Introduction More recently, it seemed to us that at the beginning of 2008 the main "hot" topic of our publications would be the comparison of new AMD processors Phenom with updated quad-core Intel Penryn processors manufactured using a 45nm process. However, these expectations are not destined to come true, and the fault lies with both AMD and Intel. Indeed, to date, AMD has not been able to offer serial quad-core processors operating at decent frequencies. The proposed Phenom models, on the other hand, show disastrous results even in comparison with Intel's previous generation quad-core CPUs, not to mention the more advanced new processors. It is quite logical that in the light of the revealed lack of worthy competitors for quite successfully sold Core 2 Quad processors on old 65-nm cores, Intel has lost incentives to update its line of quad-core processors as soon as possible. Therefore, the release of new CPUs in the Core 2 Quad line, known today under the code name Yorkfield, has been postponed indefinitely: at least until February or March. And although Intel hides behind a message about a problem found in promising processors caused by interference in the 1333-MHz front-side bus that occurs when they are used in hypothetical boards with a four-layer PCB design, it looks completely unconvincing. We are forced to state a sad result: it has become completely pointless to compare Phenom and Penryn, because the first is uncompetitive, and the second is still illusory and does not intend to lose the uncertain status of a promising product.
But, nevertheless, topics worthy of our attention can be found in today's processor market. Despite the fact that Intel decided to postpone the release of quad-core processors based on 45-nm cores, the line of dual-core Core2 Duo CPUs will still be updated. In the coming days, three new processors belonging to this model range and codenamed Wolfdale should be announced: Core 2 Duo E8500, E8400 and E8200. These processors are based on a redesigned 45nm core and belong to the same Penryn family as the deferred Yorkfield. The appearance of serial Wolfdales cannot be ignored: these processors promise to take the performance of Intel's dual-core offerings to a new level, because they have higher frequencies, a larger L2 cache, as well as other improvements. At the same time, which is especially nice, their price is set at the same level as the old Core 2 Duo.
Thus, for the second half of January, Intel has planned a massive update of its dual-core offerings in the price range from $160 to $260. It was this event that became the main topic for our new article, in which we will introduce you to what should be expected in reality from such promising new products aimed at using in desktop computers middle level.
Wolfdale processors: Core 2 Duo E8500, E8400 and E8200
So, Wolfdale is the code name for dual-core processors in the Penryn family. Like the delayed quad-core Yorkfield processors, Wolfdale processors are manufactured using a 45-nm process. Moreover, Yorkfield and Wolfdale are based on exactly the same semiconductor crystals: Yorkfield, according to the established tradition, is a gluing of two dual-core Wolfdale crystals, made in one processor package. Thus, Wolfdale can be considered as the basic building material for the formation of the entire Penryn family, which is why it is interesting in itself.The core of Wolfdale processors has an area of 107 square meters. mm and consists of 410 million transistors. These figures clearly suggest that the Wolfdale has made significant changes compared to Conroe's 65 nm predecessor, which contained 291 million transistors. Actually, this can be seen from the photo of the Wolfdale and Conroe cores: the layout of the functional blocks has changed somewhat.
Left - Wolfdale, right - Conroe (images not scaled)
Thus, the Wolfdale core is not just the Conroe core reduced due to the transition to a more advanced technical process. In the new processors, Intel engineers have made a number of improvements (more details about the features of the Penryn family of processors can be found in our material "").
Based on the new 45nm cores, Wolfdale's line of dual-core processors announced these days will initially include three models of Core 2 Duo processors: E8500, E8400 and E8200 clocked at 3.16, 3.0 and 2.66 GHz respectively. In addition, a model with the number E8190 will be available, similar to the Core 2 Duo E8200, but devoid of virtualization technology. Later, they will be joined by another, fifth, Core 2 Duo E8300 processor with a frequency of 2.83 GHz, but this will happen no earlier than the second quarter of this year.
A complete picture of serial Core 2 Duo with 45nm cores can be obtained from the table below.
The technical information indicated in the table must be accompanied by at least important information about the manufacturer's selling prices for new CPUs:
Core 2 Duo E8500 - $266
Core 2 Duo E8400 - $183
Core 2 Duo E8200 - $163
Core 2 Duo E8190 - $163
It's nice to see that Intel continues to stick to the user-approved pricing policy, where new processors are sold at the same cost as the old ones, evolutionarily ousting them from the market. This time Core 2 Duo E8500 replaces Core 2 Duo E6850, Core 2 Duo E8400 replaces Core 2 Duo E6770, and Core 2 Duo E6550 gives way to Core 2 Duo E8200. In other words, starting from the next few days, buyers of dual-core CPUs will have the opportunity to purchase more advanced and high-frequency processors at the old price.
Let's take a look at the Wolfdale processors themselves.
As you can see from the photo, the new processors with 45nm cores have almost the same appearance, as their 65 nm predecessors.
Left - Wolfdale, right - Conroe
However, the location of attachments on the belly of dual-core CPUs of different generations is different.
The CPU-Z diagnostic utility is already familiar with new processors. There are no problems with the correct identification of the Core 2 Duo E8500, E8400 and E8200.
Please note that our test samples of the new processors are based on the cores of far from the first revision C0, and it is this revision that will go into production models.
Only a single comment remains to be added to the information on the screenshot. Wolfdale processors have received support for fractional multipliers, which gives Intel the opportunity to make the clock frequency grid denser. This is exactly what we see on the example of the Core 2 Duo E8500 - this processor has a multiplier of 9.5. It should be noted that for the normal functioning of such a CPU, support for fractional multipliers from the BIOS is required. motherboard. However, in the near future, all leading motherboard manufacturers should release the corresponding updates.
