Introduction

Quad-core processors are progressively more interesting from a technical perspective. Many customers from the mainstream market were discouraged but yet of the price. This will officially change AMD today. For this purpose, the manufacturer sends the small offshoot of the very successful acting this year Phenom II processors in the race and calls this “Athlon” descendants. The complete name of the CPU, AMD today presents, is “AMD Athlon II X 4 620″.

This new model with 2.6 GHz and (in contrast to the Phenom II) without L3 cache are we on the following pages closer look. At a cost of just 85 euro, four cores Prize regions penetrate before, in which so many cores into a processor have been never seen. In addition, succeeding the new processor to deliver a good performance so that at the end of the whole package is true?

In addition to AMD Germany is thanks to ASUS, Corsair, OCZ and Noctua, who helped us with all kinds of hardware for this test.

The new Athlon II X 4

What exactly is the name AMD Athlon II X really 4 620? “Athlon” has been the new edition of Phenom in the form of Phenom II earlier this year for a processor without L3 cache and thus the preferred variant. The “X 4″ represents four cores, because there we know already X 2 models of Athlon. X 3 variants with three cores will follow shortly. The numeral “II” stands for the new generation of AMD processors, which was introduced at the beginning of 2009. This refers primarily to the 45 nm manufacturing has ironed out many minor weaknesses of the last 65-nm fabrication. Through this production method, the measure could be increased significantly, at the same time, the power consumption of the subjects fell noticeably. Overall, AMD has brought more competition in the market by this step.

Officially, the “Propus” is based on the new Athlon with four cores. This code name a new behind the who completely renounces L3 cache. Succeed AMD for some time very skillfully and efficiently, to present a maximum of models with “minimal” effort. Almost all desktop processors this year are based on only three different production masks. The “Deneb” mask, which is the identical to one of the “Shanghai”-server processors goes above all. There are currently the most spin-off in the trade of this quad-core mask. There the entire flagship series of the Phenom II is 4 900, but also the variants with the model numbers 800 X on the one hand, 700 and 500 from her. The second already available Panel is designed for the dual-core processors code-named “Regor” and most likely also for the new Sempron, which will again be without a core.

The third and today officially presented mask should be the most important after the Deneb. After all, not just four-core processors are to be considered from this, the Dreikerner of “Rana” and low-power variants with a TDP of 45 Watts are expected. Special on the Propus (and already on the Regor manufacturing mask) is that chip area and number of transistors thanks to completely missing L3 cache is much lower and thus the effective yield of a wafer is much higher than for the Phenom II. This is also the cost of the AMD next to the customers (lower prices and low power consumption) good.

< div >

The above table shows significantly on the differences. The a new Athlon II with four and three seeds is as great as II, the number of transistors used is a Phenom with 300 to 758 million even below 40 percent of the original two-thirds. Without naming specific figures, AMD says itself that this new Rana is “less expensive to manufacture” and therefore enables the best prices for the processor family. By the Heatspreader can be seen directly the nowadays but hardly a. A clever scribe in the Forum of Xtremesystems.org has three types of CPU from AMD but once carried out and gets rid of the Heatspreaders. To light the following came:

< div >

Source: Xtremesystems.org

In theory everything is clear. But the theoretical information is always only one side of the coin, in our hands the Athlon II must show X 4 620, what he can. Already in the BIOS we noted with dismay that the new Athlon II exacts 1.375 volt himself with four cores at only 2.6 GHz. But unfortunately several BIOS updates brought no change, the tension seems intended by AMD so really. On paper the Athlon II would be from X 4 620 with 1.375 volts then the number two in terms of highest default voltage work, only of the 800 MHz higher clocked flagship AMD Phenom II X 4 965 Black Edition outbid – it takes 1.4 volts. 4 Speak specifications the day before the end of the NDA of the press provided the new Athlon II X of a workspace of tension, which is to be between 0,952 and 1,425 V. Given this high voltage and the specified not just low TDP of 95 Watts it seems already questionable whether it hoping for something with the low power consumption. The following pages will bring clarity on this point.

