Comparison of performance of Intel processors of different generations. Comparison of AMD processors of different generations: Phenom II, Bulldozer and Vishera

AMD processors of the Phenom II, Bulldozer and Vishera architectures were tested repeatedly within the laboratory, both their overclocking and performance levels were studied. But comparative tests“face to face” is not enough, so it is difficult to assess the presence (or absence) of progress when moving from one generation of CPU to another. It's time to fill in the gaps.

IN this review top AMD solutions will take part recent years- This AMD Phenom II X6 1100T, AMD FX-8150 and AMD FX-8350. To complete the picture, all CPUs will be tested not only in normal mode and at equal frequencies, but also at maximum overclocking. In addition, a comparison will be made of the power consumption of processors. The Core i7-2600K was taken as a reference from Intel's camp.

Test bench and software

Testing was carried out on the following configuration:

• Motherboards:
• ASUS Crosshair V Formula;
• ASUS Sabertooth Z77;
• Processors:
• AMD Phenom II X6 1100T 3.3 GHz (16.5x200);
• AMD FX-8150 3.6 GHz (18x200);
• AMD FX-8350 4.0 GHz (20x200);
• Intel Core i7-2600K 3.4 GHz (34x100);
• Cooling system: Zalman CNPS10X Performa (120*120*25, ~2000 rpm);
• Thermal interface: Prolimatech PK-1;
• RAM: G.Skill TridentX F3-2400C10D-8GTX;
• Video card: ASUS ARES II, CrossFireX Disabled;
• HDD: Western Digital Caviar Blue (WD500AAKS), 500 GB;
• Power supply: Corsair CMPSU-750HX, 750 W;
• Housing: open test bench.

Software

• Operating system: Windows 7 Ultimate SP1 x64;
• Video card driver: Catalyst 13.5 beta 2;
• Additional software:
• FRAPS 3.5.9, build 15586;
• AutoHotkey 1.0.48.05.

Testing methodology

The following applications were used for the performance test:

• LinX 0.6.4 + Linpack 11.0.1.005;
• TrueCrypt 7.1a;
• SVPmark 3.0.3a;
• Fritz Chess Benchmark v.4.2;
• Maxon Cinebench 11.5 x64;
• POV-Ray v3.7 RC7;
• x264 HD Benchmark 5.0.1;
• TOC F@H Bench v.0.4.8.1;
• WinRar 4.2 X64;
• 7-Zip 9.30 X64.

The following games used built-in performance measurement tools:

• Batman: Arkham City;
• Hitman: Absolution;
• Metro 2033;
• Sleeping Dogs;
• Tomb Raider (2013).

In these games, performance measurements were carried out using Autohotkey scenes:

• Crysis 3 (Welcome to the Jungle);
• Far Cry 3 (Harvest The Jungle);
• The Elder Scrolls V: Skyrim (Goldflower Estate);
• The Witcher 2 (On the Front Lines).

To analyze gaming performance results, both Min/AVG FPS numbers and performance ratings based on frametimes results were used.

The frametimes analysis was carried out using the Fraps-Calc utility, which allows you to see the main features associated with system performance in a particular application. This program Based on the average (AVG) FPS and its stability characteristics, it calculates the value of the so-called performance rating. We can say that the performance rating is a numerical characteristic of the comfort of the gameplay, where a value of 1 and above means the absence of visible “brakes” when displaying an image.

Each of the processors was tested in three operating modes:

• Normal mode: completely factory system settings obtained by reset BIOS, respectively – the results “out of the box”, which will be if the system is not touched. I note that both motherboards used the Turbo Core/Turbo Boost functions in this mode, and for all CPUs the memory operating mode was set to DDR3-1600 11-11-11;
• Comparison of CPUs at the same frequency (4 GHz). Memory operating mode DDR3-1600 7-8-8-21-1T, Turbo Core/Turbo Boost disabled, HT/CPU_NB frequencies AMD processors set to standard values;
• Maximum overclocking mode for each processor.