How We Tested
To study the performance of the new Core 2 Duo E8500, E8400, and E8200 processors and compare them to their predecessors and competitors, we built several systems that included the following hardware.AMD platform:
CPU: AMD Athlon 64 X2 6400+ (Socket AM2, 3.0 GHz, 2x1024 KB L2, Windsor core).
Motherboard: ASUS M2R32-MVP (Socket AM2, AMD 580X chipset).
Memory: ).
Graphic card:
Disk subsystem:
Operating system:
Intel platform:
Processors:
Intel Core 2 Duo E8500 (LGA775, 3.16 GHz, 1333 MHz FSB, 6 MB L2, Wolfdale core);
Intel Core 2 Duo E8400 (LGA775, 3.0 GHz, 1333 MHz FSB, 6 MB L2, Wolfdale core);
Intel Core 2 Duo E8200 (LGA775, 2.66 GHz, 1333 MHz FSB, 6 MB L2, Wolfdale core);
Intel Core 2 Duo E6850 (LGA775, 3.0 GHz, 1333 MHz FSB, 4 MB L2, Conroe core);
Intel Core 2 Duo E6750 (LGA775, 2.66 GHz, 1333 MHz FSB, 4 MB L2, Conroe core).
Motherboard: ASUS P5E (LGA775, Intel X38, DDR2 SDRAM).
Memory: 2 GB DDR2-800 with 4-4-4-12-1T timings ( Corsair Dominator TWIN2X2048-10000C5DF).
Graphic card: OCZ GeForce 8800GTX (PCI-E x16).
Disk subsystem: Western Digital WD1500AHFD (SATA150).
Operating system: Microsoft Windows Vista x86.
We especially note that the ASUS P5E motherboard with BIOS version 0502 that we used to test Wolfdale processors fully supports them, allowing you to change the multiplier of these CPUs in 0.5 steps.
Performance
General performanceThe SYSmark 2007 test we have chosen uses typical performance scenarios in the most common real-world applications to determine performance.
SYSMark 2007 reveals on average about a 4% advantage for Wolfdale processors over Conroe processors running at similar clock speeds. However, due to the fact that Intel has increased the frequency of its processors in the updated CPU line, the older Wolfdale model is ahead of the older Conroe model by 7%. The cost of these processors of different generations according to the official Intel price list is the same.
An analysis of the intermediate results of SYSMark 2007 shows that the new processors provide the greatest increase in performance in a scenario in which the preparation of a training website containing a variety of media content is simulated. This scenario uses the following applications: Adobe Illustrator cs2, Adobe Photoshop CS2, Macromedia Flash 8 and Microsoft PowerPoint 2003. The smallest difference in performance between Core 2 Duo on 45nm and 65nm cores is observed in the production and processing of videos, which involve Adobe after effects 7, Adobe Illustrator CS2, Adobe Photoshop CS2, Microsoft Windows Media Encoder 9 and Sony Vegas 7.
3D games
Players should take the introduction of the new Core 2 Duo E8000 series processors with great enthusiasm. As you know, the speed of gaming applications responds well to changing the size of the cache memory, which is noted in this case. In some games, the youngest from Wolfdale, the Core 2 Duo E8200, even manages to outperform the former top dual-core model E6850 on a 65-nm core. The older AMD dual-core processor, Athlon 64 X2 6400+, which used to look bad in games before, is now in a deep knockout. It loses significantly in performance even to the younger representative of the Wolfdale line.
Media content encoding
The state of affairs is quite expected: the superiority of the Core 2 Duo E8000 family over its predecessors in the face of the Core 2 Duo E6000 is approximately at the same level as in other tests. Although this picture may change radically soon: codecs are among the applications that should benefit significantly from optimization for the SSE4 instruction set that appeared in the E8000 processor line. So for the time being, it is premature to draw any final conclusions about the work of Wolfdale in this group of tasks.
final rendering
In general, the observed picture looks quite "in the spirit" of the previous results. Well-parallelized rendering algorithms benefit from moving to a new core. Here I would like to draw attention to one curious fact that is not reflected in the graphs. The fact is that although it seems somewhat fantastic, the performance of the dual-core Core 2 Duo E8500 processor in the final rendering has almost grown to the performance level of the youngest of AMD's quad-core processors, the Phenom 9500. According to our tests, this AMD processor in 3ds max 9 scores 5, 61 points, and in Cinebench R10 - 7114 points.
Other applications
For this section, we have chosen four more interesting common problems that do not fit thematically into any of the previous parts of the presentation. However, there is nothing fundamentally new on the diagrams either: Core 2 Duo E8500, E8400 and E8200 clearly outperform models with 65nm cores at the same frequency, and even more so, at the same cost.
Power consumption and heat dissipation
Since the new 45nm manufacturing process should be reflected in the electrical and thermal characteristics of the new CPUs, we decided to pay attention to practical tests of these indicators as well.First of all, we resorted to measuring the operating temperature of processors at idle and under load. During testing, the processors were cooled by the same Zalman CNPS9700 LED cooler. The energy-saving technologies Enhanced Intel SpeedStep and Cool "n" Quiet 2.0 have been enabled. By the way, Wolfdale processors, just like their predecessors, reset their multiplier to 6x in low load states.
Processor loading was performed using the Prime95 25.5 utility, temperature readings were taken using the CoreTemp 0.96 utility. The results obtained are shown in the table.
