home Magazine Contact promo codes and discount coupons. How to find a svyaznoy promo code and use a svyaznoy promo code

Contact promo codes and discount coupons. How to find a svyaznoy promo code and use a svyaznoy promo code

Daily progress in the field information technologies only speeds up its pace. The volumes and speeds of transmitted data are growing. However, to meet the needs of modern software Do not forget about the improvement and development of the hardware component.

The connector is widely used for data transfer between devices. USB, which appeared in 1996. However, not everyone has the idea that today many modern devices are equipped with the third generation of this connector - USB 3.0. In this article we will try to figure out what changes and improvements the developers have “invested” in the 3.0 generation and what the differences between USB 2.0 and USB 3.0 are.

backward compatibility

In theory, devices equipped with 3.0 ports are backwards compatible with devices that have previous generation USB connectors. The only limitation will be speed indicator. While the 2.0 will operate at the limit of its speed capabilities, its “big brother” will not use half of its resources.

Increased performance

In the now outdated but still widely used USB 2.0 standard, the data transfer rate was within 460-490 Mbit/s. With the new standard 3.0 this figure can reach 8 times greater value – up to 5 GB per second. What do these numbers mean for the average user? Here's what: now, to transfer large files, such as movies, archives, etc., you will need to spend 10 times less time. However, not all so simple. These indicators characterize only the 3.0 connector standard, and in order to transfer, for example, files to flash memory on high speeds their support is also required by the controller chip, the “flash drive” itself.

Technical features

As written above, connectors 2.0 and 3.0 are compatible with each other. But there are still a number of differences, both design features, and in technical specifications. Both connectors, as before, have four contacts for mutual backward compatibility purposes, however, the cord used in conjunction with the 3rd generation connector has two additional contacts for organizing operation at high speeds, increasing the current used to power various devices, as well as for implementing other benefits. As a result, the cord became slightly thicker, and its recommended length was reduced from five to three meters. In addition, the cord has become a little stiffer due to the introduction of a special shielding coating into the cable to protect against electromagnetic fields induced in it.

It is also worth noting that now the current present in the connector has increased to 950 mA, while in connector 2.0 this figure was 500 mA. As a result, it is now possible to use a higher charge current for charging smartphones and other devices, which significantly reduces the time required to fully charge this class of devices. In addition, the number of devices simultaneously receiving charge from one connector can now be increased.

External differences

At first glance, it is actually very easy to distinguish between USB 2.0 and 3.0 connectors. It's all about the color of the plastic insert on which the four contacts of the connector are attached. In the 3.0 standard, this plastic insert is blue, sometimes even red, while in 2.0 it is black or gray. These two standards have no other external differences.

Price

The average cost for flash memory equipped with a USB 2.0 connector is approximately $10 for 8 GB volume, And $5 for 4 GB. This price is basically not very expensive and suits most buyers. However, it is worth paying for the increase in speed, and not very little.

The price of a flash drive with a 3.0 connector is an order of magnitude more expensive than one with a 2.0 connector. Average cost is $40 or more. This is where the question should arise: are you ready to “spend” that amount of money out of your pocket for an increase in speed. If the purpose of the purchase is a cheap tool for transferring small files, then the choice should still be made in favor of 2.0, but if speed is a fundamental factor in using a flash drive, then you cannot do without the capabilities of 3.0

How to choose the right one

Of course, the characteristics of the 3.0 connector allow you to get significant increase in speed, but before choosing it for purchase, you must carefully read the package included with the device technical description. In some cases, it happens that the device is equipped with a 3.0 connector, but the central processor (controller chip) is not at all designed to work at such high speeds. So it looks like the connector is blue, but there is no significant increase in speed.

In addition, the 3.0 connector can achieve the highest data transfer speed when using the same generation of USB connector at the other end of the wire. If a device with a 3.0 connector is running on one side, and a 2.0 connector on the other, then the speed will be limited by the capabilities of the second generation connector.

