for the block, and miners search for it. They take information from the block and begin to process it, performing certain mathematical operations in order to ultimately obtain a short sequence of letters and numbers that meets predetermined properties. This sequence is called a hash.

Miners mine bitcoins

In order for a block to be written into the blockchain chain, it is necessary to find a special hash parameter whose indicator is lower than a predetermined value. Until the miner finds this parameter by random search, the block is in progress.

If the miner finally solves the problem, then he informs the entire network that he has received a new block. The found block is verified by full nodes of the network, and after verification is included in the blockchain. To “adjust” the processing speed to the increase in power of the entire computer network Every 2016 blocks the difficulty is recalculated so that the time to find a new block is approximately 10 minutes.

This is what creating a new block looks like. The hash of the last block found during the recalculation process becomes a kind of “seal”, that is, it seals the block and confirms the reliability of the entire preceding chain. If someone tries to carry out a fictitious transaction by changing one of the blocks, then its hash will change, and anyone who recalculates the hash of this block will immediately detect the fake.

Now let's briefly describe the structure of the block.

Bitcoin block structure

The block consists of a header and a list of operations.

The header, as we already know, contains a hash (created using the SHA-256 algorithm); it also includes the hash property of the previous block, which creates continuous continuity between network blocks, a list of hashes of operations, block size, etc.

A special place is occupied by the Bits parameter - a shortened version of the hash value. A block will be added to the chain only when miners find a hash less than bits in size.

So, the header is unique and protects the block from forgery. The block is filled with a list of transactions, each of which shows the source and recipient of the transfer.

The recipient is identified using a public key, and a new transaction is created using money confirmed in one of the previous transactions. Used to confirm ownership digital signature, which verifies absolutely every transaction on the network.

Of course, the structure of the network looks complex, especially for a beginner, but as you delve into the essence of its work, the creative genius of its creator begins to emerge, solving the problem of a security flaw for the first time in history. Bitcoin cannot be copied or used twice, and the likelihood of an attack on the network tends to zero, since the attacker must have at his disposal the power of most of the network nodes, which becomes extremely difficult given the decentralized nature of the network.

So, we come to the most important thing. How is the miner’s work structured and what does he get paid for?

Block size and miner reward

If the system as a whole pays for performing certain actions, then pools will perform those actions in order to get paid. This mechanism looks like this.

The miner (mining pool) receives payment for the work performed from two sources:

Firstly, this is the reward for finding a new block, which is this moment is 12.5 BTC (the reward will be halved in 2020).
Secondly, once a miner finds a new block, he automatically receives payment for all transactions that are included in that block.

At the dawn of Bitcoin’s development, blocks were far from completely filled, often containing less than 10 transactions, but as the network’s popularity grew, block fill also began to increase, which led to an increase in the queue of unprocessed transactions. To increase the speed of transactions, they began to apply increased commissions, which led to another problem - the inability to use Bitcoin for small payments.

Many options have been proposed to solve this problem, from increasing blocks to creating higher-level protocols on top of the Bitcoin protocol. Until recently, developers tended to use a modified Segregated Witness (SegWit) protocol, called Segwit2x. With its help, some of the information was supposed to be moved outside the block, that is, stored separately from the blockchain chain, and the size of the block itself was to increase to 2 MB, which theoretically made it possible to significantly speed up transactions and increase anonymity.

However, the hard fork planned for November 16 did not take place, since after the publication of its code the community was unable to reach a consensus.

Where do empty blocks come from?

The miner, as logic dictates, should strive to include the maximum number of transactions in a new block, since in this case his income increases. It is even more surprising to see empty blocks created during mining. Where do they come from?

Let's assume that the miner has found the hash of the next block, let's call it N. Then he must immediately begin searching for block N+1, so as not to have idle capacity. At the same time, the miner must transmit block N to other network participants, who must download it and verify the transactions included in the block. Accordingly, the miner at this moment solves two tasks simultaneously - checking transactions of block N and searching for block N+1.

If a miner finds block N+1 before block N is verified, does he have the right to fill it with transactions? No, it doesn't. After all, these new transactions may include those that rely on transactions included in block N, which has not yet been confirmed. Even if the mempool has accumulated a queue of a large number of unconfirmed transactions that need to be included in block N+1, the miner cannot do this until block N has been confirmed. And if so, then the miner closes block N+1 empty, it will contain only one coinbase transaction, which is generated automatically and carries information about the reward for creating a block. Receives a reward and begins searching for block N+2.

This is where empty blocks come from - this is how the blockchain algorithm works. Empty blocks are obtained due to a mismatch between the speeds of confirming blocks and searching for the next ones, so work to improve the network architecture does not stop for a moment.

Solution

So, the main problem that leads to the creation of empty blocks is the speed of information exchange. Each new block must be “presented” by the pool to other full network nodes, which, in turn, must download it for themselves, and the download speed is different for everyone, and then check all transactions in this block. All these operations take time.

At the time of writing, the number of unconfirmed transactions exceeded 160 thousand, and the volume of raw data was 117 MB.

In 2018, it is planned to introduce several technological solutions that can relieve the Bitcoin network and increase the speed of transactions.

A common misconception most of us have is that the system unit is protected from all sides, and therefore there is no need to worry about its safety. In fact, if we compare the structure of a computer, the screen is the eyes, and the “system unit” is the brain. That is why you need to behave as correctly as possible with this part of the structure; this is the only way the equipment will last a long time.