For AMD Phenom II X6 1100T This:

• Processor frequency: 4174 MHz (260.88x16);
• CPU_NB operating frequency: 2870 MHz;
• HT operating frequency: 2609 MHz;
• Memory operating mode: DDR3-2087 8-10-10-25-1T.

For AMD FX-8150 This:

• Processor frequency: 4615 MHz (200.66x23);
• CPU_NB operating frequency: 2609 MHz;
• HT operating frequency: 2609 MHz;
• Memory operating mode: DDR3-2140 8-10-10-25-1T.

For AMD FX-8350 This:

• Processor frequency: 4592 MHz (199.66x23);
• CPU_NB operating frequency: 2396 MHz;
• HT operating frequency: 2596 MHz;
• Memory operating mode: DDR3-2396 10-11-11-28-1T.

For Intel Core i7-2600K This:

• Processor frequency: 4700 MHz (47x100) with HT enabled and 4800 MHz (48x100) with HT disabled. Both settings have been tested;
• Memory operating mode: DDR3-2133 8-10-10-25-1T.

Energy consumption measurements were made using a Mastech MY64 multimeter and a 50 A 75 mV shunt (75SHIP1-50-0.5) at the positive break of the 8-pin power cable. Measurements were made for two modes: for standard mode and for maximum processor overclocking mode. LinX 0.6.4 was used as a load for measurements.

Performance testing

LinX

Testing was carried out with a task volume of 20014 (3072 MB of memory). The final result is the best based on the results of five passes.

Gflops
Regular mode

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

When comparing participants in standard mode among AMD models, the results were arranged in order of processor release, the main progress was observed in the transition from Phenom II to Bulldozer, while the advantage of the FX-8350 is more likely explained by the 400 MHz difference in frequencies than by architectural features.

I note that this balance of power was not always the case: at the time of the release of the FX-8150, the Linpack test did not yet support instruction sets of new CPUs, and the results of AMD FX processors were several times lower than in modern versions Linpack, accordingly, the results of Phenom II were unattainable for AMD FX. As for the results of the i7-2600K, we can say that it does not work at its full potential - turning on HT in this test reduces performance, as a result, the Intel CPU is in the middle between the FX-8150 and FX-8350.

When compared at equal frequencies, the difference between AMD processors decreases significantly: the results of the Phenom II X6 have increased noticeably (no wonder, since it has the lowest operating frequency among the test subjects - 3300 MHz), and the FX-8150 and FX-8350 are equal. The i7-2600K is already ahead in this mode.

After switching to the maximum frequency mode, the status quo did not change. The Phenom II is behind, and the FX-8150 and FX-8350 have swapped places according to the frequencies achieved, but the frequency potential of the Vishera is slightly lower. As for the Core i7-2600K, its advantage only increased thanks to higher overclocking, and is especially noticeable when HT is turned off.

TrueCrypt 7.1a

A multi-threaded benchmark that measures the speed of encryption operations. The result is taken as Mean Speed ​​in the AES-Twofish-Serpent test. The result is the best based on the results of five measurements.

True Crypt AES-Twofish-Serpent

MB/s
Regular mode

Please enable JavaScript to see graphs

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

For the normal operating mode, the balance of forces is approximately similar to what was observed when considering the LinX results - both when moving from Phenom II to Bulldozer, and when moving from Bulldozer to Vishera, performance varies significantly, and the progress cannot but rejoice. What has changed is the performance level of the Core i7-2600K, which this time is already between the Phenom II X6 1100T and FX-8150, and not between the FX-8150 and FX-8350.

The transition to comparison at equal frequencies this time does not change the conclusions - the status quo this time remains the same at 4 GHz. The AMD FX-8350 remains unrivaled, and the i7-2600K becomes even closer to the Phenom II X6 1100T than the FX-8150.

The results with maximum overclocking do not bring any surprises. It's interesting to see the performance of the i7-2600K with HT turned off, where the performance level becomes even lower than that of the Phenom II X6 1100T, despite the more than 600 MHz difference in processor frequency. It may be noted that, contrary to more low frequencies The FX-8350 remains the leader compared to the FX-8150; apparently, architectural optimizations have benefited.