As expected, 45nm processors are generally cooler than their Core microarchitecture predecessors, but the temperature difference at full load is only 4-5 degrees. The fact is that the core of Wolfdale processors has a smaller area and, accordingly, a much higher density of transistors on a semiconductor chip, which makes it somewhat difficult to remove heat from it. That is why, at rest, Wolfdale and Conroe show approximately the same temperatures. As for the relatively low temperature of the Athlon 64 X2 6000+ processor, whose TDP is, by the way, twice as high as that of the Core 2 Duo, it is due to the not very good location of the temperature sensor on the core, which is far from the hottest parts of the semiconductor crystal this CPU.
From what has been said, it is quite clear that measuring the temperature of processors provides too much subjective information. Therefore, we also paid attention to power consumption tests, which should show the benefits of the new 45nm core in full. In our experiments, we measured the current passing through the processor power circuit, which allows us to estimate the power consumption of the CPUs themselves (without taking into account losses in the processor power converter).
The results shown by the new processors, released on the 45-nm process technology, are more than impressive. However, nothing else was expected, because the new technological process allowed not only to reduce the size of the elements, but also to significantly reduce leakage currents - for this, Intel switched to using transistors with a metal gate and a high-k dielectric in it. As a result, the power consumed by Wolfdale processors under load is comparable to the power consumption of a CPU two or three years ago at rest. Actually, it is this striking contrast between processor generations that is emphasized by the results of Athlon 64 X2, a processor whose microarchitecture has not yet been optimized for high "performance per watt" indicators.
conclusions
Actually, everything is clear and so. Summarizing the above, we can say that the new dual-core processors Core 2 Duo E8500, E8400 and E8200, based on 45-nm cores, are good in everything. They are not only faster than their predecessors at the same clock speeds - the maximum frequencies they have reached are also higher than previous Intel processors. If we add to this the fact that Intel is going to sell new products at the same prices as the Core 2 Duo E6850, E6750 and E6550, then we can talk about a "free" increase in the performance of dual-core Intel processors by 10...15%.
In addition, the transfer of Core 2 Duo processors to production using a new technological process gives users additional bonuses. Firstly, they can include support for the promising SSE4.1 instruction set, which will prove itself in the future as the software is optimized. Secondly, Wolfdale processors are extremely economical. Thirdly, new processors promise excellent opportunities overclocking, for which they will surely find recognition among overclockers.
In other words, the second version of dual-core processors based on the Core microarchitecture is extremely successful. The only frustrating thing is that the appearance of these CPUs on store shelves will once again hit the positions of AMD, which is this moment cannot offer similar performance options. All dual-core processors from this manufacturer are unequivocally slower than the new Core 2 Duo E8000 series, which automatically "forces" them out of the price range "more than $150", where from now on Intel's dual-core offers will dominate on an uncontested basis.
Specify the availability and cost of Intel Core 2 Duo E8000 processors
Other materials on this topic
Phenom: gift for New Year by AMD
The second iteration of the Core microarchitecture: review of the Core 2 Extreme QX9650
AMD K10 microarchitecture
Testing Intel Core 2 Duo Processors Based on 45nm Wolfdale
Despite the release last year of a new Intel platforms Nehalem, Core 2 processors for Socket LGA775 are still "alive" and are not going to give up, as evidenced by the release of new models. The high price of motherboards based on the X58 Express chipset and DDR3 memory is holding back the expansion of the Core i7. Therefore, for at least a year we will have to use "outdated" processors with an LGA775 socket, two of which, related to the middle level, we will consider in this material.
As we already wrote in the introduction to the article about, the transition from the NetBurst architecture to the Core was due to a dead end in the development of the Pentium 4 family processors: the increase in frequency and power consumption did not correspond to the final level of performance. The change in architecture allowed Intel to get rid of its "Achilles heel" and push AMD with its solutions from the Athlon 64 and Sempron families, which began to gain a decent market share. In addition, with the release of the Core 2 processors based on the new architecture, Intel decided to follow the tick-tock strategy, which involves switching to a new microarchitecture every two years, and improving the previous one every intervening year. Suppose, if another architecture was presented last year, then this year there will only be a transition to a new technical process with some improvements, and next year something completely new will be waiting for us, but on an already run-in technical process, and so on.
If our readers remember, Core 2 processors (dual-core Conroe and quad-core Kentsfield), made according to the technological standards of 65 nm, were introduced in 2006, and a year later 45-nm solutions from the Penryn family (Wolfdale and Yorkfield) came out. Among the main innovations, we note an increase in the amount of shared cache memory of the second level from 4 to 6 MB (the quad-core Wolfdale has a core twice as large, since the processor consists of a pair of crystals) and the emergence of a new set of SSE4.1 instructions, designed to speed up the processing of streaming data, video encoding and mathematical calculations. There have also been some changes made to improve the performance of 45nm cores, but at home they are unlikely to be able to perform as they should.
Wolfdale core
With the transition to a new process technology, the supply voltage was reduced, thereby reducing the level of energy consumption. But, despite this, the level of TDP remained at the same place, which will allow Intel to increase the frequency of processors if something happens without increasing the requirements for motherboards available on the market. Moreover, the potential of 45-nm cores makes it possible to reach frequencies of the order of 3.5-3.6 GHz without any problems, and Intel will have a chance to hold the high-frequency CPU market until the Nehalem platform is widely used.