Why you can’t place the system unit on the floor without a stand:

Large amount of dust. Largest quantity accumulations of dust - on the floor. It settles on nearby parts, tables and forms an imperceptible haze on the wallpaper. But in any case, the dust settles to a greater extent on the floor. The system unit contains fans that are responsible for stabilizing the temperature of the blocks, motherboard and video cards. If you place it directly on the floor, then all the dust in even greater quantities will settle on the fan blades, which in the future will cause the fan to stop and some structural element to burn out.
Smooth surface. To ensure stability of the system unit, you need to place it on a perfectly flat surface. Unfortunately, 80% of all floor coverings have certain irregularities, and therefore it is impossible to guarantee stability without delivery.
Temperature changes. The system unit should not be exposed to constant temperature changes. If you place it on a windowsill or near a battery, you cannot count on the equipment to last long. Floors are capable of different time accumulate heat, moisture, and cold throughout the year.
Mechanical damage. Any scratch on the surface of the block is a potential threat of corrosion, and therefore you should be more careful about where you place the processor. You cannot place it near the aisle, in a place where there is a risk of damaging it or overturning it. Particular attention should be paid to children's rooms. It is better to install the computer near the wall, but not close to it, so that condensation does not form.

These are the main reasons why programmers do not recommend placing the computer unit directly on the floor without a stand. But there are other common PC user errors - shocks, mechanical damage, exposure to moisture, accumulation of dampness on systems. All this contributes to the fact that after a short time of use, the computer breaks down and has to be repaired or replaced.

Microchips system unit are very susceptible to static, and therefore placing equipment near sources of static will result in failure. Also, you should not install the device in your cat’s favorite resting place, and you should not allow her to sleep near the computer.

Where to put?

The first thing that comes to mind when placing a system unit is to buy a table with special stands. What if there is already a table and there is no desire to change it? What to do in this case? In this situation, there are special stands for the system unit, which are universal in their application, easy to use and not expensive.

The main advantage of the stand is its maneuverability. The wooden base can be placed anywhere under the table; it will not interfere with work, and if necessary, you can easily change its location.

Stand for computer system unit

The universal and only practical option for arranging a workplace with a table that does not have a stand or space to place the processor is the Barsky wooden stand. Externally, it is a simple H-shaped design. But despite its simplicity, it will make your life at your desk incredibly easier. Advantages of using a stand for the system unit:

installed exactly relative to the surface;
the system unit is secured using side borders;
you can change the location of the processor: to the left or right side, forward or move back to the wall;
dust accumulates under the wooden base of the bottom, and not on the processor itself;
portable and does not require fastening to the base of the table, which does not contribute to deformation of the main structure;
Light natural wood without chemical impregnations will fit into any room interior.

The main task of such a stand is to ensure the stability of the block and protect it from the accumulation of dampness from the floor surface.

How to determine dimensions

System units differ not only in memory size, but also in external parameters: some are smaller, others are larger. How, then, to determine the required dimensions of the stand? A special addition to the computer desk, the Barsky stand is universal. Its dimensions allow you to place both large devices and non-standard system units: width-depth-height - 540x270x120 mm.

Near the side there is an opportunity to put a carrier or install a tee for connecting from the network. This helps to organize correctly workplace at home or in the office.

Barsky offers

A black and white stand for a computer system unit from Barsky is a combination of style, simplicity and harmony. It can be installed in any convenient place, which is important for left-handers (often you need to adapt to furniture designs designed for right-handers). A durable wooden stand with ideal shapes will help organize your workplace as conveniently and correctly as possible, and the black and white colors will suit any table color scheme.

I was prompted to write this article by constant questions about the materials in the “” section, which quite often begin with the word “ Why». Why is a power supply recommended in such and such an assembly?N watt? Why do you offer such expensive solutions when you can save a lot? Why is a one-kilowatt power supply recommended for an extreme build? This is just a small list of questions that came to mind immediately when I started writing this article. Indeed, users who do not yet have the proper experience in assembling and assembling system units want to know the exact and obvious criteria for choosing the “breadwinner” of the entire PC. In addition, the choice of power supplies on our market is very, very wide. Thus, at the time of writing this article, the Regard store website listed 676 models of computer power supplies - fewer central processors are sold. Therefore, it is necessary to help beginners understand this issue.

It is important to note that in this article I will not be recommending any specific power supply models. For these purposes on our website periodically. IN this material The features of modern power supply models will be considered, as well as the criteria and formats of modern PC platforms that allow you to assemble a full-fledged gaming system.

⇡ How the energy consumption of gaming components has changed

Before starting to analyze the main and secondary parameters of any computer unit power supply, in my opinion, it is necessary to understand which PC components affect the level of energy consumption. More precisely, it is clear that the central processor and discrete video card are Stakhanovites in this matter, but how much does this hardware affect power consumption?

Let's keep it simple. The graphs below show the parameters of all processors and video cards that the 3DNews laboratory has tested over the past five years and which, in the opinion of the author of this material, can at least conditionally be classified as gaming solutions (taking into account their relevance in a certain period of time, of course). In this case, we are talking about such a parameter as TDP - design thermal power. The fact is that many people associate this value with energy consumption.

Intel believes that thermal design power (TDP) is a parameter that “ indicates the average performance in watts when processor power is dissipated (running at base frequency with all cores engaged) under a challenging workload definedIntel" We see that the TDP level of modern - and not so modern - central processors varies over a fairly wide range. The statistics I collected indicate chips with estimated power from 35 and up to 250 W, respectively. If we look at the most popular devices of their time, we will see that gaming computers are mostly equipped with chips with a TDP in the range from 65 to 105 W.

