Hardware Upgrade - Logo
« Back   -  Print
The Last of the K8: The AMD Athlon 64 X2 6000+
The Last of the K8: The AMD Athlon 64 X2 6000+
"With the end of the K8 nearing, AMD releases its last Athlon 64 X2 processor based on the K8 architecture. The new processor is the fastest Athlon 64 to date, and reaches 3GHz, all at a cost lower than $500. "
Page 1 - Introduction

Near the end of 2006, AMD began releasing new versions of its Athlon 64 X2 processors. These new processor versions were the first AMD desktop chips to be manufactured using a 65 nanometer process. The move to 65nm has allowed AMD to reduce the costs associated with manufacturing its chips. In addition, it has also allowed AMD to produce 65nm Athlon processors which are more power efficient than their 90nm counterparts. AMD, however, has chosen not to release 65nm versions of every one of its processors. Instead, it has only released 65nm Athlon 64 X2 chips up until its Athlon 64 X2 5000+ model. Any model number higher than 5000+ is still constructed using a 90nm process. There are a number of reasons behind this move. For one, it allows AMD to reduce the costs of its already cheaper processors, therefore creating more aggressive pricing. In addition, as 65nm chips consume less power, AMD has been able to reduce the power consumption of these chips to 65 watts.

As we mentioned earlier, any model higher than 5000+ is still being manufactured using a 90nm process. This includes the Athlon 64 6000+ we are reviewing today. AMD will not begin creating 65nm high-end processors until the release of its Athlon 64 X2 dual and quad core chips based on the K10 architecture. The release of K10, however, is not expected to take place until halfway through 2007. This means that AMD's current 90nm Athlon 64 X2 processors will have to counter Intel's Core 2 Duo chips, all of which use a 65nm die, for another six months.

athlon64_6000+_cpu_s.jpg (45984 bytes)

Today AMD releases the Athlon 64 X2 6000+; the last processor of the Athlon 64 X2 series which is based on the well known K8 architecture. As we said in our review of AMD's Quad FX platform, the Athlon 64 X2 series has become AMD's flagship desktop processor line. Keeping this in mind, the Athlon 64 X2 6000+ is AMD's fastest and most powerful K8-based desktop chip yet. The processor is also faster than the Athlon 64 FX62, which up until recently was AMD's high-end desktop processor.

cpuz_6000+_1.gif (17487 bytes)

The new CPU comes officially clocked at 3 GHz, making it the fastest processor available for AM2, with 1 MB of L2 cache per core. This is the same speed that AMD's Athlon 64 FX-74 chip, which was introduced last year near the end of November, runs at. The core is the same Windsor core which AMD has already used for a couple of its processors, namely the Athlon 64 X2 and Athlon 64 FX-62 solutions. As we briefly discussed earlier, the Athlon 64 X2 6000+ still uses a 90nm manufacturing process. The official price for a single Athlon 64 X2 6400+ in a quantity of 1,000 is $464. The next AMD part with the most similar amount of features, the FX-74, is available at a price of $1,000. Keep in mind, however, that the FX-74 figure is for two processors since a complete package must be bought. In addition, the FX-74 can only be used with AMD’s Quad FX platform.

AMD declares the maximum thermal power of the Athlon 64 X2 6000+ at 125 W, a figure that is identical to the maximum thermal power of the Athlon 64 FX-62 which also happens to be 200 MHz slower. Nominal voltage is between 1.35V and 1.40V, the same value for other Athlon 64 X2 90nm processors. The integrated memory controller uses a divider set at 8.

Page 2 - Intel/AMD Processors

Before delving further into the review, we want to provide a brief recap of the processors currently available for sale and discuss their features.

The following table shows the technical characteristics of Intel's current lineup of processors.