Now for the Wolfdale derivatives. As before, to cover all market segments based on the main core, either high-level solutions (quad-core Yorkfield in our case) or affordable processors with a reduced L2 cache are produced. The core with untouched L2 cache goes to the production of CPUs of the E8xxx series (FSB 1333 MHz), with 3 MB of cache is the basis of the E7xxx model range (FSB 1066 MHz), and with 2 MB - for the most affordable Pentium Dual-Core E5xxx (FSB 800 MHz ). The latest line in terms of its characteristics corresponds to the Core 2 Duo E4xxx series. With the transition to the 45-nm process technology, fractional multipliers that have not been used for a long time were introduced, which made it possible to reduce the frequency step between models. The quad-core Core 2 Quad processors are also available with different total L2 cache sizes: 12 MB models Q9x50 and QX9x70 ( Extreme versions Edition with FSB 1600 MHz), with 6 MB Q9x00 and 4 MB Q8xxx.
The table below includes processors from the Penryn and Conroe families.
Core 2 Quad QX9xxx / Q9xxx / Q9xxxS* |
Core 2 Quad Q8xxx / Q8xxxS* |
Core 2 Duo E8xxx |
Core 2 Duo E7xxx |
Pentium Dual Core E5xxx |
Core 2 Quad QX6xxx / Q6xxx |
Core 2 Duo X6xxx / E6xxx |
Core 2 Duo E4xxx |
Pentium Dual Core E2xxx |
|
Allendale (Conroe-1M) |
|||||||||
Production technology, nm |
|||||||||
Number of Cores |
|||||||||
Rated frequencies, GHz |
3,2-3,0/3,0-2,66/2,83-2,66 |
2,93-2,66/2,4-2,66 |
|||||||
Bus frequencies, MHz |
1600-1333 / 1333 |
1333-1066 / 1066 |
1066 / 1333-1066 |
||||||
Second level cache, MB |
|||||||||
TDP level, W |
|||||||||
connector type |
* - energy efficient solutions
** - younger Core 2 Duo E6xxx models were equipped with 2 MB L2 cache
As you can see, the younger quad-core Yorkfields are obtained by "gluing" either a pair of E5xxx or a pair of E7xxx. We can safely assume that if applications are poorly optimized for multithreading, a small L2 cache will greatly affect the performance of a processor that is much more expensive than its dual-core counterparts. Over time, the attractiveness of multi-core solutions will only grow, and the "bottleneck" in the form of a small amount of cache memory in the second level will come to naught. We will return to this in our future materials, but for now let's see how we can pass the time before the release of mass products based on the Nehalem architecture.
Intel Core 2 Duo E8200 and E7200 processors
Despite the significant offer of high-frequency models of dual-core Core 2, we settled on processors and, as the most affordable ones. Considering that the Core 2 Duo family has always had a good overclocking potential, the question naturally arises - why pay more?
The older model came to us in a boxed version, the younger one - in an OEM version, intended exclusively for assemblers of finished systems. The only disadvantage of such processors is the lack of a complete cooler and a 12-month warranty, which is supported only by the seller. Previously, there was an opinion that boxed versions of processors run better; perhaps, but if so, it was a long time ago.
The Core 2 Duo E8200 processor, like all 45 nm models, comes in a small blue box with an updated design, on the front side of which there is a sticker indicating the use of the 45 nm process technology in the production of the CPU and 6 MB of shared cache memory of the second level. For other series of processors, of course, a different amount of L2 cache will be indicated.
Core 2 Duo E8200 packaging
The supplied cooler with Core 2 Duo is shaped like a Celeron 4xx cooling system based on Conroe-L, but unlike the latter, it is significantly simplified (and this is for a processor with a frequency of 2.66 GHz and 6 MB of L2 cache!). The low-profile heatsink is now devoid of a core, and the mounting system is a single unit with the frame on which the fan is installed, i.e. completely plastic.
Cooling system Core 2 Duo E8200 vs. Core 2 Duo E4600
Such a design does not inspire much confidence, although, if you look at it this way, they also said at one time about the LGA775 connector: "no more than 20 times to install and remove the processor." But, as practice has shown, there is no danger in the frequent change of the processor; same with this cooler. And it is unlikely that someone will often put it on and take it off, especially ordinary users. Enthusiasts, of course, will first of all turn their attention to a more efficient cooling system, and the complete cooler will be left to gather dust in the box.
Externally, the Core 2 Duo E8200 and E7200, like all representatives of the LGA775 platform, differ in the presence and arrangement of elements on the "belly" of the processors. The marking indicates the model, main characteristics (CPU frequency, L2 cache size, FSB frequency) and batch number. Also, the sSpec Number is applied to the cover, by which you can determine the core stepping (on the manufacturer's website or with specialized utilities, for example, CPU-Z). The stepping itself will be useful when looking for a processor with good overclocking potential.
Core 2 Duo E8200 and E7200 processors
The Core 2 Duo E8200 processor, made in Malaysia and received by us for testing, had the very first C0 stepping, the overclocking potential of which is at the level of 3800-4000 MHz. The E8400, E8500 and E8600 models have already managed to switch to the E0 stepping, which is characterized by lower heat dissipation and higher potential - about 4200 MHz at a voltage of 1.4 V and air cooled. Naturally, overclocking is a lottery and any processor can achieve either higher or lower frequencies. It all depends on the specific instance.
Specifications Core 2 Duo E8200
With a light load, the frequency of the Core 2 Duo E8200 drops from the nominal 2.66 GHz to 2.0 GHz due to energy-saving technologies, while the multiplier decreases to 6x.
The Core 2 Duo E7200 is also made in Malaysia and has an early M0 stepping. The overclocking potential of the E7xxx is at the level of 3800 MHz, but with the transition to R0 (E7400 and E7500), the bar can move up.