If you plan to connect, for example, devices such as a computer keyboard or mouse to the 3.0 connector, then you will not feel any differences from 2.0.

Conclusion

The new third generation offers a lot of new technical features, but today you have to pay for them and pay not so little. Of course, over time and as distribution continues, the cost of a new generation of connectors will decrease and all devices will be equipped only with this type of connector.

Before purchasing devices equipped with 3.0 connectors, you need to weigh the pros and cons. Do you need an increase in speed or will the capabilities provided by the USB 2.0 connector be sufficient?

IntroductionOver the years, the USB 2.0 interface has become something familiar - no one has been asking for a long time whether there are such ports in the system unit. Even the question of their number is no longer so relevant: on any more or less modern motherboard there are a dozen or even more connectors. Several factors ensured such popularity of the interface: ease of connection (drivers for the main types of devices have been built into all popular operating systems for ten years already), prevalence, compactness of connectors, versatility, and the ability to power a connected device from the same connector. External drives, sound cards, printers, scanners, modems, mice and keyboards - all of this has a USB interface, not to mention all kinds of accessories, from table fans to Christmas trees with backlight, which only need power from the port. But nothing lasts forever - the speed of the interface, developed ten years ago, has recently increasingly turned out to be insufficient. In principle, the theoretical throughput of 480 Mbps (60 MB/s) is quite high, but in practice it is virtually impossible to achieve speeds greater than 35 MB/s. If all kinds of mice don’t care about this, then in the case of external drives the USB 2.0 interface has long become a bottleneck - modern hard drives, including 2.5-inch ones, have a much higher reading speed from the platters. What can I say, even the performance of modern fast flash drives exceeds the capabilities of USB 2.0, forcing manufacturers to create “flash drives” with an e-SATA interface, despite the fact that they still have to be powered from a USB connector, since in the current The e-SATA version does not provide this option.

One way or another, the appearance of the next version of the USB interface is long overdue - and now we have USB 3.0. Today it is already present on more than a dozen models of motherboards, but there are practically no peripheral devices with this interface on sale yet - however, we still managed to get a couple of samples into our laboratory.

USB 2.0 and 3.0

Speaking about the features of the new interface, one cannot help but touch on its history, which dates back a decade and a half. The first version of the USB protocol, whose name stands for “Universal Serial Bus,” was introduced in 1995.

Its development was supported by such giants as Microsoft and Intel, who well understood the need to create a new universal interface that could replace the variety of external interfaces that existed at that time (parallel port, serial port, joystick port, external SCSI - and in the end they really disappeared from motherboards). However, USB was also intended to become a fast and at the same time inexpensive external interface - at that time there was a clear drawback in them. Three years later, in 1998, an updated version of the protocol 1.1 was released, and already in 2000 the specification version 2.0 appeared, with which the global distribution of this interface began. It is in this version that the modes Low Speed(speed up to 1.5 Mbit/s) and Full Speed(speed up to 12 Mbit/s) added Hi Speed, providing speeds of up to 480 Mbit/s and allowing the new interface to compete on equal terms with FireWire IEEE1394a with its 400 Mbit/s. However, there wasn’t much competition - thanks to its simpler implementation and licensing scheme, USB 2.0 quickly pushed FireWire into the secluded niche of connecting video cameras, despite some of the latter’s technical advantages.

The USB interface is quite simple to understand. At the head of everything is the host controller - the root device that controls the entire data transfer process. “Hubs”, which are splitters, and end devices are connected to it, directly or through hubs. The total number of devices in this tree can reach up to 128. Splitters can be either passive or active; the latter are distinguished by the fact that they have their own power source, which means they are able to power connected devices without consuming current from the host. By the way, hubs are not “passive” in the truest sense of the word - in practice, they are quite complex electronic devices.

