Page 1: Ivy Bridge Test and Review: Intel Core i7-3770K and All Core i5 Models
Meet the next tick: April 23, 2012 Intel announces new Ivy Bridge processors. A total of 14 new desktop and notebook models are introduced. Add to this eight varieties of chipsets that have already been partially presented, as well as five options wireless communication. In this article, we will take a closer look at desktop processors, and our second review is dedicated to mobile CPUs.
The third generation of the Intel Core microarchitecture is a "tick" from the Intel "tick-tock" model, that is, it implies a reduction in the technical process. In theory, Intel should have taken the micro-architecture of the previous generation "Sandy Bridge" and reduced the process technology. As a result, the CPUs would be made on a 22nm processor, incorporating the new Tri-Gate transistors from Intel.
But Intel decided to optimize the CPU a bit and make a number of improvements. Thus, Ivy Bridge processors are not just Sandy Bridge models with a smaller process technology. As emphasized in Intel's marketing materials, the new CPUs are "tick+". However, as we will see below, the improvements have mainly affected only the graphics core.
For tests, we received the following processors, which clearly show that Intel decided to change the naming scheme again.
- Intel Core i7-3770K
- Intel Core i5-3570K
- Intel Core i5-3550
- Intel Core i5-3450
The Intel Core i7-3770K is the new top model in the family, designed to replace the Core i7-2700K. The clock speeds are shown in the table below - they have shifted slightly compared to the previous generation. The Core i5-3570K has been announced to replace the very popular Core i5-2500K processor. Hyper-threading support not to be expected, as with all Core i5 models, but the CPU clock speeds are quite high. The third and fourth Core i5-3550 and Core i5-3450 models we received are the current processors entry level in the Ivy Bridge family (Core i3 models to be announced later). All Ivy Bridge processors have at their disposal a DDR3-1600 memory controller, which allows you to count on a slightly faster dual-channel memory interface.
Introducing the test: new flagship Intel for the desktop segment, the Core i7-3770K processor, as well as the younger models of the Ivy Bridge family.
The following table summarizes the performance of the desktop CPUs in our testing:
| Ivy Bridge Desktop Processors (Quad Core) | |||||
|---|---|---|---|---|---|
| Processor | Core i7-3770K | Core i5-3570K | Core i5-3550 | Core i5-3450 | For comparison: Core i7-2700K |
| Price | $313 10 900 rub. in Russia |
$212 6 900 rub. in Russia |
$194 6 300 rub. in Russia |
$174 5 700 rub. in Russia |
289 euros in Europe 10 300 rub. in Russia |
| Thermal Package (TDP) | 77 W | 77 W | 77 W | 77 W | 95 W |
| Kernels/ streams |
4 8 |
4 4 |
4 4 |
4 4 |
4 8 |
| CPU frequency | 3.5 GHz | 3.4 GHz | 3.3 GHz | 3.1 GHz | 3.5 GHz |
| Turbo 4 cores | 3.7 GHz | 3.6 GHz | 3.5 GHz | 3.3 GHz | 3.6 GHz |
| Turbo 2 cores | 3.9 GHz | 3.8GHz | 3.7 GHz | 3.5 GHz | 3.8GHz |
| Turbo 1 core | 3.9 GHz | 3.8GHz | 3.7 GHz | 3.5 GHz | 3.9 GHz |
| Memory interface | Two DDR3-1600 channels (Support Low Voltage) |
Two DDR3-1333 channels | |||
| L3 cache | 8 MB | 6 MB | 6 MB | 6 MB | 8 MB |
| Intel HD Graphics | HD 4000 | HD 4000 | HD 2500 | HD 2500 | HD 3000 |
| GPU frequency | 650 MHz | 650 MHz | 650 MHz | 850 MHz | |
| GPU Turbo frequency | 1150 MHz (Max: 1350 MHz) | 1100 MHz | 1100 MHz | 1350 MHz | |
| PCIe 3.0 | Yes | Yes | Yes | Yes | No |
| Intel Secure Key | Yes | Yes | Yes | Yes | No |
| OS Guard | Yes | Yes | Yes | Yes | No |
| vPro, VT-d, TXT, SIPP | No, only non-K models | Yes | No | Yes | |
| Unlocked multiplier | Yes | Yes | No | No | Yes |
In addition to the models listed, Intel also sells the Core i7-3770 processor. This model does not have an unlocked multiplier, and the frequencies are somewhat different: the Core i7-3770 has the same frequency response in Turbo modes as the Core i7-3700K processor, but the base frequency is 3.4 GHz. So the processor will be slower than the "K" model, but only if Intel Turbo Technology is turned off. Among the advantages of "non-K" processors, one can note support for vPro, VT-d, TXT and SIPP technologies - however, overclockers are unlikely to be interested in these functions.
Screenshots of CPU-Z processors i7-3770K, i5-3570K ...
...and i5-3550 and i5-3450
Intel processors The Core i5-3550 and i5-3450 use lower clock speeds (Turbo peaks at 3.7 and 3.5 GHz). In addition, these models do not use the "older" graphics core Intel HD Graphics 4000, but the "younger" version of the Intel HD Graphics 2500, which has reduced computing units. The "younger" i5-3450 processor also lacks support for vPro, VT-d, TXT, and SIPP.
In addition to the standard 77W TDP models, Intel also introduced four processors with a lower TDP: the Core i7-3770S and i7-3770-T are very close to the Core i7-3770 processors, although they differ in clock speeds, but the TDP is reduced to 65 or 45 W, reduced voltage and frequencies. The Core i5-3550S and i5-3450S processors have a TDP of 65W and match the Core i5-3550 and i5-3450 processors, but again, the voltage and frequencies are reduced.