Specifications Core 2 Duo E7200
The junior representative of the new family of processors in terms of energy-saving technologies is no exception, and its frequency drops from 2.53 GHz to 1.6 GHz when the system is idle.
Overclocking
To overclock the Core 2 Duo E8200 and E7200 processors, the following configuration was assembled:
Cooler: Noctua NH-U12P
Memory: Mushkin HP2-8500 996612 (2x1024 MB, DDR2-1066)
Video card: ASUS EN8800GS TOP 384MB
The memory operation mode was set to the minimum possible with timings of 5-5-5-18, the voltage to the processors was supplied at the level of 1.4 V; it is safe for 45nm solutions in continuous use. In general, according to Intel technical documentation maximum voltage can be 1.45 V, but no more, because then degradation is possible (the consequences are not widespread, but isolated cases have taken place). Used as a stress test OCCT program 2.01 for an hour.
With these settings, we managed to overclock the Core 2 Duo E8200 processor to 4200 MHz, while the FSB frequency was 525 MHz, and the memory was 1050 MHz.
Overclocking level Core 2 Duo E8200
Overclocking is excellent, but requires a high-quality motherboard, since not everyone will be able to function stably at frequencies above 475-500 MHz. In this case, we were lucky and, perhaps, by raising the voltage on the processor, we could achieve even better results. But we still decided to refuse this, since we need the result for everyday use, and not for a minute benching session.
The next processor, thanks to its high multiplier, is suitable for any motherboard, since for overclocking it is no longer necessary to raise the FSB frequency too much. Despite all our hopes, the ceiling of the Core 2 Duo E7200 turned out to be 3752 MHz, while the FSB worked at 395 MHz. The memory frequency was slightly less than 800 MHz, which could hardly have affected the result. To make the test more informative, we decided to take a couple more of the same processors, but their potential turned out to be even worse than the first copy - the FSB Wall affects around 390 MHz, when the processor loses stability after this mark.
Overclocking level Core 2 Duo E7200
Well, we can state the fact that indeed the processors of the younger family do not have a good overclocking potential, unlike the older models. We can also assume that we came across specimens that were not very successful for overclocking, since there are a lot of results on the Web with an FSB bus over 400 MHz (especially since even Pentium Dual-Core E5xxx conquer frequencies of the order of 4.0-4.2 GHz).
test stand
To test the Core 2 Duo E8200 and E7200 processors, the following configuration was assembled:
Motherboard: ASUS P5K Deluxe/WiFi-AP (Intel P35)
Cooler: Noctua NH-U12P
Memory: Mushkin HP2-8500 996612 (2x1024MB, DDR2-1066, 5-5-5-15-2T)
Video card: Point of View GF9800GTX 512MB GDDR3 EXO
HDD: Samsung SP2504C (250 GB, SATA2)
Power supply: FSP FX700-GLN Epsilon (700 W)
This is the same configuration as when overclocked, only the video card was replaced, and the memory was running at a frequency of 1066 MHz (or close) with timings of 5-5-5-15.
The list of processors was as follows:
Core 2 Duo E8200 (2.66GHz, 8x333, 6MB; overclocked 4.2GHz, 8x525, 1052MHz memory)
Core 2 Duo E7200 (2.4GHz, 9.5x266, 3MB; overclocked 3.75GHz, 9.5x395, 1053MHz memory)
Core 2 Duo E7100 (2.4GHz, 9x266, 3MB, hypothetical)
Core 2 Duo E6750 (2.66GHz, 8x333, 4MB; overclocked 3.4GHz, 8x425, 1066MHz memory)
Core 2 Duo E4600 (2.4GHz, 12x200, 2MB; overclocked 3.34GHz, 10x334, 1066MHz memory)
The Core 2 Duo E6750 processor was derived from the Core 2 Extreme X6800 by reducing the coefficient to 8x and increasing the FSB frequency to 333 MHz. This mode will allow you to compare a processor based on the Wolfdale core with Conroe running at the same frequency. The hypothetical Core 2 Duo E7100 was needed for comparison with the E4600. True, in the latter case, the FSB of the 45-nm solution is higher, and therefore additional tests were carried out in which all processors worked at the same core and bus frequencies, equal to 3.2 GHz and 400 MHz, respectively, while the memory operated at a frequency of 800 MHz with timings 4-4-4-12. In order to make it possible to select a memory operation mode close to 1066 MHz when overclocking the Core 2 Duo E4600, the multiplier was reduced to 10x, and the FSB frequency was increased to 334 MHz.
Testing was carried out in Windows environment XP Pro SP2, which was tuned for maximum performance. The swap file was 2048 MB.
Test results
Memory Subsystem Tests
In most memory subsystem tests Everest programs an important role is played by the frequency of the system bus, and after that the architectural features and frequency of the processor. For example, the youngest representative of the Core 2 family had to be overclocked to the maximum possible so that it could demonstrate a result similar to high-level solutions operating in nominal mode. We also note the parity between the Core 2 Duo E8200 and E6750 at the standard frequency.
The memory latency test is sensitive to the architectural features of processors, and only a high frequency during overclocking allows solutions based on the Penryn architecture to show a low result (the lower the better).
