by Ilya Gavrichenkov
09/19/2012 | 06:31 AM
New Intel processors from the Ivy Bridge family have been in the market for a few months already, but it seems that they are not extremely popular. We have already pointed out multiple times that they do not look like a significant step forward compared with their predecessors: their computational performance has increased insignificantly, and the frequency potential revealed through overclocking has become even lower than that of the previous Sandy Bridge generation. Intel is also aware that there is no real buying craze about the Ivy Bridge processors: the life span of the previous generation processors manufactured with older 32 nm process has been extended on and on, and the forecasts on the growing popularity of the newcomers are not particularly optimistic at this point. To be more exact, Intel intends to increase the share of Ivy Bridge processors only to 30% of their desktop offerings, while the remaining 60% of all shipped processors will remain Sandy Bridge based products. Does this situation actually allow us to consider new Intel processors yet another success?
Not at all. The thing is that everything I said above refers only to the desktop processors. The mobile market segment responded totally differently to the Ivy Bridge launch, because most of the innovations were designed with notebooks in mind. The two major advantages Ivy Bridge has over Sandy Bridge are, namely, significantly lower heat dissipation and power consumption and accelerated graphics core with DirectX 11 support, are in very high demand for the mobile systems. These features not only stimulated the development of notebook models with much better combination of consumer qualities, but also encourages rapid expansion of a new class of ultra-portable devices aka ultrabooks. And on top of that, the new 22 nm process and Tri-Gate transistors helped lower the size and production cost of the semiconductor dies, which is yet another argument in favor of the new design’s success.
As a result, only desktop users can actually remain somewhat uninspired by the new Ivy Bridge, and not because of some serious issues, but mostly for the lack of radical improvements, which, however, have never been promised in the first place. Do not forget, that Intel’s classification places Ivy Bridge processors into the “tick” stage, which is simply the transition of old microarchitecture onto new semiconductor base. Either way, Intel knows that desktop users are less excited about the new generation processors than notebook users, therefore, they are not in a hurry to completely refresh the lineup just yet. Until recently new microarchitecture in the desktop segment has only found its way into top quad-core Core i7 and Core i5 CPUs. Core i3 and Pentium processors on Ivy Bridge microarchitecture have just appeared in Intel’s product lineup. Moreover, Ivy Bridge based processors co-exist side by side with their Sandy Bridge based fellows and do not push them off the stage in any way. The new microarchitecture should become more aggressive in late fall, and until then it is up to the users to decide which processors they would rather go with: second generation (2000-series), or third generation (3000-series).
This is actually the reason why we decided to undertake this test session. Today we are going to compare Core i5 processors from the same price range designed for the same LGA 1155 platform but using different microarchitecture: Ivy Bridge and Sandy Bridge.
Back one and a half years ago, when Intel launched second generation Core processors, they introduced clear classification of processor families, which they have maintained since then. According to this classification, the fundamental features of the Core i5 are quad-core design without “virtual multi-threading” technology – Hyper-Threading and 6 MB L3 cache. These peculiarities were typical of the previous generation Sandy Bridge processors and were carried on in the new Ivy Bridge CPUs.
It means that all Core i5 processors with the new microarchitecture are very similar. To some extent this allows Intel to unify the production: all today’s Ivy Bridge Core i5 processors use identical 22 nm semiconductor die with E1 stepping that consists of 1.4 billion transistors and is about 160 mm2 big.
Despite the similarities all LGA 1155 Core i5 processors share in formal specifications, the differences between them are obvious. New 22 nm manufacturing process and Tri-gate transistors allowed Intel to lower the TDP of the new Core i5 processors. The LGA 1155 Core i5 CPUs used to have 95 W TDP, but the TDP of the new Ivy Bridge processors has been lowered to 77 W. However, despite the lowering of the TDP, the clock frequencies of the Ivy Bridge CPUs haven’t been increased. Just like their predecessors, the top previous generation Core i5 processors work at 3.4 GHz maximum clock frequencies. It means that the performance advantage of the new Core i5 processors over the old ones results only from the microarchitectural improvements of the computational resources, which has also been confirmed by Intel developers.