As already mentioned, all information exchange along the bus “tree” is organized by the host. It does this very simply - with a certain periodicity, it takes turns polling the end devices and allocating them certain time intervals during which they can transmit data. The disadvantages of such a scheme are quite obvious: all devices share the bus bandwidth “for everyone” and the more devices there are, the less each of them will get. The picture is somewhat smoothed out by the fact that there are several types of logical communication channels created between devices and the host: a control channel, designed to transmit short commands; interrupt channel, for short commands with guaranteed delivery time; isochronous, with a guaranteed delivery rate of a certain number of packets during a given period, and a streaming channel, in which delivery is guaranteed, but speed and delays are not regulated. Accordingly, different channels are created for different devices (mice and keyboards have an interrupt channel, drives have an isochronous channel). And then, during each period of the bus operation, interrupt packets are transmitted over it, then isochronous packets in the required quantity, and in the remaining time in the period, control and, lastly, flow packets are transmitted.

Once again, I would like to remind you that the “head” of everything is the host controller: it is he who organizes all polls, “listens” for interruptions in the time intervals allocated for this, and sends the devices to sleep. End devices cannot at will go into or out of sleep mode, initiate data exchange, or urgently inform the host of something important (for example, a buffer overflow). Moreover, all organized channels are half-duplex - simultaneous transmission and reception of data is impossible, only in turn. There is no equality in USB: no matter what devices you connect to each other, one of them must play the role of a host, while the rest must obey it.

As the popularity of devices on motherboards grew, so did the number of USB ports. And the manufacturers, in a rather curious way, came out of the unpleasant situation when one bus is shared by everyone - they organized several buses. Thus, in the now popular Intel P55 chipset, upon in-depth examination, we find as many as seven UHCI controllers (responsible for working with Low Speed and Full Speed devices), combined with seven dual-port hubs, and two EHCI controllers working with Hi Speed devices - yes, this no longer a tree, but an intricately woven bush with several roots and several trunks.

Finally, it’s worth mentioning separately about the power provided by the USB bus. The load capacity of one port is limited to 0.5 A, so when connecting multiple devices to it, you need to determine whether they will overload the port. This is achieved quite simply: after connecting, the device must inform the host about how much current it consumes from the port - and remain in sleep mode until it receives permission from the host to turn it on. If the total current consumed by devices exceeds 0.5 A, the host will not give permission to turn on the last connected device. This implementation has one vulnerability: although in principle it is possible to check whether the device really consumes as much as it asked for, such a scheme will complicate and increase the cost of the USB controller, so in the vast majority of cases no check is carried out - the host blindly trusts the devices. On the one hand, this can lead to power overload of the host and even its failure, on the other hand, it allows USB devices whose consumption exceeds 0.5 A to operate, but not too much. The latter include external hard drives: as follows from our measurements, when spinning up the spindle, they consume about 0.7-0.9 A. However, formally they report to the host a consumption of 0.5 A (in fact, they report higher consumption cannot even theoretically: this is not provided for by the USB protocol), and their further operation depends on whether the host is able to provide their actual power consumption, or whether the supply voltage under such a load drops below the minimum permissible. All sorts of USB fans, USB lights and similar devices behave even more “wrongly”: since they usually simply do not have any USB controller inside, they do not report anything to the host about their consumption. No matter how many such devices are plugged into the connector, the host will consider the load to be zero.

Obviously, the situation when a large and very popular class of devices - external hard drives - takes advantage of the fact that the severity of the law is mitigated by the optionality of its implementation is not normal, therefore the low load capacity of USB 2.0 ports can also be attributed to their disadvantages. Other consumers, such as scanners, compact speaker systems, mini-monitors and various chargers, would also benefit from additional power.

Finishing the conversation about USB 2.0, it’s worth remembering the physical level, or more precisely, the cables. They have four wires: two for data transmission, ground and +5 V for power circuits. The original specification specified the use of standard Type A flat connectors on the host side and Type B connectors on the device side. But subsequently, a host of compact connectors were quickly added to them - several variants of mini-USB and micro-USB.