AMD at the power wants to do something / must show the also planned models of Athlon II X 4, but not officially presented. As a variant with a TDP of 45 watts could fix hardly the voltage at 1, 3 x volts. Here should be selected on the part of the manufacturer very strongly, especially as the clock rate goes down at the same time significantly. The section of Undervolting in our test shows already that this is very well possible. With about 20 percent less voltage, we could run the processor equally lasting clock which has of course dramatically positive consequences.

Parallel to the present of featured Athlon II X 4 620 we simulate the upcoming model with three cores at an identical time of 2.6 GHz. This is easily via the BIOS, in which you already can choose, how many active cores of the processor is designed to provide for. A simple change of the number from four to three and even CPU-Z is at least partially convinced that we have a “Rana” ahead of us. The so generated 2.6 GHz fast three-core model X may be directly with 3 710 the Phenom II identical clock but make along with 6 MB L3 cache.

Test system

We had sent our previous test system in the earned pension with the launch of Nehalem, the new AMD test system will benefit from this. Both have seen last year a total 36 processors, 21 of which made by Intel, AMD 15. In the course of the months in which we no changes undergo such a test system, many changes in the market have taken place, it was time for a change not only in the light of the new generation of processors. Of course a complete new generation for the replacement of the complete system offers always the best, so that we have moved the formerly early plans from summer into fall. Here we have used also many suggestions of the reader, which we received in the last few months. At this point, we want to give again some notes to the new test system and explain why we have chosen for these components.

One of the largest and often installed was the older tagan power supply 480 Watts. This is sent in the probably well-deserved retirement and replaced by a brand new power supply of cooler master. The UCP series has a “Silver 80Plus” at the beginning of the summer of 2008 as one of the first desktop power supplies get certification. This is a very high 89 percent efficiency at a duty of 50 percent, but even under low load (20%) and full load (100%) is evident over the UCP power supply with 87 and 85% respectively. For our test, we rely on the smallest version with 700 watts, because sometimes clearly fall a stronger power supplies especially under low load in the efficiency. The 20% threshold is located in the 700-Watt volunteers at 140 Watts, it is a value of not even achieved with the new test system for smaller processors under full load with Prime in the 1,100-Watt model already on 220 watt high screwed.

The second important point is the new graphics card. After we complete last year had a not just energy-efficient ATi Radeon HD 2900 XT in use, we use 9800 GTX + NVIDIA’s first (relatively low-power) 55-nm solution in the high-end segment, the GeForce. In particular the reduced line recording (sh. also in the article of HT4U), also with regard to the competitors of ATi in the form of the HD 4850 and the consistently faster performance were important reasons for the selection. All other eligible models (HD 4870, GTX 260, GTX 280) provide higher performance, but only with disproportionate rising power consumption while – not matter in a processor test but also.

In order to meet the new triple-channel RAM, the transition to Windows Vista in the 64-bit variant was necessary almost imperative on the software side. Also this step we went with the new test system, so that from now on be – fitted the three slots on the X 58 boards with each two gigabytes all other Intel boards including new P55 chipset for socket LGA1156 fall back to two of these modules for the dual-channel mode. The difference of 4 to 6 GByte course exists on paper, in benchmarks but virtually not to remember. In addition, we believe that such constellations in the near future are likely to be the most common variants, or in the form of DDR2 memory with the old established core 2 are already Duo and core 2 Quad. Therefore, all (older) AMD processors also come along with potent DDR2 memory, provides the maximum 1,066 MHz. Because so far only the Phenom and Phenom II take advantage of this store, working older models with a lesser frequency, but with better timing. Socket AM3, a Board from ASUS to which opts for the 790FX and the SB750 is us at the same time. There is also a virtually identical spin-off for socket AM2, which logically sets on DDR2. These two boards we already presented news in one in images, should differences in addition to the Referenzmainboard on the basis of 790GX.