And here we immediately see a certain catch. Undoubtedly, the central processor and video card are the main consumers of energy in any computer system. At first glance, it may seem that choosing a power supply with the required power is very simple: add up the TDP of the processor with the TDP of the graphics accelerator, plus take into account that any system unit also contains other components (drives, motherboard and hardware with fans). Only, using Intel's definition, we see that the thermal design power is the average performance value in watts when the CPU is running at base frequency. Quite often you can encounter operating scenarios where the central processor for a desktop PC goes beyond the level specified by the manufacturer. In general, TDP is not an indicator of the actual power consumption level of a particular component.

Let me give you a simple example. Above is a screenshot that clearly demonstrates how the central processor works under load in the form of the Prime95 program. According to technical specifications, the base frequency of this 6-core chip is 2.8 GHz and the rated power is 65 W. Only in a program that uses AVX instructions, all cores operate at a frequency of 3.8 GHz - this is how Turbo Boost technology works. Our measurements showed that the processor consumes more than 95 W, which means it is clearly outside the limits defined by Intel in the specification. It turns out that in many boards, the MultiCore Enhancements function, which is responsible for the operation of the CPU within TDP, is enabled by default - therefore, restrictions on maximum power consumption are removed.

And we just recently learned that at a similar TDP level - 65 W - it works in a similar way. , the chip frequency varies in the range from 4.1 to 4.4 GHz with a base value of 3.6 GHz. Naturally, we are not talking about any 65 W: under serious load, the processor sets a completely different power consumption bar - 100+ W. Again, we are talking about the system operating in the default mode, without manual overclocking or increasing the voltage, that is, the manufacturer specifically makes sure that the actual power consumption significantly exceeds the declared TDP level. As you can see, both chipmakers have been acting similarly lately.

A similar situation is observed among video cards. Here is the most productive gaming model to date, the GeForce RTX 2080 Ti, with a stated TDP of 260 W at maximum load.

This is the catch. You can’t just take and add up the calculated power of the main components of the system. Thus, the sum of the TDP of the Core i9-9900K and GeForce RTX 2080 Ti is 345 W. Some more will be “eaten up” by other components of the system. However, looking ahead, I will say that I managed to load the system so that it consumed more than 450 W.

And don't forget about overclocking. You can judge its benefits from the point of view, for example, of obtaining additional FPS in games from our reviews - 3DNews does not miss interesting and popular models of central processors and video cards. But you will find out how the system’s power consumption changes after overclocking in the second part of the article.

The phrase “other system components” naturally refers to hardware such as a motherboard, RAM, other discrete devices (in addition to the video card), as well as components of cooling systems (cooler and case fans, cooling system pump, and so on). But practice shows that all the listed components do not consume very much - compared to the same processors and video cards.

*The graph above shows the power consumption of the entire system (described below), not just the RAM

Let's look at RAM. Unfortunately, I do not know of a method that will accurately measure the power consumption of individual RAM modules. So I took two Samsung M378A1G43EB-CRC modules with a total capacity of 16 GB and installed them in a system with a Ryzen 5 1600 processor and motherboard. We know that this kit can comfortably overclock to 3200 MHz while maintaining latency but increasing voltage slightly. For the load, I used the Prime95 29.8 program with the Large FFT test enabled, which loads the RAM to the maximum. Well, the difference between DDR4-2400 and DDR4-3200 was only 14 W when comparing peak power consumption.

There is no particular point in measuring the power consumption of drives, because compared to the same processors and video cards, it is extremely small. For example, a review was published on our website hard drives with a capacity of 14-16 TB - that these monsters in read mode do not consume more than 9.5 W, and yet such drives have 7-9 plates installed. It turns out that only a combination of several HDD/SSD can seriously affect the power consumption of a PC, and even then we must take into account that storage devices must work simultaneously, and this is not very typical for desktops. Typically, when it comes to a home PC, the system uses 1-2 SSDs and the same number of mechanical drives.

The situation with energy consumption is approximately the same for fans - parameters such as current, voltage and power are often indicated on their housing. Standard desktop-grade impellers rarely draw more than 5 watts. Typically, the system uses 3-4 case fans and one or two Carlson fans, which are included with the processor cooling. It turns out that even installing six impellers will increase the power consumption of the system unit by only 20-25 W.

Strictly speaking, we come to where we started. The main energy consumption in any system unit falls on the central processor and video card. We have already found out that you cannot trust the specifications of the CPU and GPU and choosing a unit based on the sum of the TDP components is not the best idea. We will tell you how to understand which block is needed in the second part.

All of the above allows us to draw another conclusion: we see that energy consumption computer equipment does not change much from year to year and is within certain limits. That is, the power supply purchased now will last a long time and will be useful when assembling the next system, or maybe two. In this vein, buying a known good power supply looks like a very rational idea.

⇡ About cable management of the system unit

Continuing the topic of choosing a power supply of a certain power, we definitely need to talk about cable management in modern PCs. The fact is that one important rule works here: the greater the power of the power supply, the more cables it has. If we talk about gaming systems, then in modern realities the power source may require at least two wires that will be connected to the motherboard. On average, four to five cables are used. But power supplies often have much more of them.

Let's start with video cards, because in most gaming PCs they require the most electricity. As you know, slot PCI Express x16 motherboard capable of transmitting up to 75 W of electricity to a discrete device (actually a little more, but the standard describes exactly this value). For example, such power supply is sufficient for most video cards of the GeForce GTX 1650 level, which can safely be classified as gaming. But on more powerful video cards you can often find 6- and 8-pin power connectors. In the first case, up to 75 W of energy is transmitted, in the second - up to 150 W.