If you've been following our news, you've probably noticed the message
The line of Core i5 processors from the global manufacturer Intel is represented on the market by a wide variety of models with different characteristics, so many users who are fond of overclocking the CPU face various problems when choosing this device.
First of all, this is due to the fact that not all novice IT specialists know what parameters to pay attention to when choosing a CPU. This article will tell you about all the nuances associated with choosing an Intel Core i5 processor and how to properly overclock it.
General information about Core i5 processor models
Intel processors Core third generations are represented on the market not by one model, but by a huge line designed for the average consumer who needs a reliable and productive processor at an affordable price. 3rd generation Intel Core processors combine second generation affordability with high specifications fourth generations. Thus, Intel Core i5 is the optimal solution for most personal computers.
CPU overclocking
In most modern models of Intel processors, including the Core i5, the innovative patented Turbo Boost technology is implemented, which allows you to overclock the processor in order to increase CPU performance. automatic mode. These processors have a microchip responsible for graphics processing.
All models of Intel Core i5 processors are available with clock speeds ranging from 1.2 to 3.6 GHz. The data transfer rate over the serial bus is 2.5 GP/s. Depending on the year of release, Core i5 processors can be assembled on crystals with a thickness of 32 or 45 nm, depending on the architecture implemented in the CPU. In addition, models are divided into dual- and quad-core with 256 KB L2 cache and 4 to 6 MB L3 cache.
According to many experts, you can significantly increase the performance of most Intel processors model range Core i5 through overclocking. Therefore, it is necessary to study in more detail the dependence of technical characteristics and overclocking.
Intel Core i5 Processor Architecture
Before starting to consider the characteristics of the CPU, it is necessary to understand how the processors are arranged. All Intel Core CPU models are based on the modern Sandy Bridge architecture, which provides processors with high performance and good technical characteristics. However, when it comes to overclocking, this architecture does not show the best results.
Sandy Bridge architecture features
The thing is that the manufacturer has blocked most of the key multipliers that affect the quality of overclocking in processors built on the Sandy Bridge architecture. Therefore, the maximum frequency increase threshold will not exceed 900 MHz, which is quite small from the point of view of professional specialists in the field of computer hardware.
However, Intel releases processors of the Core i5 family, in which, if necessary, you can independently adjust the value of multipliers. One of these CPUs is the Core i5-2500K model, which has gained great popularity among the masses. According to many IT people, the clock speed of these processors can be overclocked up to 5 GHz. Many users prefer this particular model also because it has a low cost. Thus, for only $200, you can get a processor whose performance will not be inferior to top models.
Ideal processor for overclocking
According to many experts, if you decide to purchase an Intel Core i5 processor for the purpose of further overclocking, then it is best to opt for the 2500K modification. This model has a low cost and allows you to significantly increase the performance of the CPU. According to some technical characteristics, it is not inferior even to Core i7 models, the cost of which is several times higher. Thus, the Intel Core i5-2500K is the perfect combination of performance, affordability and overclocking characteristics.
Of course, more modern and advanced technologies are implemented in Core i7, for example, the size of the third-level cache in them is 8 MB, versus 6 MB in Core i5, however, in practice, this difference will be practically not noticeable when performing most tasks.
General information about the Intel Core 2500K processor
Continuing consideration Intel specifications Core i5 2500K, one cannot fail to mention the adaptation of the processor to modern computer games. In addition, the CPU has four cores, which has already become a kind of standard in the manufacture of gaming processors, because a large multi-core processor is far from always a justified step, since multitasking has no effect on gaming processes and is, by and large, useless.
The iron factor
Starting to overclock your processor, you need to choose high-quality components for the CPU. The performance of the accompanying hardware should not be inferior to those of the processor, otherwise the experiment may not end very well or even lead to CPU failure. At Intel overclocking Core i5-2500K computer specifications should not be inferior to the following:
- a powerful processor cooling fan with a speed of at least 1800 per minute;
— graphics system Radeon HD 5870 and above;
- at least 4 GB random access memory class DDR3 and throughput 1600.
Besides system unit it is desirable to equip with additional coolers.
Factor software
According to some advanced users, overclocking the processor is something like an experiment, so the quality of the software also plays a big role. As for the operating system, Windows 7 and above is quite suitable for these purposes. In addition, it is very important that all devices have compatible drivers latest version.
Results require measurements
Performing such a complex procedure as overclocking a processor involves the use of special software, among which one of the most important is a utility for measuring the result. For this purpose, 3D Mark can do an excellent job. According to many experts, it displays up-to-date information about the performance of each core during overclocking, and also provides detailed information at the end of the process. comparative characteristic CPU performance before and after overclocking.
No less popular among many professional IT specialists is the SiSoft Sandra utility, which allows you to get detailed information not only about the performance of the chip, but also about each individual computer device. The program has a wide range of various tools, however, the most relevant of them are arithmetic and cryptographic tests.
Some experts argue that the processors of the Core i5 family are adapted to work with programs that work on the principle of a chess algorithm. One of these utilities is Fritz Chess Benchmark, which measures the performance of processor cores based on the processing speed of incoming commands and the time it takes to complete the tasks assigned to it.
Programs at your fingertips
Fritz Chess Benchmark is a professional software designed specifically for CPU overclocking and allows you to evaluate how successful CPU overclocking has been. However, in order to obtain more reliable information, it is recommended to use the set together with this utility. standard programs. For example, a conventional archiver that allows you to evaluate the speed of packing or unpacking data, as well as the stability of the processor during the execution of tasks.
In addition, it can be very effective graphics editor Photoshop, which creates a large load on the CPU during image processing and the use of various filters. If during image processing the processor normally copes with its tasks, then in this case the overclocking was successful and you don’t have to worry about the stability of the CPU.