SVPmark 3.0.3a

SVP software, designed to increase the smoothness of the played video by calculating additional frames. The test is multi-threaded and very demanding on processor performance. The result is the best based on the results of five measurements.

SVPmark 3.0.3a

Points
Regular mode

Please enable JavaScript to see graphs

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

When comparing processors at stock frequencies, the results are not much different from the two previous performance tests: AMD CPUs are once again lined up according to the hierarchy, with the greatest difference observed when moving from the Phenom II X6 1100T to the FX-8150. The Intel Core i7-2600K is located between the FX-8350 and FX-8150, closer to the latter.

At equal frequencies, you can see that, despite the leveling of the difference in frequencies, the balance of power in the case of AMD solutions remains the same. But at the same time, the i7-2600K is already in the lead, albeit with a minimal difference.

The same situation among AMD models remains the same with maximum CPU overclocking, while the i7-2600K strengthens its leadership. True, with HT turned off, it is no longer in the lead, but only competes with the FX-8150.

Fritz Chess Benchmark v.4.2

This test forces the processor to process chess algorithms and puts a lot of stress on modern CPUs. The final result is the best based on the results of five measurements.

Fritz Chess Benchmark

knodes/s
Regular mode

Please enable JavaScript to see graphs

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

The balance of power here looks much more interesting than in previous tests, and interesting numbers can be seen already for the normal operating mode of the processors.

It shows that the Phenom II X6 1100T and FX-8150 are very close, despite the fact that the FX-8150 operates at 300 MHz higher, while the FX-8350's advantage over them is not so significant. This is confirmed by comparison results at equal frequencies, where Phenom II turns out to be the most productive among the three AMD flagships. Apparently, a full six cores for this test are preferable to four “dual-core” modules from AMD FX.

Do not forget that the frequency potential of the Phenom II is much lower, and as a result, with maximum overclocking, all three AMD processors show approximately the same result. With more/less luck of specific CPU instances, any of the participants could show the best result.

As for the results of the Core i7-2600K, it leads in all modes except maximum overclocking with HT turned off.

Maxon Cinebench 11.5 x64

This benchmark analyzes the rendering speed of a test scene, assigning performance points to the processor. The final result is the best based on the results of three measurements.

Cinebench R11.5

CPU
Points
Regular mode

Please enable JavaScript to see graphs

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

The results are very similar to the Fritz Chess Benchmark. Again, when comparing AMD solutions at standard frequencies, the Phenom II X6 1100T holds up well, especially compared to the AMD FX-8150. The only difference is that this time the FX-8350 shows more confident results. When moving to 4 GHz, the six-core “old man” expectedly begins to lead among AMD CPUs, although the FX-8350 is still closer to it than it was in the case of the chess test.

When moving to maximum overclocking, the results of the processors are already consistent with generations of architectures, although the difference in the final results is still small, especially for the X6 1100T and FX-8150.

The i7-2600K, as in the chess test, leads in all modes except maximum overclocking with HT turned off.

POV-Ray v3.7 RC7

A program for constructing images using the ray tracing method. The built-in benchmark was used. The final result is the scene rendering time, the best result based on three measurements.

POV-Ray v3.7 RC7

Seconds
Regular mode

Please enable JavaScript to see graphs

4000 MHz

Please enable JavaScript to see graphs

Overclocking

Please enable JavaScript to see graphs

Already from the results of the standard mode it is clear that, unlike the two previous tests, Phenom II no longer shines here, and such values ​​​​cannot be explained by clock frequencies alone.

This is confirmed by the results at 4 GHz, where the processors are arranged according to the generations of their architectures, with a similar level of performance gain when moving from one to another. At maximum overclocking, the difference between the FX-8150 and the Phenom II X6 1100T only increases.

As for the Intel Core i7-2600K, it shows poor results in this test. At equal frequencies, its performance level is close to the Phenom II X6 1100T, and the FX-8150 and FX-8350 are more productive than the Intel processor in all respects test modes. At the same time, with maximum overclocking and HT turned off, the Core i7 representative remains an outsider.