Mid-price video cards (with a TDP no higher than 200 W) are usually equipped with one 6- or 8-pin connector. More powerful video cards usually have a pair of connectors.

Continuing the topic of cable management, we can say with confidence that in some cases other power supply cables may not be needed at all. For example, if you use M.2 form factor drives in your system and do not install various peripherals (for example, optical drive). In this case, you will only need to power the motherboard and video card from the power supply. NVMe SSDs, installed on boards and not requiring additional connectors, have long been recommended in most “Computer of the Month” builds.

However, any power supply will provide connection to at least four SATA devices. The kit also includes MOLEX wires, which are rarely used today. In cheap cases, fans can be powered from them, for example. In principle, video cards can also be powered through adapters from MOLEX (but I categorically do not recommend doing this in the case of expensive 3D accelerators!).

In particularly advanced cases, when it is necessary to connect a large number of wires, it is better to take a partially or fully modular power supply. This approach will make life much easier when assembling the system. It’s funny, but if only three or four wires are required from the power supply, then in this case it is also better to use a device with modular cable management - so that the extra “tail” does not stick out and get in the way.

And yet, aesthetically, assembling a system with a non-modular power supply is not a tragedy. Excess wires are easily hidden under the hard drive cage. And now even the most inexpensive cases are equipped with a curtain (metal or plastic) on the bottom. Behind it are hidden both the power supply itself and a bunch of unused cords.

A fully modular power supply will be needed if you want to not only build a neat PC, but do it beautifully - using a braid, for example. Corsair also sells sets of braided wires, but you can make the braid yourself.

A small announcement: I will tell (and show) more detail about cable management in another article, which will soon be published on our website.

Cable length is another important operational parameter of any power supply. Of course, a lot here depends on the computer case. But for most Midi-Tower models with a height of 400 to 500 mm with a bottom-mounted power supply, it is enough for the 4/8-pin CPU power cable to be 500-550 mm long. For Full/Ultra Tower with a height of 600-800 mm, a minimum of 600 mm is needed. It turns out to be a fairly simple rule: The length of the EPS cord should be equal to the height of the case if we are talking about the bottom location of the power supply. Then there will be no surprises during assembly. In general, the length of other power supply cables in the case of tower cases is of little interest to us. In some models, the length of the cord with a 24-pin port reaches 700 mm - in this case, it is even more problematic to properly lay it behind the chassis of the case.

The attentive reader probably noticed that I in no way touched upon the form factor of the power supply units themselves - they are different, sometimes a computer case. But this article is tied to the “Computer of the Month” section, and it recommends assemblies in classic tower cases. I promise that I will devote a separate detailed article to assembling compact gaming PCs.

Still, before purchasing, make sure that your power supply fits the length of the case. For example, the previously listed Corsair PSU models will fit 99% of Midi-Tower cases. But for some Corsair AX1200i with a length of 225 mm (and the connected wires will also take up 50-100 mm), you will have to look for a more spacious computer “home”.

⇡ How much does a new power supply cost?

I will be brief in this paragraph. Quite often in the comments to “Computer of the Month” or to any other article related to power supplies, you see a message in the style of “ Why is such a power supply here? Here there is enough model forNW" On the one hand, such commentators are right. On the other hand, the table below clearly shows that a lower-power power supply does not always cost significantly less than a model with a higher number of declared watts. This rule is especially true for models with a power of 400-600 W.

Cost of ATX form factor power supplies, rub.
		400-450 W	500-550 W	600-650 W	700-750 W	800-850 W	1000-1050 W
80 PLUS	Min.	2 850	2 940	3 560	3 850	No current models
	Max.	2 940	3 380	3 760	4 260
	Average	2 900	3 163	3 600	4 073
80 PLUS Bronze	Min.	3 090	3 420	4 500	4 800	7 080	No current models
	Max.	4 850	5 870	6 540	7 670	7 460
	Average	4 206	4 896	5 849	6 300	7 200
80 PLUS Silver	Min.	There are only two models in the store
	Max.
	Average
80PLUS Gold	Min.	4 270	5 380	5 850	6 370	8 140	8 250
	Max.	6 190	10 850	10 760	12 270	1 3460	17 530
	Average	5 280	7 547	7 780	8 636	10 560	12 738
80 PLUS Platinum	Min.	No current models	8 840	10 930	10 800	12 440	12 470
	Max.		11 250	13 420	15 420	17 620	20 860
	Average		10 500	12 392	13 255	14 088	15 653
80 PLUS Titanium	Min.	No current models		15 560	17 700	17 870	19 690
	Max.			19 900	18 750	20 230	25 540
	Average			17 730	18 215	19 050	22 615

We see that more powerful devices of a similar class (for example, those with an 80 PLUS Bronze certificate) cost quite a bit more, if at all. Comparing average prices, we see that the difference between power supplies with a power of 400-450 W and 500-550 W is a little more than 600 rubles. In this situation, it is definitely worth paying this amount, but getting a more powerful device in return. The difference in price between units with a power of 600-650 and 700-750 W is even less.

And looking at the table, you can make quite a large number of such comparisons. Therefore, another question arises: if it is possible to buy a power supply for the same or a slightly larger amount more power, then why not use it? The question, however, is rhetorical.

To collect statistics, I went to the Regard store website, selected six popular manufacturers and calculated the average cost of power supplies of a certain power and a certain 80 PLUS standard.