Speaking of the strengths of the new processor design, we first of all would like to stress the modifications applied to the graphics core. Third generation Core i5 processors use a new modification of the Intel’s graphics accelerator – HD Graphics 2500/4000. It supports DirectX 11, OpenGL 4.0 and OpenCL 1.1 software interfaces and in some cases may offer better 3D performance and faster HD video transcoding into H.264 format using Quick Sync technology.
Moreover, Ivy Bridge processor design also has improved “uncore” - memory and PCI Express bus controllers. As a result, systems with the new third generation Core i5 processors can fully support PCI Express 3.0 graphics cards and can clock DDR3 memory at higher frequencies than their predecessors.
The third generation desktop Core i5 processor family (that is Core i5-3000) has barely changed since its official public debut. Only a couple of intermediate models have been added to the lineup, which adds up to a total of five models (if we do not take into consideration energy-efficient models with restricted TDP). If we add here a few Core i7 CPUs on Ivy Bridge microarchitecture, we will get the entire lineup of 22 nm processors for LGA 1155 form-factor:
The table above, obviously, requires some comments regarding the functioning of the Turbo Boost technology that allows processors to increase their clock frequency as the power and thermal conditions permit. In Ivy Bridge this technology has been slightly modified, and the new Core i5 processors can overclock a little more aggressively than their predecessors from the Sandy Bridge family. With minimal improvements in the microarchitecture of the computational cores and no clock frequencies increase this is exactly the feature that can guarantee certain advantage over the predecessors.
The maximum frequency that Core i5 processors can reach with only one or two cores fully utilized is 400 MHz higher than the nominal. If the load is multi-threaded, then Ivy Bridge Core i5 processors may raise their clock speed by 200 MHz above the nominal provided the thermal conditions are favorable. In this case Turbo Boost efficiency will be the same for all above mentioned processors, and the only difference from the previous generation CPUs will be higher frequency increase when utilizing two, three or four cores: Core i5 from the Sandy Bridge generation head 100 MHz less headroom for frequency growth in this case.
Using the diagnostic CPU-Z utility let’s take a closer look at all the members of the Core i5 Ivy Bridge processor family.
Core i5-3570K processor is the crown of the entire third generation Core i5 line-up. It boasts not only the highest clock frequency in the family, but also has a unique distinguishing feature stressed by the “K” letter in the model name – unlocked clock frequency multiplier. This allows Intel to regard Core i5-3570K as a special overclocker product. And it does indeed look very appealing against the background of the top overclocker CPU for the LGA 1155 platform, Core i7-3770K, due to more affordable price. Therefore, Intel Core i5-3570K is currently almost the best option for enthusiasts out there from the market prospective.
At the same time Core i5-3570K is appealing not only due to its overclocking-friendliness. Non-overclockers may also turn to this product, because it features the top graphics core model from Intel – HD Graphics 4000 with much higher performance than the other Core i5 processors have to offer.
The same model name as that of the previous processor model but without the “K” letter in the end gives us a hint that this is a non-overclocker modification of the previous processor. And this is indeed so: Core i5-3570 works at the same frequencies as its more advanced brother, but doesn’t support unlimited adjustment of the clock frequency multiplier parameter, which the enthusiasts value so much.
However, there is another thing. Core i5-3570 doesn’t have the fastest graphics core, and we are left with Intel HD Graphics 2500, which is significantly weaker in many aspects as we will demonstrate later in this review.
As a result, feature-wise Core i5-3570 is actually closer to Core i5-3550 than to Core i5-3570K. And there are very good reasons for that. While this processor came out a little later than the first group of Ivy Bridge CPUs, it does signify certain evolution of the family. With the same MSRP as the Core i5-3550 model, it is destined to replace the latter eventually.