Well, now it’s time to talk about USB 3.0. The new version of the standard brought us a new operating mode, Super Speed, the most main feature which was an increase in the maximum data speed by an order of magnitude - up to 4.8 Gbit/s. The main requirements when developing the new standard were compatibility with all existing USB-enabled equipment and maintaining a simple interface.

The developers chose a path that can be called “growth in breadth.” To the existing parallel UHCI and EHCI controllers, another one was added, which is responsible specifically for the operation of devices in Super Speed mode. In this way, it was possible to maintain compatibility and add a new data transmission channel, which will not be affected by “slow” old devices.

The cables with connectors also changed accordingly: to the already existing four wires, two more pairs of signal wires were added, one of which is responsible for transmitting data to the controller, and the second from it, and one additional ground. The hardest thing was for the connectors, which became quite intricate - five more contacts were introduced into them, while maintaining compatibility with the old connectors. However, there is a certain advantage to this: devices with USB 3.0 support are easy to recognize, just look at the connector.

USB 3.0 type A

USB 3.0 type B

USB 3.0 type Micro-B

However, in addition to increased speed, the new standard brought many more interesting things. Firstly, it has increased the current that the device can request - now the upper limit has dropped to 0.9 A. External drives on 2.5-inch hard drives will be especially happy about this - their manufacturers can finally refuse Y-shaped cables that collected power from two ports at once, and from the method of violating the standard described above in order to ensure the operation of the device. Secondly, the two data lines clearly hint that the USB 3.0 standard allows you to simultaneously transmit and receive data. Thirdly, the standard brought a full-fledged interrupt mechanism, which made it possible to abandon device polling, which was so disadvantageous in terms of wasting precious time. Fourth, devices are now allowed to create more than one data channel. Energy saving has not been forgotten: with the advent of interrupts, it has become possible to implement power management for devices, with low-consumption modes initiated by the devices themselves. In fact, the entire architecture was radically redesigned, and compatibility with the previous standard was, one might say, “glued up.”

Well, perhaps, at this point we will stop getting acquainted with the theory (those who want to do this in more detail can study the documentation on the site), it’s time to evaluate how good the new interface is.

Test participants

Buffalo HD-H1.OTU3

Externally, the Buffalo drive is not anything special: we have a neat plastic parallelepiped inside which hides a 3.5-inch Samsung HD103SJ hard drive. At one of the ends of the “brick”, which is supposed to be placed vertically (but I want to put it down - the stability of the device is too low, in the absence of any legs), there are connectors. It is this end that is most interesting for us, because in addition to the power connector (alas, “large” 0.9 A drives are still not enough for full-fledged operation) and a small fan, there is a USB 3.0 type B connector located here, noticeably different from the usual connector of the old standard.

Vantec NextStar 3

The second sample was a container from a well-known manufacturer of such devices, Vantec. This example also stands on the end, although longer, and uses a small stand. However, its stability still raises concerns.

Alas, the Buffalo container was non-separable, but inside the Vantec we found an ASMedia ASM1051 chip.
As for the root part of USB 3.0, in this case its role is played by virtually the only root controller chip common today - the NEC µPD720200.

What was especially pleasant for us was that the USB 3.0 controller we received from ASUS uses four PCI-Express lanes, which means we can be sure that it will have enough channel width. Alas, at the moment using a separate controller is the best option, since on motherboards we see the same NEC controller, but with an unknown channel width to it (one PCI-Express 1.1 line is not enough for the controller - its bandwidth is less than USB 3.0), and there are no controllers built into the chipset yet.

Testing methodology

The following programs were used during testing:

IOMeter version 2003.02.15;
FC-Test version 1.0;

The test system was as follows:

ASUSTeK P5WDG2 WS Pro motherboard;
Intel Core 2 Duo E2160 processor;
IBM DTLA-307015 15 GB hard drive as a system drive;
Radeon X600 video card;
1 GB DDR2 800 MHz system memory;
Operating system Microsoft Windows XP Professional SP2 (Windows Vista in the case of the PCMark Vantage test).