⇡ Methodology and stand

Today's testing used a large amount of computer hardware to show how much power real-life gaming systems consume. In this regard, I relied on the collections of the “Computer of the Month” section. A complete list of all components is given in the table below.

Test bench, software and auxiliary equipment
CPU	Intel Core i9-9900K Intel Core i7-9700K Intel Core i5-9600K Intel Core i5-9500F AMD Ryzen 5 1600 AMD Ryzen 5 2600X AMD Ryzen 7 2700X
Cooling	NZXT KRAKEN X62
Motherboard	ASUS ROG MAXIMUS XI FORMULA ASUS ROG STRIX B450-I GAMING
RAM	G.Skill Trident Z F4-3200C14D-32GTZ, DDR4-3200, 32 GB Samsung M378A1G43EB-CRC, DDR4-2400, 16 GB
Video card	2 × ASUS ROG Strix GeForce RTX 2080 Ti OC ASUS Radeon VII ASUS DUAL-RTX2070-O8G NVIDIA GeForce RTX 2060 Founders Edition ASUS ROG-STRIX-RX570-4G-GAMING AMD Radeon RX Vega 64 ASUS PH-GTX1660-6G
Storage device	Samsung 970 PRO MZ-V7P1T0BW
power unit	Corsair CX450 Corsair CX650 Corsair TX650M Corsair RM850x Corsair AX1000
Frame	Open test bench
Monitor	NEC EA244UHD
operating system	Windows 10 Pro x64 1903
Software for video cards
NVIDIA	431.60
AMD	19.07.2005
Additional software
Removing drivers	Display Driver Uninstaller 17.0.6.1
FPS measurement	Fraps 3.5.99
	FRAFS Bench Viewer
	Action! 2.8.2
Overclocking and monitoring	GPU-Z 1.19.0
Overclocking and monitoring	MSI Afterburner 4.6.0
Optional equipment
Thermal imager	Fluke Ti400
Sound level meter	Mastech MS6708
Wattmeter	watts up? PRO

The test benches were loaded with the following software:

Prime95 29.8— Small FFT test, which loads the central processor as much as possible. A very resource-intensive application, in most cases programs that use all cores are not able to load the chips more.
AdobePremierPro 2019— 4K video rendering using the CPU. An example of resource-intensive software that uses all processor cores, as well as available reserves random access memory and storage.
"The Witcher 3: Wild Hunt"— testing was carried out in full screen mode in 4K resolution using maximum settings graphics quality. This game heavily loads not only the video card (even two RTX 2080 Ti in the SLI array are 95% loaded), but also the central processor. As a result, the system unit is loaded more than, for example, using FurMark “synthetics”.
"The Witcher 3: Wild Hunt" +Prime95 29.8(Small FFT test) - a test for maximum system power consumption when both the CPU and GPU are 100% loaded. And yet it should not be ruled out that there are more resource-intensive connections.

Energy consumption was measured using a watts up? PRO - despite such a comical name, the device can be connected to a computer, and with the help of special software it allows you to monitor its various parameters. So, the graphs below will show the average and maximum energy consumption levels of the entire system.

The period of each power measurement was 10 minutes.

⇡ What power is needed for modern gaming PCs

Let me note again: this article is to a certain extent tied to the “Computer of the Month” section. Therefore, if you are visiting us for the first time, I recommend that you at least familiarize yourself with. Each “Computer of the Month” covers six assemblies, mostly gaming ones. I used similar systems for this article. Let's get acquainted:

The combination of Ryzen 5 1600 + Radeon RX 570 + 16 GB of RAM is an analogue of the starting assembly (35,000-37,000 rubles per system unit, excluding the cost of software).
The combination of Ryzen 5 2600X + GeForce GTX 1660 + 16 GB of RAM is an analogue of the basic assembly (50,000-55,000 rubles).
A combination of Core i5-9500F + GeForce RTX 2060 + 16 GB of RAM is an analogue of the optimal assembly (70,000-75,000 rubles).
A combination of Core i5-9600K + GeForce RTX 2060 + 16 GB of RAM is another option for an optimal build.
A combination of Ryzen 7 2700X + GeForce RTX 2070 + 16 GB of RAM is an analogue of an advanced build (100,000 rubles).
The Ryzen 7 2700X + Radeon VII + 32 GB RAM combination is similar to the maximum build (130,000-140,000 rubles).
A combination of Core i7-9700K + Radeon VII + 32 GB of RAM is another option for the maximum build.
A combination of Core i9-9900K + GeForce RTX 2080 Ti + 32 GB of RAM is an analogue of an extreme build (220,000-235,000 rubles).

Unfortunately, I was not able to get Ryzen 3000 processors at the time of conducting all the tests, but the results obtained will not become less useful. The same Ryzen 9 3900X consumes less than the Core i9-9900K - it turns out that within the framework of an extreme build, studying the power consumption of an 8-core Intel processor will be even more interesting and important.

And also, as you may have noticed, the article uses only mainstream platforms, namely AMD AM4 and Intel LGA1151-v2. I did not use HEDT systems such as TR4 and LGA2066. Firstly, we abandoned them in the “Computer of the Month” a long time ago. Secondly, with the advent of the 12-core Ryzen 9 3900X in the mass segment and in anticipation of the imminent release of the 16-core Ryzen 9 3950X, such systems have become extremely highly specialized. Thirdly, because the Core i9-9900K still gives everyone a run for their money in terms of power consumption, once again proving that the calculated thermal power declared by the manufacturer tells the consumer little.

Now let's move on to the test results.