Another lowering of the model number once again indicates that the computing power will get lower, too. In this case Core i5-3550 is slower than Core i5-3570 because of slightly lower clock speed. However, the difference is only around 100 MHz or 3%, so it is not surprising that Intel priced Core i5-3570 and Core i5-3550 similarly. They anticipate the Core i5-3570 to slowly oust Core i5-3550 from the stores. That is why all other specifications of these two processors, except the clock frequency, are identical.
The MSRP of the two junior Core i5 processors on the 22 nm Ivy Bridge core is set below $200. They are currently retailing at around this price point. However, Core i5-3470 is not that much inferior to the top Core i5 processors: it has all four computational cores, 6 MB L3 cache and over 3 GHz clock frequency. Intel decided to use a 100 MHz increment to differentiate between the modifications in the refreshed Core i5 lineup, therefore, you can hardly expect different processor models to demonstrate dramatically different performance in real applications.
However, Core i5-3470 has different graphics performance than the top processors in the family. HD Graphics 2500 core in this processor works at a slightly lower clock speed: 1.1 GHz vs. 1.15 GHz by the more expensive CPU models.
The very junior 3rd generation Core i5 processor, Core i5-3450, slowly leaves the market, just like Core i5-3550. Core i5-3450 is being replaced by the above described Core i5-3470, which works at a slightly higher clock frequency. There are no other differences between these processors.
In order to paint a complete picture of the contemporary Core i5 performance, we subjected all five above described 3000-series Core i5 processors to our extensive testing routine. They will be competing against Sandy Bridge LGA 1155 processors from the same price range released earlier: Core i5-2400 and Core i5-2500K. They current price allows a fair comparison between these CPUs and the new 3000-series processors: Core i5-2400 is priced at the same level as Core i5-3470 and Core i5-3450, while Core i5-2500K is priced just below Core i5-3570K.
Besides these processors, we also included the performance numbers of higher-end Core i7-3770K and Core i7-2700K, as well as an AMD FX-8150 CPU. By the way, it is remarkable that after yet another price slashing this Bulldozer representative ended up at the level of the least expensive Core i5 3000-series CPUs. In other words, AMD is no longer exercising the idea of making their eight-core product a worthy competitor to Intel’s Core i7.
As a result, we put together test platforms with the following hardware and software components:
For our tests of the AMD FX-8150 based system we installed KB2645594 and KB2646060 OS patches.
We used Nvidia GeForce GTX 680 graphics card to test the CPUs performance in a system equipped with a discrete graphics accelerator. We used AMD Radeon HD 6570 was utilized as a reference point during the tests with integrated graphics involved.
Intel Core i5-3570 processor didn’t participate in the tests performed in a system with a discrete graphics card, because its computational performance is identical to that of the Intel Core i5-3570K working at the same clock frequencies.
As usual, we use Bapco SYSmark 2012 suite to estimate the processor performance in general-purpose tasks. It emulates the usage models in popular office and digital content creation and processing applications. The idea behind this test is fairly simple: it produces a single score characterizing the average computer performance.
Overall, the results of Core i5 processors from the 3000-series are quite what we expected them to be. They are faster than the previous generation Core i5. Moreover, the Core i5-2500K, which is the fastest Core i5 CPU on Sandy Bridge microarchitecture, yields even to the most junior Core i5-3450. However, the new Core i5 CPUs are still unable t catch up with the Core i7, because of the missing Hyper-Threading technology.
Let’s take a closer look at the performance scores SYSmark 2012 generates in different usage scenarios. Office Productivity scenario emulates typical office tasks, such as text editing, electronic tables processing, email and Internet surfing. This scenario uses the following applications: ABBYY FineReader Pro 10.0, Adobe Acrobat Pro 9, Adobe Flash Player 10.1, Microsoft Excel 2010, Microsoft Internet Explorer 9, Microsoft Outlook 2010, Microsoft PowerPoint 2010, Microsoft Word 2010 and WinZip Pro 14.5.