This article presents only the best AMD processors in 2017.

If you don’t want to independently understand all the characteristics of each processor model or are not sure what you can choose the best option, take a look at our AMD CPU rating.

Contents:

Good processor- this is the main indicator of power and. AMD is one of the leaders in the processor market.

AMD produces the following types of processors:

• CPU – central computing units
• GPU – separate device, which renders the video. Often used in gaming computers to reduce the load on the central unit and to ensure best quality video sequence;
• APU – central processors with a built-in video accelerator. They are also called hybrid, because such a component is a combination of the central one and in one crystal.

№5 - Athlon X4 860 K

The AMD Athlon line is designed for the Socket FM2+ socket. The X4 860K is the best and most powerful model of the entire series, which comes with three processors:

• Athlon X4 860K;
• Athlon X4840;
• Athlon X2

The Athlon family is designed for desktop personal computers. All models in the line are distinguished by good multi-threading.

Best results The Athlon group showed the X4 860K model.

The first detail to note is the support for virtually , which consumes no more than 95 watts along with quiet operation and no loss in performance.

If the processor has been overclocked using special programs, an increase in noise in the operation of the cooling system may be observed.

Main characteristics:

• Family: Athlon X4;
• Number of processor cores: 4;
• Clock frequency – 3.1 MHz;
• There is no unlocked multiplier;
• Core type: Kaveri;
• Approximate cost: $50.

There is no integrated graphics in the CPU.

The X4 860K processor is capable of supporting fast operation of general-purpose systems only.

Testing of the CPU operation was carried out using the AIDA64 utility. Overall, the model shows good results for a mid-class processor.

If you're looking for an affordable, multitasking CPU for your home computer, the Athlon X4 860K is one of the suitable options.

Fig.3 – testing Athlon X4 860K

No. 4 – AMDFX-6300

AMD's FX-6300 is a CPU that supports the Piledriver architecture. Processors with this architecture have already become worthy competitors to new products from Intel.

All processors from AMD FX group have excellent overclocking potential.

FX-6300 Features:

• Series: FX-Series;
• Supported connector: Socket AM3+;
• Number of cores: 6;
• No integrated graphics;
• The clock frequency is 3.5 MHz;
• Number of contacts: 938;
• The cost of the model is on average $85.

A characteristic feature of the processor is its flexibility.

The clock frequency declared by the developer is 3.5 MHz, which is a rather mediocre figure among.

However, this CPU provides the ability to overclock the frequency to 4.1 MHz.

Fig. 4 – boxing of FX series devices from AMD

Acceleration of work occurs during intense loads. Most often in the process of rendering videos or working with games.

It should be noted that this CPU model is equipped with a dual-channel memory controller.

Processor speed testing was carried out in Just Cause 2.

The final results showed that the Athlon X4 860K supports maximum resolution graphics at 1920 x 1200 pixels.

The computer also used an integrated GTX video card 580.

In the picture below you can see comparative analysis performance and other processors that have been tested under identical software and hardware environment conditions.

Fig. 5 – test result of Athlon X4 860K

№3 - A10-7890 K

The A10-7890K is a hybrid CPU from AMD. Despite the announcement of the development of a fundamental new technology and generations of processors, AMD decided to release another model in the A10 line.

The company positions this series of devices as great choice for desktop PCs.

The A10-7890K is a best-in-class playback solution.

Of course, you will have to lower the graphics settings, but as a result you will get good performance without severe overheating PC hardware.

Fig.6 – packaging of model A10-7890K

This processor has a built-in Radeon graphics unit that allows you to:

The processor comes with a Wraith cooler, which features very quiet operation. Also, the cooler supports backlight mode. Specifications A10-7890K:

• CPU Family - A-Series;
• Clock frequency: 4.1 MHz;
• Type of connector: Socket FM2+;
• Number of cores: 4 cores;
• There is an unlocked multiplier;
• Number of contacts: 906;
• Estimated cost – $130.

The main advantage of the A10-7890K is improved interaction with Windows 10.