To be honest, I present the results of testing in programs such as Prime95 and Adobe Premier Pro 2019 more for information purposes - for those who do not play or use discrete video cards. You can safely rely on these data. Basically, here we are interested in the behavior of test systems under loads close to maximum.

And here some very interesting things are observed. In general, we see that all the systems considered do not consume very much energy. The most voracious, which is quite logical, was the system with Core i9-9900K and GeForce RTX 2080 Ti, but even it in stock (read - without overclocking) consumes 338 W when it comes to games, and 468 W at maximum PC load. It turns out that such a system will have enough power supply for an honest 500 W. It is so?

⇡ It's not just about watts

It would seem that we can end the article here: recommend to everyone a power supply with a capacity of 500 honest watts - and live in peace. However, let's conduct a few additional experiments to get a complete picture of what is happening with your PC.

In the screenshot above we see that the power supplies operate most efficiently at 50% load, that is, half of the declared power. It may seem to some that the difference between a device with a basic 80 PLUS certificate with a peak efficiency of about 85% on a 230 V network and, say, a “platinum” power supply with an efficiency of about 94% is not so great, but this is misleading. my colleague Dmitry Vasiliev points out quite accurately: “An energy source with an efficiency of 85% wastes 15% of its power on heating the surrounding air, while a “breadwinner” with an efficiency of 94% converts only 6% of its power into heat. It turns out that the difference is not “ some there"10%, but x2.5." Obviously, in such conditions, a more efficient power supply is quieter (there is no point in the manufacturer setting the device’s fan to maximum speed) and heats up less.

And here is the proof of the above words.

The graphs above show the efficiency of some power supplies participating in the tests, as well as the rotation speed of their fans at different load levels. Unfortunately, the equipment used does not allow us to accurately measure the noise level, but by the number of revolutions per minute of the built-in fans we can judge how noisy the power supply will be. It is imperative to note here that this does not mean at all that the power supply will stand out “from the crowd” under load. Still, usually the noisiest components of a gaming computer are the processor cooler and video card.

Practice, as you see, agrees with theory. Power supplies really operate at their most efficient at about 50 percent load. Moreover, in this regard, I would like to note the Corsair AX1000 model - this power supply reaches its peak efficiency at a power of 300 W, and then its efficiency does not fall below 92%. But other Corsair blocks on the graphs have a completely expected “hump”.

At the same time, Corsair AX1000 can operate in semi-passive mode. Only under a load of 400 W does its fan begin to spin up at a frequency of ~750 rpm. The RM850x has the same characteristic, but in it the impeller begins to rotate at a power of ~200 W.

Now let's look at the temperatures. To do this, I disassembled all the power supplies. The fans from the top cover were removed and installed on a homemade tripod so that the distance between it and the rest of the power supply was approximately 10 cm. I am sure that in terms of cooling the device did not work any worse, but this design allowed me to take pictures with a thermal imager. In the graph above, "Temperature 1" refers to the maximum internal temperature of the power supply when the fan is running. “Temperature 2” is the maximum heating of the power supply... without additional cooling. Please do not repeat such experiments at home on your equipment! However, such a bold move allows you to clearly show how the power supply heats up and how its temperature depends on the rated power, build quality and the component base used.

Heating of the CX450 model to 117 degrees Celsius is a completely logical phenomenon, because this power supply operates at almost maximum with a load of 400 W, and is not cooled in any way. The fact that the power supply passed this test at all is a great sign. Here is a high-quality budget model.

Comparing the results of other power supplies, we can come to the conclusion that they seem quite logical: yes, the Corsair CX450 model heats up the hottest, and the RM850x the least. At the same time, the difference in maximum heating rates is 42 degrees Celsius.

It is important to define the concept of “honest power” here. Here the Corsair CX450 model can transmit 449 W of energy via a 12-volt line. It is this parameter that you need to look at when choosing a device, because there are models that do not work as efficiently. In cheaper units of similar power, noticeably fewer watts can be transmitted over a 12-volt line. It comes to the point that the manufacturer claims support for 450 W, but in fact we are talking only about 320-360 W. So let’s write it down: when choosing a power supply, you need to look, among other things, at how many watts the device produces via a 12-volt line.

Let's compare the Corsair TX650M and CX650, which have the same claimed power but are certified to different 80PLUS standards: gold and bronze, respectively. I think the thermal imager images attached above speak louder than any words. Really, support for a specific 80 standardPLUS indirectly speaks about the quality of the power supply element base. The higher the certificate class, the better the power supply.

It's important to note here that the Corsair TX650M delivers up to 612 watts over a 12-volt line, while the CX650 delivers up to 648 watts.

In the pictures above you can compare the heating of the RM850x and AX1000 models, but already at a load of 600 W. There is also an obvious difference in temperature here. Overall, we see that Corsair power supplies cope well with the load placed on them - and even in stressful situations. At the same time, I think it is now clear why the graph above did not show the temperature of the AX1000 - it does not heat up much, even if you remove the cover with the fan.

Considering the results obtained, you can see that it would be absolutely no shame to use in the system a power supply with a power twice the maximum power of the PC itself. In this mode of operation, the power supply heats up less and makes less noise - these are facts that we have just proven once again. It turns out that for a starter assembly a power supply with an honest power of 450 W is suitable, for a basic one - 500 W, for optimal - 500 W, for advanced - 600 W, for maximum - 800 W, and for extreme - 1000 W. Plus, in the first part of the article we found out that it’s not so a big difference in price between power supplies, the declared power of which differs by 100-200 W.