Media Creation scenario emulates the creation of a video clip using previously taken digital images and videos. Here they use popular Adobe suites: Photoshop CS5 Extended, Premiere Pro CS5 and After Effects CS5.
Web Development is a scenario emulating web-site designing. It uses the following applications: Adobe Photoshop CS5 Extended, Adobe Premiere Pro CS5, Adobe Dreamweaver CS5, Mozilla Firefox 3.6.8 and Microsoft Internet Explorer 9.
Data/Financial Analysis scenario is devoted to statistical analysis and prediction of market trends performed in Microsoft Excel 2010.
3D Modeling scenario is fully dedicated to 3D objects and rendering of static and dynamic scenes using Adobe Photoshop CS5 Extended, Autodesk 3ds Max 2011, Autodesk AutoCAD 2011 and Google SketchUp Pro 8.
The last scenario called System Management creates backups and installs software and updates. It involves several different versions of Mozilla Firefox Installer and WinZip Pro 14.5.
In most usage scenarios the situation is pretty typical: 3000-series Core i5 is faster than predecessors, but slower than Core i7 processors based on Sandy Bridge as well as Ivy Bridge microarchitecture. However, there are a few situations when things become out of the ordinary. For example, in Media Creation scenario Core i5-3570K processor manages to outperform Core i7-2700K; AMD FX-8150 does unexpectedly well in 3D modeling suites; the previous-generation Core i5-2500K processor almost catches up with the new Core i5-3470 in System Management scenario, which generates mostly single-threaded loads.
As you know, it is the graphics subsystem that determines the performance of the entire platform equipped with pretty high-speed processors in the majority of contemporary games. Therefore, we do our best to make sure that the graphics card is not loaded too heavily during the test session: we select the most CPU-dependent tests and all tests are performed without antialiasing and in far not the highest screen resolutions. In other words, obtained results allow us to analyze not that much the fps rate that can be achieved in systems equipped with contemporary graphics accelerators, but rather how well contemporary processors can cope with gaming workload. Therefore, the results help us determine how the tested CPUs will behave in the nearest future, when new faster graphics card models will be widely available.
During our numerous preceding test sessions we pointed out multiple times that Core i5 processors were a really good fit for gaming systems. And today we are going to stick by this statement again. Highly efficient microarchitecture, quad-core design and high clock frequencies make Core i5 processors really fast in games. And the lack of Hyper-Threading technology may actually become a benefit in games that aren’t that well-optimized for multi-threaded configurations. However, there remain fewer games like that every day, which is exactly what we see from the obtained results. Core i7 processor with Ivy Bridge inside is always faster than similarly designed Core i5 processors in all the diagrams. As a result, the gaming performance of the 3000-series Core i5 CPUs rest where we would expect it to: these processors are undoubtedly better in games than the 2000-series Core i5, and sometimes can even compete successfully against Core i7-2700K. At the same time I would like to point out that the top AMD processor doesn’t stand a chance against contemporary Intel products: its performance lag in games is nothing short of a true catastrophe.
In addition to our gaming tests we would also like to offer you the results of the Futuremark 3DMark11 benchmark (Performance profile):
The synthetic Futuremark 3DMark11 benchmark doesn’t reveal anything new. The performance of the third generation Core i5 processors falls right between that of the previous generation Core i5 and any Core i7 CPUs supporting Hyper-Threading and working at slightly higher clock speeds.
To test the processors performance during data archiving we resort to WinRAR archiving utility. Using maximum compression rate we archive a folder with multiple files with 1.1 GB total size.
The multi-threading support has been significantly improved in the latest WinRAR versions, so now the archiving speed became seriously dependent on the number of computational cores inside the CPU. Therefore, Core i7 processors fortified with Hyper-Threading technology, as well as the eight-core AMD FX-8150, do best of all in this test. As for the Core i5 series, things haven’t really changed here. Ivy Bridge based Core i5 is indisputably better than the old CPUs, showing about 7% advantage for the processors working at identical frequencies.