What other software can I use when testing the performance of overclocked Intel Core i5 processors? The Linpack utility, which allows you to test hardware for the speed of solving algebraic equations, may be useful. In addition, any program that allows you to monitor changes in CPU temperature is useful. As such, Real Temp is quite suitable. Another and very useful utility is CPU-z, which provides a detailed report on the performance of all devices in the computer.
Game Factor
As mentioned earlier in this article, microprocessor intel Core i5-2500K belongs to the gaming type. This is also evidenced by high technical characteristics, for example, clock frequency, which in basic modifications of the CPU is 3 GHz. Therefore, when overclocking your processor, it is worth testing its performance in modern mode. computer games or using specialized software that simulates a computer environment. This can be done using the powerful FRAPS utility.
However, before proceeding directly to the test, you must first run the processor for default settings. By measuring the factory parameters, you can gradually increase the frequency of the multipliers. Read more about how to do this in the article.
BIOS Features
One of the easiest and most reliable ways to overclock a processor is to change the appropriate multiplier settings in BIOS. To do this, you need to restart the computer and enter the BIOS settings menu using the DEL key, which is set by default by the manufacturer. Some BIOS versions can use other keys or their combinations to enter, however, as a rule, a hint is displayed on the screen at the very beginning of the launch, which you can use.
After the BIOS starts, you need to go to the Frequency Control tab, which contains the control of the base multipliers, of which only two are important - the operating frequency of the central bus, measured in MHz, as well as its coefficient. It is worth noting that it is not recommended to adjust the bus frequency, since the architecture of the microchips implemented in the Intel Core i5 processors does not provide for manual adjustment of this parameter. In addition, increasing the frequency of the system bus will not give a noticeable increase in performance.
How to achieve a progressive increase in productivity?
You can achieve an increase in CPU performance by adjusting the bus frequency factor. In the Intel Core i5-2500K processors, the developers have provided for an increase in this parameter to 57. However, in order not to burn the processor during overclocking, it is not recommended to set the coefficient immediately to the maximum, but to increase it progressively. Experts claim that CPU overclocking you can start with the initial value of the coefficient value of 40. If the processor becomes unstable, it should be lowered.
According to some experts, to achieve more high performance when overclocking the processor, you can also use the CPU voltage control. However, the opinions of various experts on this issue differ. If you still decide to take advantage of this opportunity, then you should start with a voltage of 1.15 V. Many experts argue that this value is optimal for achieving the best results when overclocking the Intel Core i5-2500K processor.
Practical results
What maximum performance processor can be obtained by overclocking? How much will the installation of Intel Core i5 affect the gaming experience? According to IT-specialists, the results can exceed any expectations, so installation and overclocking justify themselves. During numerous experiments, it was found that the technical characteristics of the processor justify themselves in terms of reliability and protection against overheating.
In other words, having overclocked the processor to the maximum available frequency, its maximum temperature will not exceed the maximum allowable limit. Therefore, if necessary, users get the opportunity to control the values of the multipliers at their discretion. However, in this case, the operation of the system may be unstable. Its testing should take place simultaneously using several specialized utilities.
Processor overclocking is an opportunity to increase the performance of your computer without any financial costs. Each user gets the opportunity to get a powerful game processor, the cost of which does not exceed the budget options. With proper overclocking and judicious use of multiplier values, Intel performance Core i5 will be comparable to higher generation processors.
Introduction
Recently, we began to acquaint our readers with the results of a large-scale testing of modern processors carried out in our laboratory. To date, there are already two materials on our site covering this topic:
Thus, we have already tested all the desktop processors offered by AMD and Intel, the official cost of which is below the $200 mark. Now it's time to move on, and today we will meet with more expensive and more productive models. As mentioned earlier, two hundred dollars is the border, when crossing which the level of processor performance grows far less quickly than their price. Accordingly, the choice of expensive models when creating desktop computers general purpose, in our opinion, is by no means the most rational solution. Much higher efficiency in terms of final complex performance can be obtained by additional investments in the graphics subsystem or, for example, by switching the disk subsystem to use SSD.
But this does not mean at all that the purchase of processors belonging to the upper price category is a waste of money. There are numerous situations where processor power is very important, and increasing it gives good results. Usually, in this vein, we can talk about the areas of computer use, in which a significant part of the tasks being solved involve heavy computational work. With regard to home and semi-professional users, this, first of all, can be video processing - simple transcoding, non-linear video editing or creating visual effects. Also, various packages for three-dimensional modeling or systems for editing and mastering audio content create a fairly serious load. Some players do not disdain powerful processors, however, only that layer of them that already has graphics subsystems from the highest price segment at its disposal (otherwise it is more profitable to invest in updating a video card). In other words, expensive processors are not a luxury, but a quite demanded class of products, although not as popular as inexpensive or mid-budget processors.
It should be noted that the general structure of offers in the upper price segment is qualitatively different from what we saw in the lower and middle price segments. The main feature is the almost complete absence of AMD products among expensive processor models. AMD did not have to leave the top market sector of its own free will: while Intel engineers have been increasing the potential of their microprocessor architecture over the past few years, AMD has been “marking time”, in fact, only slowly improving technological production processes. As a result, today AMD simply cannot offer real competitors to Intel's older processors - so the main part of today's article will deal with Intel solutions, and Intel has spent a lot of effort on the development of high-performance solutions and can even offer a special platform as high-end solutions. However, about everything - in order, but we will start with those models of AMD processors that still managed to jump above their heads and get an objectively justified cost above the $ 200 mark.