Processor Number of Cores Clock Bus
L2 Cache 		multiplicator manufacturing process Socket TDP
Core 2 Extreme QX6700 4 2.67 GHz 1,066 MHz 2x4M 10x 65 nm 775 LGA 130 Watt
Core 2 Quad Q6600 4 2.4 GHz 1,.066 MHz 2x4M 9x 65 nm 775 LGA 105 Watt
Core 2 Extreme X6800 2 2.,93 GHz 1,066 MHz 4M 11x 65 nm 775 LGA 75 Watt
Core 2 Duo E6700 2 2.67 GHz 1,066 MHz 4M 10x 65 nm 775 LGA 65 Watt
Core 2 Duo E6600 2 2.4 GHz 1,066 MHz 4M 9x 65 nm 775 LGA 65 Watt
Core 2 Duo E6400 2 2.13 GHz 1,066 MHz 2M 8x 65 nm 775 LGA 65 Watt
Core 2 Duo E6300 2 1.86 GHz 1,066 MHz 2M 7x 65 nm 775 LGA 65 Watt
Core 2 Duo E4300 2 1.,8 GHz 800 MHz 2M 9x 65 nm 775 LGA 65 Watt

All the processors use a quad-pumped 1,033 MHz front side bus. The only exception is the E4300 which uses an 800 MHz FSB. In addition, all of the CPUs use a unified cache, with a total L2 cache of 2MB for the E6400, E6300, and E4300. The remaining CPUs all have a 4MB L2 cache. The two quad-core architectures from Intel are, however, slightly different. This is because they have 2 separate 4 MB L2 caches, adding up to a total of 8 MB. The QX6700/Q6600 are this way simply because Intel placed two dual-core die side by side on the same package to create a total of four cores. Each die has a shared cache of 4 MB.

The declared TDP of all Intel Core 2 Duo models is 65 Watts. The only exception is the Core 2 Extreme X6800 which is declared at 75 watts. The TDP grows larger for the quad-core chips. The Core 2 Quad Q6600 has a TDP of 105 watts, while the Core 2 Extreme QX6700 has one of 130 watts. These values are quite impressive compared to the 130 watt TDP of the dual-core Pentium D. In theory, this means that Intel has been able to halve the TDP of its processors while also increasing their performance.

AMD Processors

In December AMD officially launched the first Athlon 64 X2 processors manufactured using a 65nm process. The models are the Athlon 64 X2 5000+, 4800+, 4400+ and 4000+. They all have 512kb L2 cache per core. The use of a smaller manufacturing process has not affected the amount of transistors on each chip. The number is still at 145 million transistors. The surface area of the die, however, has been reduced to 126 mm2 . As a result of the die shrinkage, the declared TDP has also been reduced from 89 watts to 65 watts.

Processor

 Number of Cores Clock HT Bus
L2 Cache 		Multiplier Manufacturing Process Socket TDP
Athlon 64 X2 6000+ 2 3 GHz 1 GHz 2x1 Mbyte 15x 90 nm AM2 125 Watt
Athlon 64 FX62 2 2.8 GHz 1 GHz 2x1 Mbyte 14x 90 nm AM2 125 Watt
Athlon 64 X2 5600+ 2 2.8 GHz 1 GHz 2x1 Mbyte 14x 90 nm AM2 89 Watt
Athlon 64 X2 5400+ 2 2.8 GHz 1 GHz 2x512 Kbytes 14x 90 nm AM2 89 Watt
Athlon 64 X2 5200+ 2 2.6 GHz 1 GHz 2x1 Mbyte 13x 90 nm AM2 89 Watt
Athlon 64 X2 5000+ 2 2.6 GHz 1 GHz 2x512 Kbytes 13x 90 nm AM2 89 Watt
Athlon 64 X2 5000+ 2 2.6 GHz 1 GHz 2x512 Kbytes 13x 65 nm AM2 65 Watt
Athlon 64 X2 4800+ 2 2.5 GHz 1 GHz 2x512 Kbytes 12.5x 65 nm AM2 65 Watt
Athlon 64 X2 4600+ 2 2.4 GHz 1 GHz 2x512 Kbytes 12x 90 nm AM2 89 Watt
Athlon 64 X2 4600+ EE 2 2.4 GHz 1 GHz 2x512 Kbytes 12x 90 nm AM2 65 Watt
Athlon 64 X2 4400+ 2 2.3 GHz 1 GHz 2x512 Kbytes 11.5x 65 nm AM2 65 Watt
Athlon 64 X2 4200+ 2 2.2 GHz 1 GHz 2x512 Kbytes 11x 90 nm AM2 89 Watt
Athlon 64 X2 4200+ EE 2 2.2 GHz 1 GHz 2x512 Kbytes 11x 90 nm AM2 65 Watt
Athlon 64 X2 4000+ 2 2.1 GHz 1 GHz 2x512 Kbytes 10.5x 65 nm AM2 65 Watt
Athlon 64 X2 3800+ 2 2 GHz 1 GHz 2x512 Kbytes 10x 90 nm AM2 89 Watt
Athlon 64 X2 3800+ EE 2 2 GHz 1 GHz 2x512 Kbytes 10x 90 nm AM2 65 Watt
Athlon 64 X2 3800+ EESFF 2 2 GHz 1 GHz 2x512 Kbytes 10x 90 nm AM2 35 Watt