However, let's not rush to final conclusions.

⇡ A few words about the upgrade

The builds in the “Computer of the Month” are designed not only to work in default mode. In each issue, I talk about the possibilities of overclocking some components (or the pointlessness of overclocking in the case of some processors, memory and video cards), as well as the possibilities for subsequent upgrades. There is an axiom: the cheaper the system unit, the more compromises it has. There are compromises that will allow you to use a PC here and now, but the desire to get something more productive, quiet, efficient, beautiful or comfortable (underline as necessary) will still not leave you. Captain Obviousness suggests that in such situations, a power supply with a good watt reserve will be very useful.

I will give a clear example of upgrading the starting assembly.

I took the AM4 platform. 6-core Ryzen 5 1600, Radeon RX 570 and 16 GB DDR4-3000 RAM were recommended. Even when using a standard cooler (cooling system that is sold complete with the CPU), our chip can be easily overclocked to 3.8 GHz. Let's say I did something radical and changed the CO to a noticeably more efficient model, which allowed me to increase the frequency from 3.3 to 4.0 GHz when all six cores were loaded. To do this, I needed to raise the voltage to 1.39 V, and also set the fourth level of Load-Line Calibration of the motherboard. This overclock essentially turned my Ryzen 5 1600 into a Ryzen 5 2600X.

Let's say I bought Radeon video card RX Vega 64 - on the Computeruniverse website a month ago you could buy it for 17,000 rubles (excluding delivery), and even cheaper. And in the comments to “Computer of the Month” they talk so sweetly about used GeForce GTX 1080 Ti, sold for 25-30 thousand rubles...

Finally, instead of the Ryzen 5 1600, you can take the Ryzen 2700X, which has become noticeably cheaper since the release of the third-generation AMD family of chips. There is no particular need to overclock it. As a result, we see that in both cases of the upgrade I proposed, the system’s power consumption more than doubled!

This is just an example and characters in the situation described may be completely different. However, this example, in my opinion, clearly shows that even in a starter assembly, a power supply with an honest power of 500 W, or better even 600 W, would not hurt at all.

⇡ “Gaming PCs do not need 1 kW units” - commentators under articles on the site

We often see comments like this when it comes to gaming PCs. In the vast majority of cases - and we have found this out in practice - this is so. However, in 2019 there is a system that can amaze with its energy consumption.

We are, of course, talking about an extreme build in its, so to speak, maximum combat form. Not long ago, an article “” was published on our website - in it we talked in detail about the performance of a couple of the fastest GeForce video cards in 4K and 8K resolution. The system is fast, but the components are selected in such a way that it is very easy to make it even faster. In addition, it turned out that overclocking the Core i9-9900K to 5.2 GHz is completely useful in the case of the GeForce RTX 2080 Ti SLI array and Ultra HD games. Only at its peak, as we see, such an overclocked configuration consumes more than 800 W. Therefore, for such a system under such conditions, a kilowatt power supply will definitely not be superfluous.

⇡ Conclusions

If you carefully read the article, you have identified several main points that you need to keep in mind when choosing a power supply. Let's list them all again:

Unfortunately, it is impossible to rely on the TDP indicators declared by the manufacturer of the video card or processor;
the energy consumption of computer equipment does not change much from year to year and is within certain limits - therefore, a high-quality power supply purchased now will serve for a long time and will definitely come in handy during the assembly of the next system;
the needs for cable management of the system unit also influence the choice of a power supply of a certain power;
Not all power connectors on the motherboard need to be used;
a power supply with a lower power is not always more profitable (in terms of price) than a more powerful model;
when choosing a power supply, you need to look, among other things, at how many watts the device produces via the 12-volt line;
support for a certain 80 PLUS standard indirectly indicates the quality of the element base of the power supply;
There is absolutely no shame in using a power supply whose honest power is twice (or even more) the maximum power consumption of the computer.

Quite often you can hear the phrase: “ More - no less" This very laconic aphorism perfectly describes the situation when choosing a power supply. For your new PC, take a model with a good power reserve - it definitely won’t be worse, and in most cases it will only be better. Even for an inexpensive gaming system unit, which consumes about 220-250 W at maximum load, it still makes sense to take good model with honest 600-650 W. Because this block:

it will work more quietly, and in the case of some models - absolutely silently;
it will be colder;
will be more effective;
will allow you to easily overclock the system, increasing the performance of the central processor, video card and RAM;
will allow you to easily upgrade the main components of the system;
will survive several upgrades, and also (if the power supply is really good) will live in the second or third system unit;
It will also allow you to save money during the subsequent assembly of the system unit.

I think few readers will refuse a good power supply. It is clear that it is not always possible to immediately buy a high-quality device with a large reserve for the future. Sometimes, when buying a new system unit and having a limited budget, you want to get a more powerful processor, faster video cards, and a higher-capacity SSD - all this is understandable. But if you have the opportunity to buy a good power supply with a reserve, there is no need to save on it.

We express our gratitude to the companiesASUS andCorsair, as well as the Regard computer store for the equipment provided for testing.

Measurements of system energy consumption turned out to be quite expected. The most simple system Without a discrete video card, I could probably get by with any compatible power supply at all. We can also see that it is now quite old AMD processor The Phenom II X4 965 shows a decent difference in power consumption when compared to the less demanding Intel Core i7-3770K. However, technically, all four systems would be able to function properly even on a 450 W power supply (of adequate quality with honest watts).