The detailed characteristics of the processor are shown to us in the figure below:

Fig. 7 – detailed characteristics of APU A10-7890K

Results of testing the component using the standard Cinebench R15 test:

Fig. 8 – Cinebench R15 test result

As you can see, the tested component has surpassed in its parameters some AMD models in the A-10 and Athlon line.

At the same time, the results obtained were not enough to outperform analogues from Intel.

№2 - Ryzen 5 1600 X

The first two places in our TOP are occupied by models of the Ryzen line. It is in the last few years that the architecture of these processors has become key for Advanced Micro Devices Corporation.

The presented Zen microarchitecture is gradually returning the manufacturer to its leading position in the market.

Ryzen 5 is a direct competitor to the group's processors. The CPU performs best in gaming systems. This is also stated by the CEO of AMD.

Characteristics:

• AMD Ryzen 5 Family;
• 6 cores;
• No integrated graphics;
• There is an unlocked multiplier;
• Clock frequency 3.6 MHz;
• Socket AM4 connector;
• The cost is about $260.

Most modifications of the 1600X lack the native . Users will have to purchase this component separately.

Base frequencies do not cross the established 3.6 MHz mark. When operating in turbo mode (as a result of overclocking the processor), the clock frequency reaches 4.0 MHz.

All fifth-generation Ryzen models support SMT - surface mount technology.

This way, the CPU can be easily mounted on the surface of the PCB without the need to trim parts of the component.

Fig.9 – Ryzen package 5

In the process of testing the CPU operation even with the most resource-intensive programs, Maximum temperature The CPU did not exceed 58 degrees. , Test results:

Fig. 10 – test of the 1600X model

Along with the line of powerful CPUs, AMD also released special firmware for their initial setup– AGESA.

The utility allows you to reconfigure memory to avoid delays and interruptions in work.

AMD processors first appeared on the market in 1974, following the presentation by Intel of its first 8080-type models and were their first clones. However, the very next year the am2900 model of its own design was introduced, which was a microprocessor kit that began to be produced not only by the company itself, but also by Motorola, Thomson, Semiconductor and others. It is worth noting that the Soviet microsimulator MT1804 was also made on the basis of this kit.

AMD Am29000 processors

The next generation - Am29000 - full-fledged processors that combine all the components of the kit into one device. They were a 32-bit processor based on RISC architecture, having a cache of 8 KB. Production began in 1987 and ended in 1995.

Besides own developments, AMD also produced processors manufactured under license from Intel and bearing similar markings. So, the Intel 8088 model corresponded to Am8088, Intel 80186 - Am80186, and so on. Some models were modernized and received their own markings, which differed slightly from the original, for example Am186EM - improved analogue of Intel 80186.

AMD C8080A processors

In 1991, a line of processors designed for desktop computers. The series was designated Am386 and used microcode developed for the Intel 80386. For embedded systems, similar processor models were launched into production only in 1995.

AMD Am386 processors

But already in 1993, the Am486 series was introduced, intended for installation only in its own 168-pin PGA connector. The cache ranged from 8 to 16 KB in upgraded models. The family of embedded microprocessors is designated Elan.

AMD Am486DX processors

Series K

In 1996, production of the first family of the K series began, designated K5. To install the processor, a universal socket was used, called Socket 5. Some models of this family were designed for installation in Socket 7. The processors had a single core, the bus frequency was 50-66 MHz, and the clock frequency was 75-133 MHz. The cache was 8+16 KB.

AMD5k series processors

The next generation of the K series is the K6 processor family. During their production, proper names begin to be assigned to the kernels on which they are based. So, for the AMD K6 model the corresponding code name is Littlefood, AMD K6-2 - Chomper, K6-3 - Snarptooth. The standard for installation in the system was the connector Socket type 7 and Super Socket 7. The processors had one core and operated at frequencies from 66 to 100 MHz. The first level cache was 32 KB. For some models there was also a second level cache, 128 or 256 KB in size.