&Lt; div > “Asus M4A79 & M4A79T Deluxe” Photo Gallery (20 pictures)

To enable a possible fair and realistic comparison between the competitors, all in a closed MIDI tower with factory fan Assembly tests (a fan back sucking, blowing one on the hard drive cage in front), to be attentive so also on thermal problems in the car. A “cooler master stacker 832 RC”, which was made available to us by Caseking is used. The housing allows the use of up to nine 120-mm fans, of which the two used fans are included.

• Complete breakdown of the test system and the components used.
  To keep the overview in the charts, we have many processors and us dedicated to only the environment of new processors.
  • Processor
    • Quad-core
      • AMD
        • AMD Phenom II X 4 965 Black Edition – 3.40 GHz, 2 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Deneb C2)
        • AMD Phenom II X 4 955 Black Edition – 3.20 GHz, 2 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Deneb C2)
        • AMD Phenom II X 4 920-2.80 GHz, 2 MB L2 cache, 6 MB L3 cache, HT 1.8 GHz (Deneb C2)
        • AMD Phenom II X 4 810-2.60 GHz, 2 MB L2 cache, 4 MB L3 cache, HT 2.0 GHz (Deneb C2)
        • AMD Phenom X 4 9950 Black Edition – 2.6 GHz, 2 MB L2 cache, 2 MB L3 cache, HT 2.0 GHz (Agena B3)
        • AMD Athlon II X 4 620-2.60 GHz, 2 MB L2 cache, HT 2.0 GHz (Rana C2)
      • Intel
        • Intel Core i7-870 – 2.93 GHz, 1 MB L2 cache, 8 MB L3 cache (Lynnfield B1)
        • Intel Core i5-750 – 2.66 GHz, 1 MB L2 cache, 8 MB L3 cache (Lynnfield B1)
        • Intel Core 2 Quad Q9550 – 2.83 GHz, 12 MB L2 cache, FSB1333 (Yorkfield C1)
        • Intel Core 2 Quad Q8200 – 2.33 GHz, 4 MB L2 cache, FSB1333 (Yorkfield r0)
    • Triple-core
      • AMD
        • AMD Phenom II X 3 740 Black Edition – 3.00 GHz, 1.5 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Heka C2)
        • AMD Phenom II X 3 720 Black Edition – 2.80 GHz, 1.5 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Heka C2)
        • AMD Phenom II X 3 710-2.60 GHz, 1.5 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Heka C2)
        • AMD Athlon II X 3 420-2.60 GHz, 1.5 MB L2 cache, HT 2.0 GHz (Rana C2)(simulated)
    • Dual-core
      • AMD
        • AMD Phenom II X 2 550 Black Edition – 3.10 GHz, 1 MB L2 cache, 6 MB L3 cache, HT 2.0 GHz (Callisto C2)
        • AMD Athlon II X 2 250-3.00 GHz, 2 MB L2 cache, HT 2.0 GHz (Regor C2)
        • AMD Athlon II X 2 240-2.80 GHz, 2 MB L2 cache, HT 2.0 GHz (Regor C2)
        • AMD Athlon X 2 4850e – 2.50 GHz, 1 MB L2 cache, HT 1 GHz (Brisbane G2)

          All Athlon, Athlon-II-Phenom-and Phenom II processors have been, unless otherwise indicated, with C & Q and measured C1E disabled (out). We had a comprehensive study on this subject at the beginning of the year in the program. The memory controller was operated on all AMD systems in the mode of “UnGanged”. The particular choice of memory, and motherboard (chipset) is specified separately.