The processor performance in cryptographic tasks is measured using a built-in benchmark of the popular TrueCrypt utility that uses AES-Twofish-Serpent “triple” encryption. I have to say that this utility not only loads any number of cores with work in a very efficient manner, but also supports special AES instructions.
Everything is as usual, only the FX-8150 processor is again at the top of the chart. It definitely benefits from the ability to execute eight computational threads simultaneously and to process integer and bit-operations very fast. As for 3000-series Core i5, they are again indisputably ahead of their predecessors. Moreover, the performance difference between the processors with the same declared nominal clock speed is quite substantial and rests at about 15% in favor of the newcomers with Ivy Bridge microarchitecture.
Now that the eighth version of the popular scientific Mathematica suite is available, we decided to bring it back as one of our regular benchmarks. We use MathematicaMark8 integrated into this suite to test the systems performance:
Wolfram Mathematica is traditionally one of those applications that do not “digest” Hyper-Threading technology too well. Therefore, the diagram above shows Core i5-3570K in the first position. In fact, other 3000-series Core i5 processors also did very well in this benchmark. All of them not only outperform their predecessors, but also leave behind the top Core i7 processor on Sandy Bridge microarchitecture.
We measured the performance in Adobe Photoshop CS6 using our own benchmark made from Retouch Artists Photoshop Speed Test that has been creatively modified. It includes typical editing of four 24-megapixel images from a digital photo camera.
New Ivy Bridge microarchitecture provides the third-generation Core i5 processors with about 6% performance advantage over their predecessors working at the same clock frequency. And if we compare the processors with the same pricing, then the ones with new microarchitecture will be in an even more favorable position reaching over 10% performance advantage over the 2000-series Core i5 processors.
The performance in Adobe Premiere Pro CS6 is determined by the time it takes to render a Blu-ray project with a HDV 1080p25 video into H.264 format and apply different special effects to it.
Non-linear video editing is a well-paralleled task, so the new Core i5 processors with Ivy Bridge inside are unable to catch up with Core i7-2700K. However, they are about 10% faster than their predecessors (provided we compare models with the same clock frequency).
In order to measure how fast our testing participants can transcode a video into H.264 format we used x264 HD Benchmark 5.0. It works with an original MPEG-2 video recorded in 1080p resolution with 20 Mbps bitrate. I have to say that the results of this test are of great practical value, because the x264 codec is also part of numerous popular transcoding utilities, such as HandBrake, MeGUI, VirtualDub, etc.
The results of HD video content transcoding tests are quite common. The advantages of Ivy Bridge microarchitecture result into about 8-10% higher performance of the new Core i5 processors compared with the older ones. The only unusual thing here is the high performance of the FX-8150, which outperforms even Core i5-3570K during the second transcoding pass.
Following our readers’ requests, we’ve added a new HD video benchmark to our tests. SVPmark3 shows the computer performance in the SmoothVideo Project application which makes videos smoother by adding new intermediary frames. The numbers in the diagram reflect the speed of processing Full HD videos without the graphics card’s help.
This diagram is very similar to the results of the second transcoding pass using x264 codec. This indicates clearly that most tasks dealing with HD video content processing create very similar computational load.
We will test computational performance and rendering speeds in Autodesk 3ds max 2011 using the special SPECapc for 3ds max 2011 benchmark:
Frankly speaking, we can’t say anything new about the performance during final rendering. The results are pretty standard overall.
We use special Cinebench 11.5 benchmark to test final rendering speed in Maxon Cinema 4D suite.
The Cinebench results also reveal nothing new. New 3000-series Core i5 processors are again much faster than their predecessors. Even the very junior Core i5-3450 model is confidently ahead of the Core i5-2500K.