As the table shows, AMD has also released some more energy efficient chips that are manufactured using a 90nm process. These chips are indicated by AMD as Energy Efficient (EE) and Energy Efficient Small Form Factor (EESFF). Respectively, they have a 65 watt TDP and a 35 watt TDP. Although the 65nm processors have the same declared TDP as the EE series, AMD has chose not to use the acronym for the new processors.

It is important to remember than any AMD processor with an odd divider results in the memory actually working at a lower speed than what it is rated at. The reason behind this lies in the way the memory bus clock is calculated. For AMD Athlon processors, the divider is half the CPU multiplier. Now when the divider ends up being odd, what happens is that it is rounded up. For example, the Athlon 64 X2 5200+ has a multiplier of 13. Halved, the divider is 6.5. Since the AMD64 memory bus does not work with decimal numbers, the divider is then rounded up to the next highest number, which is in this case 7. Thus, a DDR2-800 memory module would end up working at 733 MHz DDR speeds ( (DDR2-800; 2,400 MHz / 7). This creates a situation in which the lower memory speeds could potentially have a negative impact on the overall system performance.

Below is a table that shows the processors along with their dividers and memory frequencies with DDR2-800 modules:

Processor

 Clock Divider Memory Frequency
DDR2-800
Athlon 64 X2 6000+ 3 GHz 8 750 MHz
Athlon 64 FX62 2.8 GHz 7 800 MHz
Athlon 64 X2 5600+ 2.8 GHz 7 800 MHz
Athlon 64 X2 5400+ 2.8 GHz 7 800 MHz
Athlon 64 X2 5200+ 2.6 GHz 7 743 MHz
Athlon 64 X2 5000+ 2.6 GHz 7 743 MHz
Athlon 64 X2 5000+ 2.6 GHz 7 743 MHz
Athlon 64 X2 4800+ 2.5 GHz 7 714 MHz
Athlon 64 X2 4600+ 2.4 GHz 6 800 MHz
Athlon 64 X2 4600+ EE 2.4 GHz 6 800 MHz
Athlon 64 X2 4400+ 2.3 GHz 6 767 MHz
Athlon 64 X2 4200+ 2.2 GHz 6 733 MHz
Athlon 64 X2 4200+ EE 2.2 GHz 6 733 MHz
Athlon 64 X2 4000+ 2.1 GHz 6 700 MHz
Athlon 64 X2 3800+ 2 GHz 5 800 MHz
Athlon 64 X2 3800+ EE 2 GHz 5 800 MHz
Athlon 64 X2 3800+ EESFF 2 GHz 5 800 MHz

Page 3 - Athlon 64 65 nanometer: L2 Cache and Performance

To give an idea of how a move from 90nm to 65nm affects performance, we performed some benchmarks on both versions of the Athlon 64 X2 5000+. As can be seen in our results, in most cases the 65nm version of the processor actually has less performance compared to the 90nm version. This is primarily because AMD has made some modifications to the L2 cache, and so the L2 cache latency has actually increased from 12 cycles to 14 cycles. The latency of the L1 cache, on the other hand, remains the same along with the latency of the memory controller.

AMD has stated that the larger latency will allow them to increase the size of the L2 cache in future processors. This does not, however, mean that AMD will necessarily release a 65nm CPU with a larger L2 cache, just that they have the possibility of doing so.