      • Intel
        • Intel Core 2 Duo E8600 – 3.33 GHz, 6 MB L2 cache, FSB1333 (Wolfdale E0)
        • Intel Core 2 Duo E8400 – 3.00 GHz, 6 MB L2 cache, FSB1333 (Wolfdale E0)
        • Intel Core 2 Duo E7600 – 3.06 GHz, 3 MB L2 cache, FSB800 (Wolfdale r0)
        • Pentium E6300 – 2.80 GHz, 2 MB L2 cache, FSB800 (Wolfdale r0)
  • Motherboard
    • Intel
      • ASUS P6T Deluxe (Intel X 58 chipset)-revision 1.02G, BIOS: 0703 (October 15, 2008), BIOS: 1504 (April 24, 2009) d0 stepping and new processors require
      • ASUS P7P55D (Intel P55 chipset)-revision 1.02G, BIOS: 0501 (August 26, 2009)
      • Asus P5E3 premium (Intel X 48 chipset)-needed revision 2.00G, 0605 (October 9, 2008) BIOS, BIOS 0803 (May 21st, 2009) for E7600 and E6300
    • AMD
      • ASUS M4A79 Deluxe (AMD 790FX + SB750) – socket AM2 + – revision 1.01G, BIOS 0801 (January 8, 2009)
      • ASUS M4A79T Deluxe (AMD 790FX + SB750) – socket AM3 – revision 1.01G – BIOS 0503 (January 9, 2009), Phenom II X 4 945/955 BIOS 0902, for Athlon II and Phenom II X 2 BIOS 1303, BIOS 0066 for 965 BE/955 BE
  • Memory
    • Intel
      • 3 x 2,048 MB DDR3-1600 Corsair Dominator (CL8-8-8-24-2T, 1.60 volts, Triple-channel)
      • 3 x 2,048 MB DDR3-1066 Corsair Dominator (CL7-7-7-20-1T, 1.50 volts, Triple-channel)
      • 2 x 2,048 MB DDR3-1600 Corsair Dominator (CL8-8-8-24-2T, 1.60 volts, dual-channel)
      • 2 x 2,048 MB DDR3-1333 Corsair Dominator (CL7-7-7-20-1T, 1.50 volts, dual-channel)
      • 2 x 2,048 MB DDR3-1066 Corsair Dominator (CL7-7-7-20-1T, 1.50 volts, dual-channel)
    • AMD
      • 2 x 2,048 MB DDR3-1600 OCZ Platinum (CL8-8-8-24-1T, 1.60 volts, dual-channel)
      • 2 x 2,048 MB DDR3-1333 Corsair Dominator (CL7-7-7-20-1T, 1.50 volts, dual-channel)
      • 2 x 2,048 MB DDR3-1066 Corsair Dominator (CL7-7-7-20-1T, 1.50 volts, dual-channel)
      • 2 x 2,048 MB DDR2-1066 OCZ Gold (CL5-5-5-15-2T, 2.1 volt, dual-channel)
      • 2 x 2,048 MB DDR2-800/DDR2-750/DDR2-714 OCZ Gold (CL4-4-4-12-1T, 1.8 Volt, dual-channel)
  • Graphics card
    • NVIDIA GeForce 9800 GTX +, 55 nm, 512 MB GDDR3, engineering sample
  • Peripherals
    • Samsung HD501LJ, 500 GB (SATA-II hard disk)
    • MSI DR8-A (burner)
  • Power supply
    • Cooler master UCP 700 watts
  • CPU cooler
    • Noctua NH-U12P with NF-P12 fan
  • Driver versions
    • NVIDIA GeForce 178.24
    • Intel chipset driver 9.1.0.1007
  • Software
    • Microsoft Windows Vista 64-bit, Service Pack 1
    • Microsoft DirectX August 2008

Benchmarks

In addition to the completely new designed test system, we have subjected also the benchmark course an update. Some theoretical tests are still with the part of, because these are usually a step as any real application. There, new features that have found a place in processors be checked first in their performance. Because these theoretical tests but use not much users in front of the PC, we consult many practical applications. Various multimedia applications, but also everyday things like packaging files include today in the first place. The test is rounded off by some games where we have picked out a selection of games with integrated benchmark function, as well as games, not only the average frames per second, but also the minimums and maximums were identified where using Fraps.

As in the past, each benchmark is performed five times, to be able to find any break in the performance. This happens in almost all applications unfortunately more common than expected, both the negative and the positive direction. The average of the five tests (without extreme outliers) is then applied in the single charts. Games used the resolution of 1,280 x 1,024 pixels, because almost only still the graphics card plays a role in higher resolutions, have lower resolutions but no relation to reality. Typically, we test it, if not closer described, with the highest details. The values of the AA/AF may thus vary from game to game, however, the different setups to be a small simulation for it, which these days can change with each new game. All tested processors need along with the video card in these tests to prove that they can cope with a few quality-enhancing features, as well as with a maximum.