According to Intel, one of the major advantages of the finer 22 nm process used to manufacture Ivy Bridge generation processors is the lower heat dissipation and power consumption of the semiconductor dies. This is certainly reflected in the official third-generation Core i5 specifications: their TDP has been set at 77 W instead of the 95 W used previously. So, the new Core i5 will undoubtedly be more energy-efficient than their predecessors. But how big of an advantage will they boast in real life? Should we consider energy-efficiency of the 3000-series Core i5 processors a serious competitive statement?
To answer all these questions we performed a round of special tests. The new digital power supply unit from Corsair – AX1200i – allows monitoring consumed and produced electrical power, which we use actively during our power consumption tests. The graphs below (unless specified otherwise) show the full power draw of the computer (without the monitor) measured after the power supply. It is the total power consumption of all the system components. The PSU's efficiency is not taken into account. The CPUs are loaded by running the 64-bit version of LinX 0.6.4-AVX utility. Moreover, we enabled Turbo mode and all power-saving technologies to correctly measure computer's power draw in idle mode: C1E, C6 and Enhanced Intel SpeedStep.
When idle, the systems with all processors participating in our today’s test session consume about the same amount of power. Of course, the readings are not fully identical, the numbers differ by tenths of a watt, but we decided not to transfer them onto the diagram, because these insignificant deviations can most likely be attributed to the measuring error margin than the actual physical processes. Moreover, since all tested processors demonstrated very close power readings, the efficiency of the mainboard voltage regulator configuration starts to affect the power consumption levels, too. Therefore, if you are really concerned with the power consumption in idle mode, you should make sure that you have a mainboard with the most efficient voltage regulator circuitry design. As for the processor, our tests show that any LGA 1155 CPU could do just fine.
In case of single-threaded load, when the CPUs supporting Turbo mode boost their frequency to the maximum, we see noticeable differences in power consumption. First of all, we notice enormous energy appetite of the AMD FX-8150. As for the LGA 1155 CPU models, the ones using 22 nm semiconductor dies are really much more energy-efficient. The difference in power consumption of the quad-core Ivy Bridge and Sandy Bridge processors working at the same clock frequencies is about 4-5 W.
Full multi-threaded computational load increases the power consumption delta even more. A system equipped with a third-generation Core i5 is about 18 W more energy-efficient than a similar platform with a previous-generation CPU. So, desktop Ivy Bridge processors are simply unattainable in terms of performance-per-watt.
When talking about contemporary LGA 1155 processors it is important to pay due attention to the integrated graphics cores, which has become faster and more advanced with the introduction of Ivy Bridge microarchitecture. However, at the same time Intel prefers to equip their desktop processor modifications with a cut-down graphics core featuring only 6 execution units out of 16. In fact, only Core i7 and Core i5-3570K processors have fully-functional integrated graphics cores. The majority of desktop Core i5 CPUs from the 3000-series will most likely prove pretty slow in 3D graphics applications. However, it is quite possible that even the limited graphics functionality will be quite sufficient for some users who do not regard integrated graphics core as a replacement for a discrete 3D graphics accelerator.
We decided to start testing integrated graphics with 3DMark Vantage benchmarks. 3DMark scores are a very popular way of estimating average gaming performance of the graphics cards. And we chose Vantage suite because it uses DirectX 10 supported by all the participating graphics accelerators including integrated graphics in Core processors with Sandy Bridge microarchitecture. Note that besides a full set of Core i5 processors with their respective integrated graphics cores we also included the results of a Core i5-3570K based system with a discrete Radeon HD 6570 graphics accelerator. This configuration will serve as a reference point that will allow us to rank Intel HD Graphics 2500 and HD Graphics 4000 cores among discrete graphics accelerators.
HD Graphics 2500 core that Intel uses for the majority of their desktop processors demonstrates 3D performance similar to that of the HD Graphics 3000. However, the top core from Ivy bridge processors, HD Graphics 4000, becomes a tremendous leap forward, as its performance is almost twice as good as that of the best previous generation graphics core. However, any of the existing Intel HD Graphics modifications still cannot deliver acceptable 3D performance for desktop uses. For example, Radeon HD 6570 graphics card that belongs to the lower price segment and costs about $60-$70 is still much faster.