Here are the results of our tests:

Benchmark Athlon 64 X2
5000+ 65nm		 Athlon 64 X2
5000+ 90nm		 % Variation
from 90 to 65 nm
Sysmark 2004 Rating 259 263 -1.52%
Internet Content Creation- Overall 339 343 -1.17%
Internet Content Creation- 3D Creation 312 316 -1.27%
Internet Content Creation- 2D Creation 404 408 -0.98%
Internet Content Creation- Web Publication 308 313 -1.60%
Office Productivity- Overall 194 202 -3.96%
Office Productivity- Communication 184 199 -7.54%
Office Productivity- Document Creation 238 242 -1.65%
Office Productivity- Data Analysis 167 172 -2.91%
Splinter Cell Chaos Theory - aa4x anis 16x - 1024x768 123.5 124,1 -0.48%
Splinter Cell Chaos Theory - aa4x anis 16x - 1280x1024 97 97 0.00%
Splinter Cell Chaos Theory - aa4x anis 16x - 1600x1200 76.6 76,6 0.00%
3D Mark 2006 - 1280x1024 6119 6131 -0.20%
3D Mark 2006 - 1600x1200 5369 5408 -0.72%
3D Mark 2006 - CPU test 1971 1980 -0.45%
Fear - 1280x1024 aa4x anis 16x - min 36 37 -2.70%
Fear - 1280x1024 aa4x anis 16x - average 74 74 0.00%
Fear - 1600x1200 aa4x anis 16x - min 27 27 0.00%
Fear - 1600x1200 aa4x anis 16x - average 56 56 0.00%
Far Cry - Training - aa4x anis 16x - 1024x768 106.9 116.6 -8.32%
Far Cry - Training - aa4x anis 16x - 1280x1024 102.7 109.7 -6.38%
Far Cry - Training - aa4x anis 16x - 1600x1200 95,8 99.3 -3.52%
Prey - aa4x anis 16x - 1024x768 111.1 111.5 -0.36%
Prey - aa4x anis 16x - 1280x1024 82.7 82.8 -0.12%
Prey - aa4x anis 16x - 1600x1200 64.7 64.8 -0.15%
AutoGordianKnot - Divx 6.11 - conversione traccia 9 Kill Bill volume 1 67.2 68.1 -1.32%
Itunes 7.0.1.8 - Traccia audio da 613 Mbytes 92 92 0.00%
Lame MT - traccia audio da 613 Mbytes - Constant Bit Rate - Intel Compiler - singletask 177 177 0.00%
Lame MT - traccia audio da 613 Mbytes - Constant Bit Rate - Intel Compiler - singletask 177 177 0.00%
Lame MT - traccia audio da 613 Mbytes - Constant Bit Rate - Intel Compiler - multitask 129 127 -1.55%
Mainconcept MPEG Encoder - Conversione da DV in MPEG2 bitrate variabile 327 324 -0.92%
Mainconcept H.264 Encoder - Conversione da HD 1080i a H.264 High 135.6 134,5 -0.81%
Windows Media Encoder 9 with advanced profiles - Video ATI factory Tour; 6 min 57 sec 240 240 0.00%
7-Zip 4.42 benchmark integrato (MIPS) 32 Mbytes - compressione 3119 3266 -4.50%
7-Zip 4.42 benchmark integrato (MIPS) 32 Mbytes - decompressione 2500 2525 -0.99%
7-Zip 4.42 benchmark integrato (MIPS) 32 Mbytes - totale 2810 2896 -2.97%
DVD Shrink 3.2 - Fahrenheit 9/11 - compresso a 2000 Mbytes 276 265 -3.99%
Cinebench 9.5 - CPU Benchmark (CB-SEC) 384 386 -0.52%
Cinebench 9.5 - CPU Benchmark Multiple (CB-SEC) 716 724 -1.10%
Cinebench 9.5 - Povray 3.7 - benchmark 281 281 0.00%
Viewperf 9.03 - 3dsmax-04 10.53 10.68 -1.40%
Viewperf 9.03 - catia-02 10.76 11.26 -4.44%
Viewperf 9.03 - ensight-03 10.38 9.88 5.06%
Viewperf 9.03 - light-08 10.93 11.4 -4.12%
Viewperf 9.03 - maya-02 12.03 12.1 -0.58%
Viewperf 9.03 - proe-04 6.008 6.243 -3.76%
Viewperf 9.03 - sw-01 10.88 11.23 -3.12%
Viewperf 9.03 - tcvis-01 2.531 2.58 -1.90%
Viewperf 9.03 - ugnx-01 9.401 9.4 0.01%

Page 4 - Test Configuration

For the analysis of the AMD Athlon 64 6000+, we used two main test platforms. For Intel LGA 775 processors we used the Intel D975XBX2KR while AMD Socket AM2 processors, we instead chose to use the Asus M2R32-MVP motherboard.