High end AMD processors
And again, speaking of AMD processors, focused on the upper price category, we will have to focus on the Socket AM3 platform. The situation is such that so far AMD supports only one platform, for which processors are offered with completely different characteristics and different cost levels. This imposes certain restrictions on them, for example, all modern AMD processors are forced to be equipped with a dual-channel DDR2/DDR3 memory controller, however, for the models we are considering today, these restrictions cannot be called too strict.
The problem with AMD products is different - they are based on cores with the Stars microarchitecture, which has an IPC characteristic (number of instructions executed per clock) that is too low by modern standards. Therefore, even those AMD processors that operate at impressive frequencies, in reality, are not as fast as their rivals offered by Intel. In addition, AMD has not yet been able to introduce a modern process technology with 32-nm standards, which negatively affects the power consumption and heat dissipation of processors.
All this in total does not allow AMD to confidently declare itself in the upper price segment. And as a result, in today's material, this company is represented by only two models of processors that belong to a single family.
Phenom II X6. The family of six-core processors appeared in the ranks of AMD products not so long ago. For them, a special monolithic semiconductor crystal with improved energy efficiency was developed, which, nevertheless, is produced using a 45-nm process technology. Thanks to this introduction, the clock speeds of the older Phenom II X6 processors that fall into the upper price segment are 3.0-3.2 GHz. Additionally, these processors are equipped with Turbo Core technology, which raises the frequency to 3.5-3.6 GHz if at least three of the six cores are idle. Other characteristics of Phenom II X6 are quite typical: each of the cores is equipped with a 512-KB L2 cache, in addition, Phenom II X6 has a 6 MB L3 cache shared between all cores. As for the integrated memory controller, it can work with DDR2 or DDR3 memory clocked at frequencies up to 1600 MHz. It only remains to add that the senior representative of the lineup, Phenom II X6 1090T, belongs to the Black Edition modifications, which means that it can be overclocked by increasing the multiplier.
Finding Phenom II X6 processors among expensive processors at first glance, it seems quite justified, because Intel also offers its six-core processors only in the upper price segment, and besides, it rates them many times more expensive. However, after getting acquainted with the test results, it becomes clear that the prospects Socket platforms AM3 is not too rosy here, and it is not surprising that under the pressure of progress driven by Intel forces, AMD's six-core processors are rapidly depreciating. In other words, the Socket AM3 platform may soon leave the number of expensive solutions. Ahead of us is the release of only one new model in the Phenom II X6 series, and then these processors will gradually fall to the middle level, freeing up space for promising products with the Bulldozer microarchitecture, which, according to AMD's promises, will appear around the second quarter of next year.
Nevertheless, Socket AM3 can still be considered as a platform top level. There are expensive and feature-rich motherboards on the market for Socket processors AM3, which do not limit the freedom of creativity in building high-performance systems. For example, there are even boards aimed at fanatical players and allowing you to install three or four video cards at once.
The following table shows the characteristics of the few modifications of AMD processors that can be attributed to the upper price level and which took part in the tests.
High end Intel processors
It seems that by this moment it is already quite clear that Intel plays the first violin in the upper price segment. And, surprisingly, the offers of this company are very diverse not only in terms of the proposed processors, but also in terms of platforms. As the main top-level solution, the manufacturer positions the special LGA1366 platform. But, despite this, having set aside more than $200 for a purchase, today you can buy an Intel processor not only in LGA1366, but also in LGA1156, and even in LGA775. On the one hand, internal competition and the presence of all three platforms in the upper price segment is a symptom of the ill-conceived marketing policy, but in this case, two of the three platforms may well coexist side by side due to their unique features.
Actually, the advantages of LGA1366 solutions are easily noticeable. Exclusively for this platform, the manufacturer offers six-core processors, and LGA1366-quad-core processors differ from their counterparts for LGA1156 by the presence of a three-channel DDR3 SDRAM controller. However, despite its "elite" status, this platform is not without noticeable disadvantages. Having come to desktops from servers and workstations, it assumes not the most successful system architecture: the PCI Express graphics bus controller is placed in the logic set, as a result of which there is a need for an additional high-speed QPI bus connecting the processor and chipset. Adding two copies of the QPI controller to the main chips has led to the fact that the power consumption of LGA1366 systems is significantly higher than, for example, systems with LGA1156 processors.
Core i7-900. This model range combines rather heterogeneous processors intended for use as part of the LGA1366 platform. In fact, this family includes two types of products with the Nehalem microarchitecture. The first is quad-core processors based on 45nm chips, the second is processors with six cores, which are based on crystals manufactured using 32nm technology. Quad-core processors operate at frequencies of 3.06-3.2 GHz, have an 8-MB cache in the third level and support three-channel DDR3 memory. They also have all modern technologies aimed at increasing performance - virtual multi-core Hyper-Threading and automatic overclocking turbo boost. Six-core processors have higher frequencies of 3.2-3.33 GHz, and their cache size has been increased to 12 MB. Otherwise, in terms of performance, they are similar to quad-core, also have a three-channel memory controller and support Hyper-Threading and Turbo Boost technologies.
It should be noted that in terms of model numbers, there is a very thin line between 4- and 6-core processors for LGA1366 systems. So, models with numbers up to i7-960 inclusive are quad-core models, and numbering of six-core products starts with i7-970. At the same time, the senior representative of the family, Core i7-980X, additionally bears the proud title of Extreme Edition, which means that it can be overclocked by changing the multiplier.
It is curious that of all the platforms currently on the market, it is LGA1366 that has a chance to get the longest life: its replacement may not occur until the end of next year. That is why LGA1366 can be recommended to fans of frequent computer upgrades.