As far as possible, to eliminate the graphics card, we have introduced also an additional point: the unrealistic games tests. Four applications of different ages show its result in minimum and maximum details at very small resolution.

• Synthetic
  • 3DMark06 1.10
  • 3DMark Vantage 1.01
  • SiSoftware Sandra 2009.1.15.52 beta
  • Cinebench R10
  • Super PI xmod 1.5 XS
  • MaxxPi m² 1.11ß
  • wPrime 2.00
• Office and multimedia
  • PCMark Vantage 1.00
  • Tsunami MPEG video encoder Xpress 4.5.2.255
  • DivX 6.8.4.7
  • iTunes 8.0.1
  • WinRAR 3.80
  • POV-ray 3.7 Beta 29
  • TrueCrypt 6 0a
• Games
  • Assassin’s creed 1.0
  • Bioshock 1.0
  • Company of Heroes 1.71
  • Crysis demo
  • Lost planet: colonies 1.0
  • Race driver grid 1.2
  • Sacred 2 demo
  • World in conflict 986
• other required tools
  • CPU-Z 1.52
  • Fraps 2.96
  • Prime

Synthetic tests

3DMark06

Although now, the well-known series of benchmark from Futuremark has appeared in the vantage version, but “3DMark06″ is still very popular. Measure the performance of the graphics card four sequences of the six test scenes and show a graphic splendour that is sometimes still unparalleled. To achieve that the Finns put which is why not only massively the shader model 3.0 is used on modern 3D technology, extremely elaborate textures, spectacular particle effects, shadow complex computations and as another highlight “High Dynamic Range Rendering” – short HDRR – are used. Thus, Futuremark opts FP16 HDR, which delivers a very high picture quality, but it is also expensive to calculate. There are more details about this program in our detailed article.

Download: 3DMark06

3DMark Vantage

After some years on the knob has the long-serving 3DMark06 and affects not only the graphic now somewhat dusty but in addition, the CPU limit for fast graphics cards is becoming more and more apparent, it was high time for a successor. Accordingly the Finnish manufacturer Futuremark has brought the 3DMark Vantage on the market after a long wait, who is from the outset been programmed for the Direct3D 10 API. Graphically, the two game tests accordingly much for the eye, and in particular the second test can set highlights offer. FP16 HDR, depth of field, Parallax Occlusion Mapping, a physical simulation on the GPU, various Shadereffekten, and much more, 3DMark Vantage easily brings the 3D hardware to the performance limit. We test the program in the performance preset. There are more details about this program in our detailed article.

Download: 3DMark Vantage

Cinebench R10

The popular, from Germany, Maxon Cinema4D is represented within our benchmark show in the form of Cinebench R10. The software uses up to 16 processors to the raytracing and thus benefit from all available on the market desktop processors from AMD or Intel currently. In our test, we present the version Cinebench R10 published in mid-2007. The diagrams show the test as usual on the one hand with only one processor core, on the other hand also the multi-core test, which uses Hyper-Threading.

Download: Cinebench

SiSoft Sandra 2009

No matter whether it comes to motherboard, memory, hard drive, peripherals, cards, processor, network, interfaces, BIOS, Windows or DirectX, SiSoft Sandra has extensive answers at hand. For much of the hardware in the PC, there is also a benchmark tests, the PC test can on its performance compared to some reference computers. All’ these values are however almost exclusively purely theoretical nature and have little relation to the practice, but processors in their theoretical possibilities can be compare well. Also, programs like Sandra usually significantly more offer ways to test newly integrated features of the processors before they are integrated in months or years in real programs. We decided from the extensive repertoire for the testing of arithmetic, multimedia, and cryptography, as well as for the memory, the memory latency and the cache and memory access. Still a beta version of October 22, 2008 is used, as these for the first time the new SSE4.1. uses 2 instructions of Nehalem.

Download: SiSoft Sandra