In addition to the synthetic 3DMark Vantage we ran a few real gaming tests. We used low image quality settings and 1650x1080 screen resolution, which in our opinion is the lowest acceptable screen resolution for the majority of desktop users these days.
Overall, we see about the same situation in all games. The top modification of the Intel HD Graphics core integrated into Core i5-3570K delivers pretty good average fps rate (for an integrated solution). However, Core i5-3570K is the only processor in the third-generation Core i5 family that can deliver acceptable performance in most today’s games (provided you will agree to certain image quality restrictions). All other processors with the HD Graphics 2500 core inside that has fewer execution units run at almost half the speed, which is not enough by all means.
The advantage of the HD Graphics 4000 over the previous-generation HD Graphics 3000 core varies in a pretty wide range and is around 90% on average. The junior graphics core model from the Ivy Bridge family, which is currently used in a majority of desktop Core i5 processors, HD Graphics 2500, can easily compete against the previous flagship graphics core from Intel. As for the previous mainstream graphics core, Intel HD Graphics 2000, its performance in games now strikes as extremely low. Our tests show that it falls about 50-60% behind the new HD Graphics 2500.
In other words, 3D performance of the graphics core in the new Core i5 processors has increased significantly, but it is still way to low to compete against the fps rate that you can get from a discrete Radeon HD 6570 graphic accelerator. Even HD Graphics 4000 integrated into Core i5-3570K cannot become a worthy alternative to a low-end discrete 3D accelerator. The more common modification of the new Intel HD Graphics core is totally unfit for most games.
However, not all users regard Intel’s integrated graphics cores as gaming 3D graphics accelerators. Many users are attracted to HD Graphics 4000 and HD Graphics 2500 for their multimedia capabilities, to which there is absolutely no alternative in the low-end price segment. Here we first of all mean the Quick Sync technology intended for fast hardware video transcoding into AVC/H.264 format, which second version has been implemented in Ivy Bridge processors. Since Intel promises that new integrated graphics cores will offer much higher transcoding speeds, we paid special attention to testing Quick Sync this time.
During our test session we measured the time it took to transcode one 40-minute episode of a popular TV-show from 1080p H.264 with 10 Mbps bitrate into a format compatible with Apple iPad2 (H.264, 1280x720, 3 Mbps). We used Cyberlink Media Espresso 6.5.2830 utility that supports Quick Sync.
The situation here is radically different from what we have just seen in games. If previously Intel didn’t differentiate Quick Sync in CPU with different graphics core modifications, now everything has changed. The performance of this technology in HD Graphics 4000 has almost doubled compared with its speed in HD Graphics 2500. Moreover, regular Core i5 processors from the 3000-series equipped with HD Graphics 2500 transcode HD video with the help of Quick Sync at about the same speed as their predecessors. The only way you can actually see the performance boost is by Core i5-3570K, which features “advanced” HD Graphics 4000 core.
You can overclock Core i5 processors from Ivy Bridge generations following one of the two different strategies. The first one is related to Core i5-3570K processors, which is designed as an overclocker model right from the start. This processor has an unlocked clock frequency multiplier and its frequency may be increased above the nominal value according to LGA 1155 specific algorithm: we increase the CPU clock frequency by raising the multiplier and at the same time improve CPU cooling and increase the CPU Vcore accordingly to achieve stability.
Without any Vcore adjustments our test Core i5-3570K overclocked to 4.4 GHz. To ensure stability in this mode we only had to switch the mainboard Load-Line Calibration to High.
By pushing the CPU core voltage to 1.25 V we achieved stability at higher 4.6 GHz clock frequency.