Processors used in the test:

  • AMD Athlon 64 X2 6000+AM2 90 nm ( 3,000 MHz clock - cache L2 2x1 Mbytes) - Dual DDR2-800 memory
  • AMD Athlon 64 FX62 Socket AM2 90 nm ( 2,800 MHz clock - cache L2 2x1 Mbytes) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 5600+AM2 90 nm ( 2,800 MHz clock - L2 2x1 Mbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 5200+AM2 90 nm ( 2,600 MHz clock - L2 2x1 Mbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 5000+AM2 90 nm ( 2,600 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 5000+AM2 65 nm ( 2,600 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 4800+AM2 65 nm ( 2,500 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 4600+AM2 90 nm ( 2,400 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 4400+AM2 65 nm ( 2,300 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 4200+AM2 90 nm ( 2,200 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 4000+AM2 65 nm ( 2,100 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • AMD Athlon 64 X2 3800+AM2 90 nm ( 2,000 MHz clock - L2 2x512 Kbytes cache) - Dual DDR2-800 memory
  • Intel Core 2 Extreme QX6700 ( 1066 MHz bus, 2.66 GHz - 2x4 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Extreme X6800 ( 1066 MHz bus, 2.93 GHz - L2 4 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Duo E6700 ( 1066 MHz bus, 2.66 GHz clock frequency- 4 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Duo E6600 ( 1066 MHz bus, 2.4 GHz clock frequency- 4 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Duo E6400 ( 1066 MHz bus, 2.13 GHz clock frequency- 2 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Duo E6300 ( 1066 MHz bus, 1.83 GHz clock frequency- 2 Mbytes L2 cache) - Dual DDR2-800 memory
  • Intel Core 2 Duo E4300 ( 800 MHz bus, 1.8 GHz clock frequency- 2 Mbytes L2 cache) - Dual DDR2-800 memory

Testbed Components:

Intel 975X Platform- Intel Core 2 Duo and Core 2 Extreme Processors

  • motherboard: Intel D975XBX2KR ( Intel 975X chipset)
  • memory: Corsair CM2X1024 6400 (5-5-5-15) @ 800 MHz; 2x1 Gbyte
  • hard disk: Western Digital WD1600JS - Serial ATA - 7.200 rpm, 160 Gbytes
  • videocard : ATI Radeon X1950XTX (gpu 650 MHz; memoria video 2.000 MHz)
  • operating system : Windows XP Professional, Service Pack 2
  • video driver: ATI Catalyst 6.10

Athlon 64 X2 and Athlon 64 FX processor platforms:

  • motherboard: Asus M2R32-MVP (chipset AMD580)
  • memory: Corsair CM2X1024 6400 (5-5-5-15) @ 800 MHz; 2x1 Gbyte
  • hard disk: Western Digital WD1600JS - Serial ATA - 7.200 rpm, 160 Gbytes
  • videocard: ATI Radeon X1950XTX (gpu 650 MHz; memoria video 2.000 MHz)
  • operating system: Windows XP Professional, Service Pack 2
  • video driver : ATI Catalyst 6.10

Benchmarks:

Benchmark Applicatoins

Sysmark 2004 SE

Gaming Tests

3D Mark 2006
CPU Test
1280x1024
1600x1200

Splinter Cell Chaos Theory - aa4x anis 16x
1024x768
1280x1024
1600x1200

Fear -aa4x anis 16x
1280x1024
1600x1200

Far Cry - Training - aa4x anis 16x
1024x768
1280x1024
1600x1200

Prey demo - aa4x anis 16x
1024x768
1280x1024
1600x1200

Multimedia

Auto GK 2.27 - Divx - conversione traccia 9 "Kill Bill volume 1"; audio inglese 6ch
    Divx 6.11
    scena 6
    Qualità 75% - no audio