The situation is similar in the L2 cache latency test, but in contrast to common test here the feature of the L2 cache of new processors strongly influences. The fact is that with the growth of the cache memory, the number of associativity channels was increased - from 16 (Conroe) to 24 (Penryn); this should have had a positive effect on performance, but the latency has become higher. This problem will be solved by increasing the frequency of future processors (possibly, due to the slow L2 cache, the frequency ceiling has increased, and not only due to the transition to the 45-nm process technology). In addition, the Split-load cache enhancement technology has been added, which allows you to distribute data in the cache more intelligently and compensate for the slowness of the L2 cache.
Gaming Applications
The performance of processors in gaming applications depends on both the size of the L2 cache and the frequency. So, in the game F.E.A.R. Extraction Point, which doesn't shine with complex graphics at medium settings, depends on the cache, and the overclocked Core 2 Duo E4600 turned out to be less efficient than the E8200 running at the nominal frequency. In Crysis, the opposite is true - due to more high frequency you can get more frames than with a capacious L2 cache. Naturally, as the resolution and graphics quality increase, the difference between the processors will level out.
Testing in 8x400 mode
Now let's move on to the results of "pure" performance, when all processors work with the same frequency of cores, bus and memory.
As you might expect, the amount of cache memory significantly affects performance (between 2 and 6 MB about 20%), and in combination with different system bus frequencies, Intel can release processor models designed for different market segments. But, given the overclocking potential of the Core 2, it is not necessary to pursue a solution with a large L2 cache.
conclusions
With the release of Core i7, Intel began migrating to a new architecture, but, as it already did with the AMD64 platform, LGA775 processors will be present on the market for at least a year and a half, which are so far the only available solutions from this manufacturer. Naturally, with the release of mass processors of the Nehalem architecture, the Core 2 family will lose its positions and move to the entry level, and some models will disappear altogether.
In the meantime, we will have to be content with LGA775 processors, made with 65 and 45 nm detail, and the latter look much more attractive at the same cost. Although the Core 2 Duo E8200 and E7200 reviewed by us are not the last in their family, they perfectly show the performance of solutions based on the Penryn architecture. Low heat, power consumption and high overclocking potential in the region of 3.8-4.2 GHz will allow these CPUs to become a worthy choice for enthusiasts and overclockers. The younger model allows overclocking to achieve the performance of more expensive solutions, but it does not require an overclocker motherboard, as it has a high multiplier, due to which sky-high FSB frequencies can be avoided. The E8xxx series is becoming the undisputed leader among dual-chip products, but overclocking the E8200 requires a high-quality board capable of keeping FSB frequencies around 500 MHz and higher.
Ordinary users will also not be left out. So, on the basis of the Core 2 Duo E7xxx, you can assemble a quiet media center or a mid-range gaming system, which in terms of performance will not be much inferior to systems with more expensive hi-end processors. The cost of these models is at the level of Core 2 Duo E4xxx, which is clearly not in favor of the latter. The same can be said about E8xxx and E6xxx. The only drawback of the Penryn family of processors is the need to use motherboards on Intel chipsets 3 Series and 4 Series. But when upgrading or buying new system you can't call it a hindrance anymore.
The situation with Core 2 Duo processors is no different from everything that happened before. In reviews and test stands, first of all, older and extreme models are found, and the younger ones either fade into the background, or are completely lost in the total mass of all new products entering the market and, thus, remain undeservedly forgotten. But after all, it is they who determine who is actually the winner and who, in fact, will receive most of the money, no matter how trivial it may sound. In our materials, we reflected in sufficient detail the architectural Intel features Core and conducted a significant number of tests, from the performance of new processors with math packages to the performance of the Core 2 Duo in everyone's favorite gaming applications. But it just so happened that the most affordable (consider the most popular) Core 2 Duo processors based on the Allendale core fell out of these same tests. As you may know, the main and only difference (besides, of course, clock frequency) of the junior series from processors based on the Conroe core lies in a twofold reduction in the amount of L2 cache memory (up to 2 MB). However, looking ahead, I would like to note that this did not greatly affect performance, and today we will prove it to you. Note that we do not take into account another representative of the series - Core 2 Duo E4300, which, among other things, has had its system bus frequency reduced to 800 MHz. It will be possible to talk about the performance of this processor only after its official announcement, which, according to the latest data, will take place in January 2007.
Intel Core 2 Duo E6300 and E6400 - "sweet couple"
Currently, prices for Intel Core 2 Duo E6300 and E6400 processors fluctuate around $200 and $230, respectively. And this means that the direct competitors for them are the dual-core AMD Athlon X2 3800+ and 4000+ and, with a stretch, their brother - the Intel Pentium D 945 with a clock frequency of 3.4 GHz. The distribution of prices among distributors is quite high, so at least conditionally rank the above models in the same price category. Quite a lot of words have already been said about the success of the new Intel architecture. If you still don’t understand something, then feel free to go to the study of the articles "Core 2 Duo processors: shock and awe" and "Intel Conroe dual-core processors". Today we have another task - to study the behavior of processors in practice, to identify their overclocking potential in difficult overclocking conditions (more on that later), and also to compare the manufacturer with the Core 2 Duo on the Conroe core and its closest rivals. That's what we're going to do today, but first let's take a look at pivot table with characteristics.|
CPU |
Clock frequency, GHz |
Factor |
Bus frequency, MHz |
L2 cache size, Mb |
Typical heat dissipation, W |
Cost, $ |
|
|
Core 2 Extreme X6800 |
|||||||
|
Core 2 Duo E6700 |
|||||||
|
Core 2 Duo E6600 |
|||||||
|
Core 2 Duo E6400 |
|||||||
|
Core 2 Duo E6300 |
|||||||
|
Core 2 Duo E4300 |
This table confirms our words - the differences between the processors are minimal, and those who are not used to overpaying can easily compensate for the difference in clock speed by overclocking. Now let's take a closer look at processors. Boxed versions of processors look no different from each other. Here are two processors based on the Allendale core and one on the Conroe (Core 2 Duo E6600).