Testing was carried out with basic drivers operating system. The drives were allocated for the FAT32 and NTFS file systems in one partition with the default cluster size. In some cases described below, logical partitions of 32 GB in size, partitioned under FAT32 and NTFS with the default cluster size, were used for testing. In all tests, internal drives were connected to a port on the motherboard and operated with AHCI mode activated, and external drives were connected to a USB 3.0 port on an ASUS expansion card or a USB 2.0 port on the motherboard.

We decided that the most interesting thing for both us and you would be to compare not only the two containers with each other and compare the two versions of the standard, but also compare them with what the familiar SATA 300 gives us. Therefore, you will see four sets of data at once - two for different containers on USB 3.0, one for the Vantec container when working via USB 2.0 and one for hard drive on SATA 300. In all cases, a Samsung HD103SJ drive was used. However, due to the fact that the Buffalo container was non-separable, we had to do something rather unique. We knew that Buffalo was using a Samsung HD103SJ, so we chose the same drive to use in Vantec and in the containerless version. It is clear that due to the fact that the disks themselves are somewhat different, we get some scatter in the results, caused not by the interface we are interested in, but by the disks themselves, but still, we tried to reduce the difference to a minimum.

Additionally, we used an SSD in a couple of tests Intel X25-M G2 160 GB.

IOMeter: Sequential Read & Write

Let's start with tests in IOMeter. The first, as always, will be sequential operations. In this test, a stream of requests is sent to the drives with a command queue depth of four. Once a minute, the size of the data block increases. As a result, we get the opportunity to trace the dependence of the linear read and write speeds of drives on the size of the used data blocks and estimate the maximum achievable speeds.

If you wish, you can see the numerical results of measurements here and below in the corresponding tables, but we will work with graphs and diagrams.

Superiority new version The interface above the old one is visible to the naked eye - the maximum speed when transferring data via USB 3.0 is exactly the same as when working via SATA 300, while the old interface is limited to the expected 33.5 MB/s. At least one problem has been solved - the interface speed is clearly sufficient for modern drives. True, it was not possible to completely get rid of additional delays - look at the speed of working with small blocks, it is noticeably lower for USB 3.0 than for SATA 300. What is especially curious is that when installing an SSD in a container, we see exactly the same speeds - we are clearly facing some kind of performance limitation. To be honest, it is still difficult for us to say whether we see insufficient performance of the USB chip, or a fundamental limitation of the new bus associated with its architecture.

But we were even more surprised by the SSD's results in terms of the maximum speed achieved. We specifically checked them several times and tried to take other SSDs - no, that's right, the speed is limited at 160 MB/s. Of course, this is much better than 35 MB/s, but somehow it still doesn’t seem at all like the tenfold promised speed increase! I really want to hope that we are faced with the shortcomings of the first implementations of USB 3.0 and in the future we will see speeds worthy of the declared 4.8 Gbit/s.

The picture in the recording is the same: USB 3.0 clearly outperforms its predecessor; the bandwidth of the new interface is quite enough to serve a modern 3.5-inch hard drive. Unfortunately, the drop in speed on small blocks has not gone away and is too clearly repeated to be considered an accident.

IOMeter: Disk Response Time and IOMark: Average Positioning Speed

To measure the response time, we use “IOMeter” to send a stream of requests to the drives for reading or writing data blocks of 512 bytes within ten minutes with a queue depth of outgoing requests equal to one. The number of requests processed by the drive is such that it obviously exceeds the capacity of the buffer memory. As a result, we get a stable drive response time.