The LGA1156 platform seems to be a more mature solution than LGA1366, since it appeared on the market a little later, and its structure is clearly more optimal due to the fact that the PCI Express graphics bus controller is transferred to the processor. However, older processor models for this platform have more low frequencies, can boast only a dual-channel memory controller, and the most annoying thing is that they have only a single PCI Express x16 bus, which, although it can be split into two parts, limits the performance of multi-GPU video subsystems. As a result, although many LGA1156 processors are rightfully classified as top-level solutions, they are one step lower in the hierarchy than their LGA1366 counterparts.
Core i5-600. Dual-core LGA1156 processors built on a 32-nm semiconductor crystal and containing an additional second crystal - a graphics core. Such processors are perceived by the operating system as quad-core, as they support Hyper-Threading technology, thanks to which each of the cores can perform two computing processes simultaneously. The frequencies of the processors Core i5-600, related to the proposals of the upper level, are in the range from 3.46 to 3.6 GHz, and the amount of cache in the third level is 4 MB. In addition, these processors also support Turbo Boost technology, which can automatically increase the processor frequency when one core is idle. The memory controller built into the processor can work with dual-channel DDR3 SDRAM at 1067 or 1333 MHz.
Core i5-700. The junior family of quad-core processors for LGA1156 systems has its representatives both in the middle and in the upper price category. The latter include the Core i5-760, a processor based on a 45nm crystal clocked at 2.8GHz. It supports Turbo Boost technology, but Hyper-Threading does not work in it, as a result of which, like the Core i5-600, it appears in the operating system as four cores. The Core i5-760 has 8 MB of L3 cache, and the built-in memory controller is designed to work with dual-channel DDR3-1067 and DDR3-1333. It should be noted that, unlike the Core i5 of the six hundredth series, the models of the seven hundredth series do not have an integrated graphics core.
Core i7-800. These are full-fledged quad-core LGA1156 processors, full-fledged masters of the upper market segment. Although they are based on the same semiconductor crystal as the Core i5-700, unlike their younger counterparts, they have higher frequencies in the range of 2.8-3.06 GHz and support not only Turbo Boost, but also technology hyperthreading. As a result, the Core i7-800 appears to the operating system as a processor with eight cores. In addition, the Core i7-800 boasts a more efficient memory controller, which, in addition to DDR3-1067 and DDR3-1333, is also capable of supporting DDR3-1600. However, the L3 cache size of the Core i7-800 is the same as that of the Core i5-700 - 8 MB.
On this moment the LGA1156 platform looks very attractive, but do not forget that its decline is very close. In two months it will be replaced by the new LGA1155 platform, which supports processors with a more advanced microarchitecture and operates at higher frequencies. More high-speed processors in the LGA1156 version than there are in stores now will never be, and, unfortunately, there is no compatibility between LGA1156 and LGA1155.
As for the LGA775 platform, its presence in the upper market segment at the end of 2010 is completely illogical. It can only be regarded as a tribute to the memory of very successful processors Core rulers 2 Duo and Core 2 Quad, thanks to which Intel back in 2006 was able to take the lead from a competitor in pursuit of the performance of serial processors. In other words, Core microarchitecture, which underlies LGA775 processors, is no less outdated than the Stars microarchitecture. But at the same time, Intel, unlike AMD, does not even try to increase the clock speeds of its models, offering two-three-year-old products at high prices.
Core2 Duo. A series of dual-core LGA775-processors based on a 45-nm core turned out to be represented in the upper price category, however, only by a single model - Core 2 Duo E8600. This processor is rated at 3.33GHz, uses a 333MHz system bus, and has a shared 6MB L2 cache.
Core 2 Quad. Quad-core processors in the LGA775 version are presented somewhat better in the upper price segment. Intel offers models from 2.83 to 3.0 GHz using a 333 MHz bus. Recall that the quad-core for the LGA775 platform is a gluing of two dual-core semiconductor crystals, made inside one processor package. Therefore, the second-level cache, the total size of which is 12 MB, in Core 2 Quad processors is divided in half between pairs of computing cores.
Let's summarize by giving a table with the characteristics of all current Intel processors belonging to the upper price category.
How We Tested
Simultaneous testing of a large number of processors is a rather difficult task, not only associated with serious labor costs, but requiring simultaneous access to dozens at once various models processors. Therefore, most of the databases of test results that can be found on the net either fill up gradually, which entails either differences in the platforms, drivers and software versions used, or are run on an outdated system that has not been updated for a long time. We have tried to work around these problems - our testing is done now and immediately. The tests used modern components with up-to-date firmware and latest versions drivers and operating system with all updates installed.
More specifically, the following hardware and software components were used in the test systems:
Processors:
AMD Phenom II X6 1075T (Thuban, 6 cores, 3.0 GHz, 6 MB L3);
AMD Phenom II X6 1090T (Thuban, 6 cores, 3.2 GHz, 6 MB L3);
Intel Core 2 Duo E8600 (Wolfdale, 2 cores, 3.33 GHz, 6 MB L2);
Intel Core 2 Quad Q9550 (Yorkfield, 4 cores, 2.83 GHz, 12 MB L2);
Intel Core 2 Quad Q9650 (Yorkfield, 4 cores, 3.00 GHz, 12 MB L2);
Intel Core i5-670 (Clarkdale, 2 cores, 3.46 GHz, 4 MB L3);
Intel Core i5-680 (Clarkdale, 2 cores, 3.6 GHz, 4 MB L3);
Intel Core i5-760 (Lynnfiled, 4 cores, 2.8 GHz, 8 MB L3);
Intel Core i7-860 (Lynnfiled, 4 cores, 2.8 GHz, 8 MB L3);
Intel Core i7-870 (Lynnfiled, 4 cores, 2.93 GHz, 8 MB L3);
Intel Core i7-880 (Lynnfiled, 4 cores, 3.06 GHz, 8 MB L3);
Intel Core i7-950 (Bloomfiled, 4 cores, 3.06 GHz, 8 MB L3);
Intel Core i7-960 (Bloomfiled, 4 cores, 3.2 GHz, 8 MB L3);
Intel Core i7-970 (Gulftown, 6 cores, 3.2 GHz, 12 MB L3);
Intel Core i7-980X (Gulftown, 6 cores, 3.33 GHz, 12 MB L3).