The form factor of the box itself has not changed compared to the Pentium D either, only the color scheme and arrangement of some elements have changed.
The package is pretty standard too. This is the "BOX-cooler" from Intel, which has already become popular, with the factory "gum" applied in advance and the instruction manual, which describes the basic rules for working with the processor: "Do not chew, do not throw, in microwave oven not warm up, well, and so on ... "
The marketing name of the Intel Core 2 Duo E6300 processor can be deciphered as follows.
The letter index at the beginning of the marking, in this case "E", classifies the TDP of the processor, without any relation to the form factor. Thus, the TPD level, in accordance with the official Intel classification, varies between 50-70 watts. In turn, a 4-digit digital index also carries a semantic load. In general, the larger the 4-digit number represented by the processor label, the greater the performance and power consumption it is characterized by. At the same time, the first digit means the chip belongs to a certain product family, the second digit indicates the corresponding arrangement of chips within the family. Accordingly, the larger the number, the more efficient the chip. Of course, this method of labeling chips has nothing to do with the PR ratings of AMD processors, which claim some kind of conditional correspondence to some conditional megahertz of some conditional processor, everything is much simpler: than more number, the more efficient the chip. Here's what the CPU-Z utility says about the Core 2 Duo E6300 processor.
Despite the reduced amount of L2 cache, the E6300 model, just like the E6400, is endowed with the whole package of technologies inherent in the Core 2 Duo processor line, namely:
- Intel Wide Dynamic Execution is a technology for executing more instructions per clock cycle, which increases the efficiency of application execution and reduces power consumption. Each processor core can execute up to four instructions simultaneously using a 14-stage pipeline;
- Intel Intelligent Power Capability - a technology that activates the work of individual chip nodes as needed to execute tasks, which significantly reduces the power consumption of the system as a whole;
- Intel Advanced Smart Cache - a technology for using L2 cache memory common to all cores, which reduces overall power consumption and improves performance, while, as needed, one of the processor cores can use the entire cache memory while dynamically disabling another core;
- Intel Smart Memory Access is a memory subsystem optimization technology that reduces response time and improves throughput memory subsystems;
- Intel Advanced Digital Media Boost is a technology for processing 128-bit SSE, SSE2 and SSE3 instructions widely used in multimedia and graphic applications, for one cycle.
- The SSE4 instruction set, which adds another 52 new processor instructions to speed up the processing of multimedia, text and other data (the CPU-Z application already identifies them).
Let's smoothly move on to the consideration of the second processor, whose name is Core 2 Duo E6400. All the differences of the Core 2 Duo processor following the E6300 are reduced to a clock frequency increased by 266 MHz, achieved by changing the multiplier from 7 to 8.
The CPU-Z utility confirms this.
But contrary to expectations, the processor showed the best result in overclocking. The resulting frequency was 3200 MHz.
Definitely, many readers might think that the reason for such low overclocking rates, in addition to factory cooling, could be the weak overclocking capabilities of the motherboard. However, we dare to assure you that after long tests and strength testing of the ASUS P5B Deluxe board, the ceiling of its "overclocking" was determined at around 475 MHz FSB, which, unfortunately, was not achieved with any of the processors tested today. If, nevertheless, we discuss the reasons for some “underclocking” more carefully, then first of all these are deliberately complicated testing conditions - the experimental Core 2 Duo E6300 was able to start at frequencies above 3.2 GHz, but after 5-10 minutes of testing it warmed up above 65 ° C and began to “skip cycles”. True, it is worth mentioning that not only poor ventilation and high temperature can stand in the way of a further increase in clock frequencies - the Core 2 Duo E6600 under study could not be heated above 58 ° C by any means, although further overclocking led to " blue screen death" during Windows boot, i.e. the frequency of 3.33 GHz turned out to be the ceiling of stability for him. Now that everything is more or less clear with all the processors, let's see how the twofold reduction in the amount of cache memory affected in practice, and also compare the performance of the processors with the previously described competitors and the now budget Athlon 64.
Nearly 3 times the speed: 802.11ax 2x2 160 MHz delivers a maximum theoretical data rate of up to 2402 Mbps, nearly 3 times (2.8 times) faster than 802.11ac 2x2 80 MHz (867 Mbps) ) as documented in the specs wireless standard IEEE 802.11. Requires an 802.11ax wireless router with a similar configuration.
Compared to other PC I/O technologies including eSATA, USB, and IEEE 1394 Firewire*. Actual performance figures may vary depending on the hardware and software used. The use of a device with Thunderbolt™ technology is mandatory. Additional information can be found on the website.
The software and workloads used in benchmark tests are optimized for high performance with Intel® microprocessors only. Benchmark tests, including SYSmark* and MobileMark*, are conducted using specific computer systems, components, software, operations and features. Any changes to these parameters may change the final results. Customers are also advised to refer to other sources of information and performance tests when deciding whether to purchase certain systems and components, including tests to verify the performance of specific products when used in combination with other components.
More information is available on the website.
Based on 3DMark FireStrike* workload comparison test performed on pre-production 10th Gen Intel® Core™ i7-1065G7 processor and 8th Gen Intel® Core™ i7-8565U processor. Benchmark results are based on testing as of May 23, 2019 and may not reflect all publicly available security updates. detailed information presented in the description of the configuration. No system can be completely secure.