AMD K6 Processor Family

Since 1999, the production of Athlon models began, part of the K7 series, which have received widespread and well-deserved recognition from many users. In the same line there are also budget models Duron, as well as Sempron. The bus frequency ranged from 100 to 200 MHz. The processors themselves had clock frequencies from 500 to 2333 MHz. They had 64 KB of first-level cache and 256 or 512 KB of second-level cache. The installation connector was designated as Socket A or Slot A. Production ended in 2005.

AMD K7 series

The K8 series was introduced in 2003 and includes both single-core and dual-core processors. The number of models is quite diverse, as processors have been released for both desktop computers and mobile platforms. Various connectors are used for installation, the most popular of which are Socket 754, S1, 939, AM2. The bus frequency ranges from 800 to 1000 MHz, and the processors themselves have clock speeds from 1400 MHz to 3200 MHz. L1 cache is 64 KB, L2 - from 256 KB to 1 MB. An example of successful use is some Toshiba laptop models based on Opteron processors, codenamed according to the core codename - Santa Rosa.

AMD K10 Processor Family

In 2007, the release of a new generation of K10 processors began, represented by only three models - Phenom, Athlon X2 and Opteron. The processor bus frequency is 1000 - 2000 MHz, and the clock frequency can reach 2600 MHz. All processors have 2, 3 or 4 cores depending on the model, and the cache is 64 KB for the first level, 256-512 KB for the second level and 2 MB for the third level. Installation is carried out in connectors such as Socket AM2, AM2+, F.

The logical continuation of the K10 line is called K10.5, which includes processors with 2-6 cores, depending on the model. The processor bus frequency is 1800-2000 MHz, and the clock frequency is 2500-3700 MHz. The work uses 64+64 KB of L1 cache, 512 KB of L2 cache and 6 MB of third-level cache. Installation is carried out in Socket AM2+ and AM3.

AMD64

In addition to the series presented above, AMD produces processors based on the Bulldozer and Piledriver microarchitecture, manufactured using a 32 nm process technology and containing 4-6 cores, the clock speed of which can reach 4700 MHz.

AMD a10 processors

Nowadays, processor models designed for installation in the FM2 socket, including hybrid processors of the Trinity family, are very popular. This is due to the fact that the previous implementation of Socket FM1 did not receive the expected recognition due to relatively low performance, as well as limited support for the platform itself.

The core itself consists of three parts, including graphics system with the Devastrator core coming from Radeon video cards, a processor part consisting of an x-86 Piledriver core and a north bridge, which is responsible for organizing work with RAM, supporting almost all modes, up to DDR3-1866.

Most popular models of this family - A4-5300, A6-5400, A8-5500 and 5600, A10-5700 and 5800.

The flagship models of the A10 series operate with a clock frequency of 3 - 3.8 GHz, and when overclocked they can reach 4.2 GHz. The corresponding values ​​for A8 are 3.6 GHz, with overclocking - 3.9 GHz, A6 - 3.6 GHz and 3.8 GHz, A4 - 3.4 and 3.6 GHz.

The result is banal: it is impossible to judge the performance of any central processor by only one parameter. Only a set of characteristics gives an understanding of what kind of chip it is. Narrowing down the processors to consider is very easy. AMD's modern ones include FX chips for the AM3+ platform and A10/8/6 hybrid solutions of the 6000 and 7000 series (plus Athlon X4) for FM2+. Intel has Haswell processors for the LGA1150 platform, Haswell-E (essentially one model) for LGA2011-v3 and the latest Skylake for LGA1151.

AMD processors

I repeat, the difficulty in choosing a processor lies in the fact that there are a lot of models on sale. You simply get confused in this variety of markings. AMD has hybrid processors A8 and A10. Both lines include only quad-core chips. But what's the difference? Let's talk about this.

Let's start with positioning. AMD FX processors are top chips for the AM3+ platform. Game games are assembled on their basis. system units and workstations. Hybrid processors (with built-in video) of the A-series, as well as Athlon X4 (without built-in graphics) are mid-class chips for the FM2+ platform.

The AMD FX series is divided into quad-core, six-core and eight-core models. All processors do not have a built-in graphics core. Therefore, for a complete assembly you will need either motherboard with built-in video, or a discrete 3D accelerator.