Motherboards:
ASUS Crosshair IV Formula (Socket AM3, AMD 890FX + SB850, DDR3 SDRAM);
ASUS P5Q3 (LGA775, Intel P45 Express, DDR3 SDRAM);
ASUS P7P55D Premium (LGA1156, Intel P55 Express);
Gigabyte X58A-UD5 (LGA1366, Intel X58 Express).
Memory:
2 x 2 GB DDR3 SDRAM (Kingston KHX1600C8D3K2/4GX):
DDR3-1333 9-9-9-27 when using Core 2 Duo E8600, Core 2 Quad Q9550, Core 2 Quad Q9650, Core i5-670, Core i5-680, and Core i5-760;
DDR3-1600 9-9-9-27 when using Phenom II X6 1075T, Phenom II X6 1090T, Core i7-860, Core i7-870, Core i7-880;
3 x 2 GB, DDR3 SDRAM (Crucial BL3KIT25664TG1608):
DDR3-1600 9-9-9-27 when using Core i7-950, Core i7-960, Core i7-970, and Core i7-980X processors.
Graphics card: ATI Radeon HD 5870.
Hard drive: Kingston SNVP325-S2/128GB.
Power supply: Tagan TG880-U33II (880 W).
Operating system: Microsoft Windows 7 Ultimate x64.
Drivers:
Intel Chipset Driver 9.1.1.1025;
ATI Catalyst 10.9 Display Driver.
Performance
Overall Performance
To assess the performance of processors in common tasks, we traditionally use the SYSmark 2007 test, which simulates the user's work in common office programs and applications for creating and processing digital content. The idea of the test is very simple: it produces a single metric that characterizes the weighted average speed of the computer.
The picture is very revealing. New Intel microarchitectures are ready to actually prove their superiority over old designs: LGA1156 and LGA1366 processors boast significantly more high speed work rather than offerings for LGA775 and Socket AM3 systems. Even dual-core processors like the Core i5-670 easily outperform AMD's six-core processors in SYSmark 2007. So the cheapness of the Phenom II X6 is, in general, not surprising: in practical use in everyday tasks, they can only compete with Core 2 Duo and Core 2 Quad processors.
Yes, and in general dual-core Core processors The i5-670 and Core i5-680 can boast of quite impressive performance according to SYSmark 2007. Which, I must say, is not so strange. Firstly, a large number of commonly used applications make full use of only two cores, and secondly, dual core processors The Core i5-600 families have high clock speeds, and thirdly, they have Hyper-Threading technology at their disposal, which is very useful in multi-threaded workloads.
However, the leaders in SYSmark 2007 are still processors core series i7, targeting both LGA1366 and LGA1156 systems. But at the same time, the LGA1366 platform justifies its higher level, and almost all LGA1366 processors turn out to be faster than their LGA1156 counterparts. For example, the older Core i7 of the 800th series is only slightly ahead of the Core i7-950, which, however, can hardly be called a surprise if we compare their prices.
Let's supplement the above diagram with a table with more detailed SYSmark 2007 results, sorted by application type:
As you can see, the high results of the Core i5-670 and Core i5-680 dual-core processors are explained by their fast work V office applications and in image processing. Working with video content or 3D modeling systems brings multi-core products to the first place in terms of performance.
Gaming Performance
As is known, real performance platforms equipped with powerful processors in games in the vast majority of modern games is determined by the graphics subsystem. That is why, when testing processors, we try to conduct tests in such a way as to take the load off the video card as much as possible: tests are carried out without anti-aliasing enabled and with far from the highest resolutions set. That is, the results obtained make it possible to assess not so much the level of speed achievable in systems with modern video cards, but how well processors perform under gaming load in general. Therefore, based on the above results, it is quite possible to speculate about how processors will behave in the near future, when new generations of graphics accelerators appear on the market.
The results were quite expected. In most games powerful processors, in general, are not needed, since performance is likely to be limited by the capabilities of the video subsystem. Moreover, the measures taken by us did not help at all, even at low resolutions and without anti-aliasing, the Core i7 series processors with different clock speeds provide very similar results. In other words, the players are really powerful models(such as six-core Intel) may be of interest only in exceptional cases - they reveal their advantages in a relatively small number of games and in the presence of a multi-core system graphics system last generation. In general, we can say that any of the processors belonging to the Core i7 series today has more than enough performance for a gaming system, even at the top level.
Archiving and Encryption
To measure the speed of processors when compressing information, we use the WinRAR archiver, with which we archive a folder with various files with a total volume of 560 MB with the maximum compression ratio.
Many factors influence the speed of work in WinRAR: clock frequency, number of cores, cache memory size... However, in our case, the diagram looks quite simple. Representatives of LGA1366 are located on top, followed by LGA1156-quad-core, and then quad-core by LGA775. It adds some spice and at the same time disappoints only the speed of the Phenom II X6, these processors are able to outperform only dual-core models.
The encryption performance of processors is measured by the built-in test of the popular TrueCrypt cryptographic utility. It should be noted that it is not only capable of efficiently loading any number of cores, but also supports a new specialized AES instruction set.
Encryption is a highly parallelizable process, which is why AMD and Intel processors with six cores are located at the top of the diagram. At the same time, the result of six-core Intel is much higher, which is explained by their progressive microarchitecture and their support for a specialized AES instruction set.