Over the course of the response, the situation turned out to be quite funny. On the one hand, we see a clear improvement: the new protocol introduces less latency than its predecessor, although it still lags noticeably behind SATA 300 - this is clearly visible in the results of the Vantec container, which used the same drive that we connected via SATA. But in Buffalo there was another copy of the disk, albeit of the same model, and its results were radically different. Of course, we can assume that this container uses a “slow” chip with imperfect firmware, but we are inclined to attribute most of the difference to differences in the disks themselves. Thus, the results of testing the response to an SSD at Vantec, interesting in themselves due to the extremely short response time of the drive itself, show that the increase in response time due to the influence of the interface, namely the Super Speed protocol, is very small.

IOMeter: Random Read & Write

Let us now evaluate the dependence of drive performance in read and write modes with random addressing on the size of the data block used.

We will consider the results in two versions. For small blocks, we will plot the dependence of the number of operations per second on the size of the block used, and for large blocks, instead of the number of operations, we will take the speed measured in megabytes per second as a performance criterion.

In terms of performance on small blocks, the new interface is not so different from the old one: they are both slightly worse than SATA 300, but still performance is much more determined by the disk than the interface. But for any large requests (say, 1-2 MB - consider that we are looking at photos from a fragmented disk), the new interface is already noticeably better than the old one. Moreover, its implementation in Vantec is clearly better - it is just a little behind in speed from a drive connected via SATA. As the block size increases, the difference increases even more.

But on the recording we see a slightly different picture. On small units, the SATA hard drive is clearly faster, while all external interfaces run at approximately the same speed, almost twice as fast as the drive. On larger units, the USB 3.0 lag decreases. The most interesting thing begins when the request size grows to 2 MB or more - USB 2.0 comes to maximum speed, and SATA and USB 3.0 continue to increase speed nicely. It's interesting that Vantec again turns out to be noticeably better than Buffalo, although the latter's behavior is more predictable and regular.

IOMeter: Database

Using the Database test, we determine the ability of drives to handle streams of requests to read and write 8-kilobyte data blocks with random addressing. During testing, there is a consistent change in the percentage of write requests from zero to one hundred percent (in increments of 10%) of the total number of requests and an increase in the depth of the command queue from 1 to 256.

You can view tables with complete test results at the following links: .

In this case, we will not build summary diagrams, but compare the diagrams with the results of each drive with each other.

The comparison turned out to be extremely clear, especially thanks to the interesting behavior of the Samsung drive. In the case of USB 2.0, it sharply loses ground - delayed writing almost disappears, and reordering of read requests is extremely difficult to find - any noticeable performance increase is observed only with queue 16.

USB 3.0 as implemented by Vantec looks a little more interesting - in large queues the performance increase becomes more noticeable. True, a queue of four requests is still almost no different from a single one. But in the case of USB 3.0 according to the Buffalo version, the disk draws something incredible. The shape of the graphs is such that if it were a SATA drive, we would say that its firmware is extremely imperfect. Apparently, the controller in the container itself tries to help the disk on deep queues to the best of its ability, but does so very unstable. However, one point remained unchanged: at shallow queue depths, the performance difference is still minimal.

IOMeter: Webserver, Fileserver

In this group of tests, drives are tested under loads typical of servers and workstations.

We remind you that in “Webserver” and “Fileserver” we emulate the operation of the drive in the corresponding servers, while in “Workstation” we simulate the operation of the disk under a typical workload for a workstation, with a maximum queue depth limit of 32 requests. Testing in “Workstation” is carried out both using the entire disk space of the drive, and when working only with a 32 GB address space.

Since the topic of our article is the interface for external drives, we will be brief and just compare performance ratings - nevertheless, such loads are far from the most typical.

The results were somewhat unexpected. Thus, for USB 3.0 servers, the Buffalo implementation is clearly more interesting than the Vantec version, although they both lag behind a drive connected via SATA. For workstations the picture is similar, but Buffalo demonstrates noticeable superiority only in a reduced working area. As for comparing USB 3.0 with previous version interface, then in the case of Vantec the gap is very small, but if we also take Buffalo into comparison, it is quite significant.