This is a pretty common result for Ivy Bridge processors. These CPUs overclock a little bit less than those from Sandy Bridge generation. Supposedly, it is caused by the smaller semiconductor die as a result of the migration to finer 22 nm process, which required increased heat flow density during overclocking. At the same time the internal thermal interface used by Intel inside their CPUs, just as the regular means of dissipating heat from the surface of the CPU packaging, didn’t really help much.
However, 4.6 GHz is still a very good result, especially keeping in mind that Ivy Bridge processors are about 10% faster due to their microarchitectural improvements compared with Sandy Bridge CPUs working at identical clock frequencies.
The second overclocking algorithm should be applied to all other Core i5 processors, which do not have an unlocked clock frequency multiplier. Although LGA 1155 platform doesn’t handle the increase in the base clock generator frequency too well and often loses stability at even 5% frequency increase, it is still possible to overclock non-K Core i5 processors. The thing is that Intel allows slightly increasing their multiplier by setting it no more than 4 points above the nominal value.
Keeping in mind that in this case Turbo Boost technology remains fully operational and allows overclocking Core i5 Ivy Bridge processors by 200 MHz even when all processor cores are utilized, the overall clock frequency “boost” can add up to as much as 600 MHz. In other words, Core i5-3570 may be overclocked to 4.0 GHz, Core i5-3550 – to 3.9 GHz, Core i5-3470 – to 3.8 GHz, and Core i5-3450 – to 3.7 GHz. And this is exactly what we saw during our overclocking experiments:
I have to say that the non-overclocker processors are even easier to overclock than the Core i5-3570K. Relatively low increase in the CPU clock speed doesn’t cause any stability issues even with the nominal Vcore settings. Therefore, the only thing you will most likely need to do during non-K Core i5 Ivy Bridge processors overclocking is to change the multiplier in the mainboard BIOS. Although the obtained result won’t be a record-breaker, it will most likely be more than satisfying for the majority of non-extreme users.
We have already pointed out multiple times that Ivy Bridge microarchitecture has become a successful evolution of the Intel processors. 22 nm production process and numerous microarchitectural improvements turned the new processors into faster and more energy-efficient products. This is true for any Ivy Bridge processors in general and desktop Core i5 processors from the 3000-series discussed today in particular. If we compare the new Core i5 processor family with what we had a year ago, we can easily single out a number of significant improvements.
First, new Core i5 based on Ivy Bridge microarchitecture are faster than their predecessors. Although Intel didn’t really increase their clock speeds, the newcomers are about 10-15% faster. Even the slowest third generation desktop Core i5, Core i5-3450, outperforms Core i5-2500K in most benchmarks. And the top models in the new lineup can often compete even against the higher-end Core i7 CPUs on Dandy Bridge microarchitecture.
Second, new Core i5 processors have become much more energy-efficient. Their 77 W TDP does have immediate effect on their performance. Computers with Core i5 Ivy Bridge processors inside consume several watts less power than the same systems with Sandy Bridge CPUs. Moreover, this difference may increase to as much as 20 W under peak operational loads, which translates into substantial power savings according to today’s standards.
Third, new processors have a significantly improved integrated graphics core. The junior modification of the new Ivy Bridge graphics core works at least as good as HD Graphics 3000 from the second-generation top Core processors. Besides, it supports DirectX 11 and therefore offers more up-to-date functionality. As for the flagship HD Graphics 4000 core used in Core i5-3570K processors, it is powerful enough to deliver acceptable gaming performance with certain image quality restrictions, of course.
The only arguable thing about the third-generation Core i5 processors is a slightly lower overclocking potential than that of Sandy Bridge processors. However, it is relevant only for the overclocker Core i5-3570K model, which multiplier is unlocked. Nevertheless, higher performance of the Ivy Bridge microarchitecture easily makes up for this little issue.
In other words, if you are looking for a mainstream LGA 1155 CPU, there is no real reason why you should prefer the older Sandy Bridge based processors to the new third-generation Core i5. Especially, since Intel’s pricing on the new Core i5 modifications are quite reasonable and very close to those of the previous generation processors.