Image editing
Performance measurement in Adobe Photoshop we conduct using our own test, which is a creatively redesigned Retouch Artists Photoshop Speed Test, which includes a typical processing of four 10-megapixel images taken with a digital camera.
There are no surprises here either. LGA1366 and LGA1156 processors are located on the diagram almost in full accordance with their model numbers. LGA775-oldies once again failed to compete with newer products, and AMD's six-core processors generally fell into the outsiders.
Audio and video transcoding
When testing the speed of transcoding audio, the Apple iTunes utility is used, with the help of which the contents of a CD are converted to AAC format.
Apple iTunes is an application that uses only two processor cores, so you should not be surprised at such a strange arrangement of models on the diagram. Actually, that is why the leaders are the Core i5 dual-core series, which have high clock speeds.
The x264 HD test is used to measure the speed of transcoding video to H.264, based on measuring the processing time of the original video in MPEG-2 format, recorded in 720p resolution with a stream of 4 Mbps. It should be noted that the results of this test are of great practical importance, since the x264 codec used in it is the basis of numerous popular transcoding utilities, such as HandBrake, MeGUI, VirtualDub and others.
We saw earlier that AMD processors do a pretty good job of transcoding video using the x264 codec. However, when comparing processors of the upper price category, it turns out that Phenom II X6 not only lags significantly behind six-core Intel processors, but also loses out to many quad-core Core i7 processors with Hyper-Threading technology.
performance in Adobe Premiere The Pro is tested by measuring the rendering time to H.264 Blu-Ray of a project containing 1080p25 HDV footage with various effects applied.
Qualitatively, the results of testing video encoding speed in Adobe Media Encoder do not differ much from the picture observed when using the x264 codec. Summing up, it should be emphasized that the six-core processors of the Core i7 series are a real find for users who are actively involved in processing video content. The use of a Core i7-970 or Core i7-980X in such a system instead of an expensive quad-core processor can increase performance by at least one and a half times. Which, however, is not surprising: the processes associated with video processing are perfectly parallelized.
Mathematical calculations
Performance in the Wolfram Mathematica math package is measured by running the standard MathematicaMark7 test.
Quite a typical state of affairs for today's test. And once again we have to state that the Core i7 processors do not have worthy competitors and are forced to compete with each other, and the balance of power between them also turns out to be expected and correlates well with the official Intel price list.
final rendering
Final rendering speed testing in Maxon Cinema 4D is performed using a specialized Cinebench test.
More cores - higher performance. This principle can be clearly seen when testing processors of the upper price category, if we assume that support for Hyper-Threading technology makes it possible to count each physical processor core as one and a half.
Rendering performance in Autodesk 3ds max 2011 using both Scanline and mental ray we measure using the specialized SPECapc test.
The relative rendering speed in 3ds max 2011 is similar to the results in Cinebench. Once again, we are convinced that AMD six-cores can compete only with Intel quad-cores, and LGA775 processors are hopelessly outdated.
Energy consumption
Performance isn't the only practical feature that potential buyers of high-end processors might be interested in. In many situations, their energy consumption is also important, which is directly related not only to the amounts in the electricity bills. The same parameter imposes requirements when choosing cases, power supplies and cooling systems. That is why we supplemented the performance tests with power consumption tests.
The following graphs show the total consumption of systems (without monitor) measured "after" the power supply, which is the sum of the power consumption of all components involved in the system. The efficiency of the power supply itself is not taken into account in this case. During the measurements, the load on the processors was created by the 64-bit version of the LinX 0.6.4 utility. In addition, in order to correctly assess idle power consumption, we activated all available energy-saving technologies: C1E, AMD Cool "n" Quiet and Enhanced Intel SpeedStep.
In the idle state, all the systems participating in the testing were divided into three groups. The first group, which includes the most economical models, includes the LGA1156 platform in its entirety. The second group, which includes processor models with slightly worse energy characteristics, is populated by products for LGA775 and Socket AM3. But the LGA1366 platform turned out to have a much higher idle consumption, which, in general, is not a surprise and is explained by the peculiarities of its architecture.
Under load, the situation changes somewhat, although the high power consumption of LGA1366 processors does not disappear anywhere. Systems built on these processors turn out to be significantly more “gluttonous” even under load. The dual-core Core 2 Duo and Core i5 processors can boast of the lowest power consumption, which is also quite natural. As for the power consumption of the Socket AM3 platform and the Phenom II X6 processors, their result is at the level of the older Intel quad-core LGA1156 processors.
Despite the fact that a fairly large number of models took part in today's tests of processors of the upper price category, we did not witness heated battles. Unlike the mid-range and low-end, the upper price range is very stretched, and the processors in it vary greatly in cost and rarely directly compete with each other. In other words, it is quite possible to choose an expensive processor based on the price. General rule here it is: more expensive models provide better performance.
However, there are exceptions to this rule, as to any other. Firstly, if you are looking at processors more expensive than $200, you should immediately throw out all LGA775 models from the sphere of attention. They are clearly outdated and absolutely nothing can oppose newer products, the performance of which is much higher.
Secondly, one should be wary of the expensive dual-core processors presented in the Core i5 series. Most of today's resource-intensive tasks, for which it makes sense to choose a more powerful processor, are perfectly able to parallelize. For this reason, quad-core processors in general look clearly preferable, and the older dual-core Core i5-670 and Core i5-680 models cannot be saved even by a high clock speed. So such processors can be of interest only in two cases: either for use in some specific applications, which, for some reason, have advantages multi-core processors ignored, or in cases where the buyer prioritizes not performance, but efficiency.
To further illustrate the ratio of consumer qualities of processors belonging to the upper price category, we made a diagram showing the correspondence between average performance and cost.