So who will need 1000 W power supplies then? Obviously, they can also find real use, for example, if you have a sophisticated gaming system costing about one hundred thousand rubles with three video cards. Some information storage enthusiasts have the weakness of installing about twenty hard drives with a bunch of additional controllers, but for most ordinary ones, even powerful systems An honest (read: high-quality) 550 W power source will be enough. Office computer without discrete video (or with devices entry level) will probably be able to get by with one of the most low-power devices.

Conclusion

The results we received speak for themselves. Even powerful gaming computer with overclocked components it does not consume more than 360 W at peak. That is, it is obvious that you will not need a kilowatt power supply until you decide to assemble a 3-Way SLI configuration. Of course, the results should not mislead you. They do not mean at all that you can use a 400 W power supply from cases with a total cost of 900 rubles for such an assembly. But in the end, there is no reason to buy a really high-quality 750-1000 W power supply; you can get by with a cheaper and fairly reliable model that will work on your computer - and still with a large margin.

The power supply is the most important component of any personal computer, on which the reliability and stability of your assembly depends. There is quite a large selection of products on the market from various manufacturers. Each of them has two or three lines or more, which also include a dozen models, which seriously confuses buyers. Many people do not pay due attention to this issue, which is why they often overpay for excess power and unnecessary bells and whistles. In this article we will figure out which power supply is best for your PC?

The power supply (hereinafter referred to as PSU) is a device that converts high voltage 220 V from the socket to computer-friendly values and equipped with the necessary set of connectors for connecting components. It seems to be nothing complicated, but upon opening the catalog, the buyer is faced with a huge number various models with a bunch of often incomprehensible characteristics. Before we talk about choice specific models, let’s look at what characteristics are key and what you should pay attention to first.

Main parameters.

1. Form factor. In order for the power supply to fit into your case, you must decide on the form factors, based on from the parameters of the system unit case itself . The dimensions of the power supply in terms of width, height and depth depend on the form factor. Most come in the ATX form factor, for standard cases. In small system units of the microATX, FlexATX, desktops and others, smaller units are installed, such as SFX, Flex-ATX and TFX.

The required form factor is specified in the characteristics of the case, and it is by this that you need to be guided when choosing a power supply.

2. Power. The power determines what components you can install in your computer, and in what quantity.

It is important to know! The number on the power supply is the total power across all of its voltage lines. Since the main consumers of electricity in a computer are the central processor and video card, the main power line is 12 V, when there are also 3.3 V and 5 V to power some components of the motherboard, components in expansion slots, power drives and USB ports. The power consumption of any computer along the 3.3 and 5 V lines is insignificant, so when choosing a power supply for power, you should always look at the "characteristic" power on line 12 V", which ideally should be as close as possible to the total power.

3. Connectors for connecting components, the number and set of which determine whether you can, for example, power a multiprocessor configuration, connect a couple or more video cards, install a dozen hard drives, and so on.

Main connectors except ATX 24 pin, This:

To power the processor, these are 4 pin or 8 pin connectors (the latter can be detachable and have a 4+4 pin entry).

To power the video card - 6 pin or 8 pin connectors (8 pin is most often collapsible and is designated 6+2 pin).

For connecting 15-pin SATA drives

Additional:

4pin MOLEX type for connecting older HDDs with an IDE interface, similar disk drives and various optional components, such as rheobass, fans, etc.

4-pin Floppy - for connecting floppy drives. They are very rare these days, so such connectors most often come in the form of adapters with MOLEX.

Extra options

Additional characteristics are not as critical as the main ones in the question: “Will this power supply work with my PC?”, but they are also key when choosing, because affect the efficiency of the unit, its noise level and ease of connection.

1. Certificate 80 PLUS determines the efficiency of the power supply unit, its efficiency (efficiency factor). List of 80 PLUS certificates:

They can be divided into the basic 80 PLUS, on the far left (white), and the colored 80 PLUS, ranging from Bronze to the top Titanium.

What is efficiency? Let's say we are dealing with a unit whose efficiency is 80% at maximum load. This means that on maximum power The power supply will consume 20% more energy from the outlet, and all this energy will be converted into heat.

Remember one simple rule: the higher the 80 PLUS certificate in the hierarchy, the higher the efficiency, which means it will consume less unnecessary electricity, heat less, and, often, make less noise.

In order to achieve the best efficiency indicator and obtain the 80 PLUS “color” certificate, especially top level, manufacturers use their entire arsenal of technologies, the most efficient circuit design and semiconductor components with the lowest possible losses. Therefore, the 80 PLUS icon on the case also speaks of the high reliability and durability of the power supply, as well as a serious approach to creating the product as a whole.

2. Type of cooling system. The low level of heat generation of power supplies with high efficiency allows the use of silent cooling systems. These are passive (where there is no fan at all), or semi-passive systems, in which the fan does not rotate at low powers, and starts working when the power supply becomes “hot” under load.

When selecting a power supply, you should pay attention to for the length of the cables, the main ATX24 pin and the CPU power cable when installed in a case with a bottom-mounted power supply.

For optimal installation of power wires behind the rear wall, they must be at least 60-65 cm long, depending on the size of the case. Be sure to take this point into account so you don’t have to bother with extension cords later.

You need to pay attention to the number of MOLEX only if you are looking for a replacement for your old and antediluvian system unit with IDE drives and drives, and even in a significant quantity, because even the simplest power supplies have at least a couple of old MOLEX, and in more expensive models There are dozens of them in general.

I hope this little guide to the DNS company directory will help you in this way. complex issue at the initial stage of your acquaintance with power supplies. Enjoy the shopping!