Best-in-class Wi-Fi 6 technology: Intel® Wi-Fi 6 (Gig+) adapters support additional 160 MHz channels to achieve the highest possible theoretical speed (2402 Mbps) for typical Wi-Fi 2x2 802.11ax PC adapters. Premium Intel® Wi-Fi 6 (Gig+) adapters deliver 2x to 4x the theoretical maximum speed over standard Wi-Fi adapters 802.11ax PC 2x2 (1201 Mbps) or 1x1 (600 Mbps) which only support mandatory 80 MHz channels.
Based on AIXprt workload comparison test results between pre-production 10th Gen Intel® Core™ i7-1065G7 processor and 8th Gen Intel® Core™ i7-8565U processor (INT8 results). Benchmark results are based on testing as of May 23, 2019 and may not reflect all publicly available security updates. Detailed information is provided in the description of the configuration. No system can be completely secure.
Intel is a sponsor and member of the Benchmark XPRT developer community, and the primary developer of XPRT benchmarks. Principled Technologies is the publisher of the XPRT family of benchmark tests. You must refer to other sources of information and performance tests to get a full evaluation of the product you plan to buy.
Changing the clock speed or voltage may damage or shorten the life of the processor and other system components, and may also degrade system stability and performance. If the processor specifications change, the product may not be covered under warranty. Behind additional information contact the system and component manufacturers.
Intel and the Intel logo are trademarks of Intel Corporation or its subsidiaries in the United States and/or other countries.
* Other names and trademarks are the property of their respective owners. (if third party names and trademarks are used)
AND Core 2 Extreme- dual-core desktop processors based on Intel's Conroe core.
Based on next-generation Core microarchitecture Intel processor The Core 2 Duo is the second generation of chips built using a 65nm process.
This process makes it possible to create transistors so small that about a hundred of them would fit in one human cell.
Using two powerful processors sharing resources, and with such an incredibly small size, the Intel Core 2 Duo chip allows you to achieve significantly more performance while consuming less power.
The 64-bit processor architecture allows the Intel Core 2 Duo to manipulate data and execute instructions in twice the chunks (compared to 32-bit processors), which greatly increases processing power.
The key characteristics of these processors, inherited from the predecessors of the Intel Pentium M, enriched with the best practices of the NetBurst architecture and a number of completely new technologies:
Intel Wide Dynamic Execution- technology to execute more instructions per cycle, increasing the efficiency of application execution and reducing power consumption.
Each processor core can execute up to four instructions simultaneously using a 14-stage pipeline.
Intel Intelligent Power Capability- a technology with the help of which, for the execution of tasks, the operation of individual chip nodes is activated as needed, which significantly reduces the power consumption of the system as a whole.
Intel Advanced Smart Cache- a technology for using the L2 cache memory common to all cores, which reduces overall power consumption and improves performance, while, as necessary, one of the processor cores can use the entire cache memory while dynamically turning off the other core.
Intel Smart Memory Access- technology for optimizing the operation of the memory subsystem, which reduces the response time and increases the throughput of the memory subsystem.
Intel Advanced Digital Media Boost- a technology for processing 128-bit SSE, SSE2 and SSE3 commands, widely used in multimedia and graphic applications, in one clock cycle.
The processor marking consists of five characters.
The letter index at the beginning of the marking classifies the TDP of the processor, without any relation to the form factor:
X - TDP over 75W
E - TDP from 50 W and above
T - TDP within 25W - 49W
L - TDP within 15W - 24W
U - TDP of the order of 14 W or less
The 4-digit digital index also carries a semantic meaning: the larger the 4-digit number represented by the processor marking, the greater the performance and power consumption it is characterized by.
The first digit means that the chip belongs to a certain product family.
The second digit is the corresponding layout of chips within the family.
Accordingly, the larger the number, the more efficient the chip.
Here's what the markings of modern processors look like:
Core 2 Extreme X6800 - 2.93 GHz, 4 MB L2 cache, 1066 MHz FSB
Core 2 Duo E6600 - 2.4 GHz, 4 MB L2 cache, 1066 MHz FSB
Core 2 Duo E6400 - 2.13 GHz, 2 MB L2 cache, 1066 MHz FSB
Core Duo T2500 - 2 GHz, 2 MB L2 cache, 667 MHz FSB
Core Duo U2500 - 1.06 GHz, 2 MB L2 cache, 533 MHz
Of course, this method of labeling chips has nothing to do with the PR ratings of AMD processors, which claim some kind of conditional correspondence to some conditional megahertz of some conditional processor.
Everything is much simpler: the larger the number, the more efficient the chip.
Driver AMD Radeon Software Adrenalin Edition 19.9.2 optional
A new version AMD drivers Radeon Software Adrenalin Edition 19.9.2 Optional improves performance in Borderlands 3 and adds support for Radeon Image Sharpening.
Cumulative windows update 10 1903 KB4515384 (added)
On September 10, 2019, Microsoft released a cumulative update for Windows 10 version 1903 - KB4515384 with a number of security improvements and a fix for a bug that broke Windows work Search and caused high CPU usage.
Driver Game Ready GeForce 436.30 WHQL
NVIDIA has released the Game Ready GeForce 436.30 WHQL driver package, which is designed for optimization in games: "Gears 5", "Borderlands 3" and "Call of Duty: Modern Warfare", "FIFA 20", "The Surge 2" and "Code Vein", fixes a number of bugs seen in previous releases, and expands the list of displays in the G-Sync Compatible category.
Loading...