It is easy to see that in terms of price-performance ratio, one of the most best processors upper price category is Core i7-950. At an affordable price (for a high-end product), this quad-core processor is not only faster than many rivals. Its advantage is that it is intended for the LGA1366 platform, which, firstly, is promising, and, secondly, has all the necessary tools to create arbitrarily fast graphics subsystems.
True, motherboards with an LGA1366 connector are quite expensive, and besides, a system with an LGA1366 processor will be, to put it mildly, completely uneconomical. If these factors hold you back from building a computer based on an LGA1366 processor, take a look at Core i7-870. It is quite capable of competing in speed with the Core i7-950, but at the same time it is much more economical and does not require an expensive motherboard. True, in this case, certain limitations arise when building video subsystems with several video cards, but, obviously, not everyone is concerned about them.
Six-core processors also deserve attention. Cheap - Phenom II X6 and expensive - Core i7-970 and Core i7-980X. Of course, we would not recommend them for ordinary high-performance home computers, the average ratio of their performance and price looks like ordinary users not in the most attractive way. But there are a large number of specific applications where six full-fledged cores can be very useful. This is all kinds of high resolution video processing, final rendering and some other similar computationally intensive applications. So if you are going to use a computer to solve such problems, feel free to bet on one of the six-core.
Other materials on this topic
Big test of processors. Part 1: Budget processors
Big test of processors. Part 2: Middle segment
Processor Intel Core i5-3340S Ivy Bridge (2800MHz, LGA1155, L3 6144 KB, L2 1024 KB)
3340S . The presented processor uses the Ivy Bridge Core, the CPU is built using the 22 nm process technology. The number of cores in this processor is 4, and the clock frequency is 2800MHz. This Intel-branded CPU uses a socket (connector) to connect to the (mother) system board LGA1155. The volume of the cache memory of the 3rd level is 6144 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 25.6 Gb/s. Manufacturer - by Intel this processor is presented as a computing processor for level systems: Desktop PCs. Found information about the start date of production (or release) of the processor: Q3 2013.
Processor Intel Core i5-3570S Ivy Bridge (3100MHz, LGA1155, L3 6144 KB, L2 1024 KB)
A processor from Intel from the Core i5 lineup with a processor number: 3570S . The presented processor uses the Ivy Bridge Core, the CPU is built using the 22 nm process technology. The number of cores in this processor is 4, and the clock frequency is 3100MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1155 (motherboard) motherboard. The volume of the cache memory of the 3rd level is 6144 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 25.6 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: Q2 2012.
Processor Intel Core i5-3475S Ivy Bridge (2900MHz, LGA1155, L3 6144 KB, L2 1024 KB)
A processor from Intel from the Core i5 lineup with a processor number: 3475S . The presented processor uses the Ivy Bridge Core, the CPU is built using the 22 nm process technology. The number of cores in this processor is 4, and the clock frequency is 2900MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1155 (motherboard) motherboard. The volume of the cache memory of the 3rd level is 6144 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 25.6 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: Q2 2012.
Processor Intel Core i5-655K Clarkdale (3200MHz, LGA1156, L3 4096 KB, L2 512 KB)
A processor from Intel from the Core i5 lineup with a processor number: 655K . The presented processor uses the Clarkdale Core, the CPU is built using the 32 nm process technology. The number of cores in this processor is 2, and the clock frequency is 3200MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1156 (motherboard) motherboard. The level 3 cache memory is 4096 KB, while the level 2 cache memory is 512 KB. Information about max. memory bandwidth: 21 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: Q2 2010.
Processor Intel Core i5-661 Clarkdale (3333MHz, LGA1156, L3 4096 KB, L2 512 KB)
A processor from Intel from the Core i5 lineup with a processor number: 661 . The presented processor uses the Clarkdale Core, the CPU is built using the 32 nm process technology. The number of cores in this processor is 2, and the clock frequency is 3333MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1156 (motherboard) motherboard. The level 3 cache memory is 4096 KB, while the level 2 cache memory is 512 KB. Information about max. memory bandwidth: 21 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: 1 quarter 2010.
Processor Intel Core i5-750S Lynnfield (2400MHz, LGA1156, L3 8192 KB, L2 1024 KB)
A processor from Intel from the Core i5 lineup with a processor number: 750S . The presented processor uses Lynnfield core, The CPU is built according to the 45 nm process technology. The number of cores in this processor is 4, and the clock speed is 2400MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1156 (motherboard) motherboard. The volume of the cache memory of the 3rd level is 8192 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 21 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: 1 quarter 2010.
Processor Intel Core i5-2500S Sandy Bridge (2700MHz, LGA1155, L3 6144 KB, L2 1024 KB)
A processor from Intel from the Core i5 lineup with a processor number: 2500S . The presented processor uses the Sandy Bridge Core, the CPU is built using the 32 nm process technology. The number of cores in this processor is 4, and the clock speed is 2700MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1155 (motherboard) motherboard. The volume of the cache memory of the 3rd level is 6144 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 21 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: 1 quarter 2011.
Processor Intel Core i5-2500T Sandy Bridge (2300MHz, LGA1155, L3 6144 KB, L2 1024 KB)
A processor from Intel from the Core i5 lineup with a processor number: 2500T . The presented processor uses the Sandy Bridge Core, the CPU is built using the 32 nm process technology. The number of cores in this processor is 4, and the clock frequency is 2300MHz. This Intel-branded CPU uses a socket (connector) to connect to the LGA1155 (motherboard) motherboard. The volume of the cache memory of the 3rd level is 6144 KB, in turn, the cache memory of the 2nd level is 1024 KB. Information about max. memory bandwidth: 21 Gb/s. The manufacturer, Intel, presents this processor as a computing processor for systems of the level: Desktop PCs. Found information about the start date of production (or release) of the processor: 1 quarter 2011.