IOMeter: Multi-thread Read & Write

This test allows you to evaluate the behavior of drives under multi-threaded load. It emulates a situation where from one to four applications work with the drive, and the number of requests from them varies from one to eight, and the address spaces of each application, whose roles are played by workers in IOMeter, do not intersect.

If you wish, you can see tables with complete testing results at the corresponding links, and we will consider the most indicative diagrams of writing and reading for situations with a queue depth of one request, since when the number of requests in the queue is two or more, the speed values are practically not depend on the number of applications.

A comparison of USB 3.0 and USB 2.0 on a multi-threaded load is without surprises - with one thread, the new standard clearly wins, since it allows you to realize the full speed of the disk, and even at several, although it loses in speed, it remains ahead of its predecessor, almost twice as fast.

Another thing is more interesting: in the case of three and four reading threads, a disk via USB works faster than via SATA. We don’t know what exactly allowed the disk to increase speed, but the result is clear and stable and cannot be attributed to chance.

Multi-threaded recording proceeds without incident - the drives carried the ratio of their speeds through all variants of this load. USB 2.0 clearly plays the role of a bottleneck, so much so that the drive does not pay attention to the number of threads at all, but in other cases the speed gradually decreases gradually as the number of threads increases.

FC-Test

Let’s complete the testing in our favorite “FileCopy Test”. Two 32 GB partitions are created on the drive, partitioned at two stages of testing, first in NTFS and then in FAT32, after which a certain set of files is created on the partition, read, copied within the partition, and copied from partition to partition. The time of all these operations is recorded. Let us remind you that the “Windows” and “Programs” sets include a large number of small files, and the other three templates (“MP3”, “ISO” and “Install”) are characterized by a smaller number of larger files, and in “ISO” The largest files are used.

Don't forget that the copy test not only tells you the speed of copying within one drive, but also allows you to judge its behavior under heavy load. In fact, during copying, the drive simultaneously works with two threads, one of them for reading, and the second for writing.

We will consider in detail only the values achieved in NTFS; you can find out the results of testing in FAT32 from the table at the following link:.

The diagrams are extremely similar to each other and quite predictable, so what can we say about different modes Separately it makes no sense. Overall, USB 3.0 demonstrates that, unlike its predecessor USB 2.0, it really is an interface capable of fully realizing the speed characteristics of modern hard drives under any load. The fee for external execution in the case of working with files turned out to be extremely low - yes, drives in containers with USB 3.0 still lag behind their SATA counterparts, but not by that much. Moreover, during reading this lag is very small and amounts to less than 10% of the speed, but during writing it increases to approximately 15 percent. Yes, it’s most noticeable when copying small files, but tell me, how often do you do this? Still, in most cases, an external drive is used either for reading or writing. Compared to the new interface, the old USB 2.0 looks very poor - no matter what you say, its time as an interface for storage devices has passed.

Summarizing

Frankly speaking, we were left with somewhat ambivalent impressions. On the one side, speed characteristics USB 3.0 is really enough to fully realize the capabilities of modern hard drives, and the presence of radical changes in the protocol puts us in an optimistic mood. On the other hand, we did not see the promised tenfold increase in speed - the devices that fell into our hands were unable to produce more than 160 MB/s where SATA 300 easily demonstrates 250 MB/s. But Moscow was not built right away, and early implementations of USB 2.0 were also quite flawed in terms of speed - it is possible that after some time we will see faster USB 3.0 chips. However, I would even more like to see chipsets in which support for the new standard will be native, and not implemented using a third-party chip. Until this happens, it is difficult to expect much popularity of the new standard, because at least in the field of external drives it already has a serious opponent in the form of e-SATA, which, although it does not provide power to the devices connected to it, does provide this moment is much more widespread than USB 3.0, and the speed, as we see, is higher. However, in the long term, victory will undoubtedly remain with USB 3.0 - and the only question is how long it will take.