Lame MT
    Traccia audio da 613 Mbytes
    Constant Bit Rate - Intel Compiler

Itunes 7.0.1.8
    conversione traccia audio da 613 Mbytes in formato MP3 192 Kbps

Mainconcept MPEG Encoder
    Conversione da DV a MPEG2 bitrate variabile
    Video ECS factory tour - DV 16.836 frames
    720x576

Mainconcept H.264 Encoder
    Conversione da HD 1080i a H.264 High
    Video HD 1080i da 24 secondi
    hdwatermellon

Windows Media Encoder 9 with advanced profiles
    Video ATI factory Tour; 6 min 57 sec
    320x240 - 282 Kbps

7-Zip 4.42
    benchmark integrato, 32 Mbytes

DVD Shrink 3.2
    Fahrenheit 9/11
    compresso a 2000 Mbytes

Rendering

Cinebench 9.5
    CPU Benchmark (CB-SEC)
    CPU Benchmark Multiple (CB-SEC)Cinebench 2003

Povray 3.7
    benchmark

Open GL professional

Viewperf 9.03

Page 5 - Sysmark 2004 SE

sysmark_1.png (65695 bytes)

sysmark_2.png (67480 bytes)

sysmark_3.png (65375 bytes)

Sysmark 2004 SE is a productivity benchmark which allows us to simulate various real world environments. We started with a benchmark of overall performance, and following that performed specific environment tests.

For the most part the new AMD Athlon 64 X2 6000+ is able to perform better than all other AMD processors; however, it falls short of Intel's top of the range offerings. In the office productivity test the Athlon 64 X2 6000+ scores less than the Athlon 64 X2 5600+ and Athlon 64 FX-62. This happens because of the lower memory frequency of the Athlon 6000+, which is 750 MHz, 50 MHz less than the other two processors.

We ran the following applications in these series of tests:

  • Adobe Acrobat 5.0.5
  • Microsoft Access 2002
  • Microsoft Excel 2002
  • Microsoft Outlook 2002
  • Microsoft PowerPoint 2002
  • Microsoft Word 2002
  • Microsoft Internet Explorer 6.0 SP1
  • ScanSoft Dragon NaturallySpeaking 6
  • Network Associates McAfee VirusScan 7.0
  • WinZip Computing WinZip 8.1
  • Adobe After Effects 5.5
  • Adobe Photoshop 7.01
  • Adobe Premiere 6.5
  • Discreet 3ds max 5.1
  • Macromedia Dreamweaver MX
  • Macromedia Flash MX
  • Microsoft Windows Media Encoder 9
Page 6 - Gaming

The gaming tests were run using a single ATI Radeon X1950XTX video card with 4x anti-aliasing and 16x anisotropic filtering enabled. These are the same settings we usually benchmark our video cards with. By disabling both anisotropic filtering along with anti-aliasing we could have created a situation more dependant on processing power, however, we wanted to create a more realistic situation.

To test processor gaming performance, we usually run our gaming tests at 640x480 or 800x600, therefore increasing the impact the processor has on performance. However, considering the power of many of today's video cards, including lower end cards, we decided to run the F.E.A.R. tests at 1280x1024 and 1600x1200.

3dmark_2006_1.png (76865 bytes)

3dmark.gif (2700 bytes)

3dmark_2006_2.png (66437 bytes)

3D Mark 2006 has a benchmark built specifically for testing processors. For this reason the Core 2 Extreme QX6700 is able to perform the fastest in the tests. In a real world situation, as we discovered in our review of the QX6700, most games don't take advantage of the processor's four cores.

fear_1.png (77066 bytes)

fear_2.png (76596 bytes)

In F.E.A.R. performance across all processors is relatively the same, the only major exception is the Core 2 Duo E4300 and Core 2 Duo E6300, which end up with least amount of performance. We believe that the video game has some limits with these processors. We are pretty sure, although we cannot guarantee, that this is a result of the clock frequency of these processors being below the 2 GHz line.

On the whole, the AMD Athlon 64 X2 6000+ does not perform differently from other processors in F.E.A.R. Its performance stays relatively the same and remains on par with competing processors.

Page 7 - Gaming-2

far_cry.png (117338 bytes)

prey.png (86777 bytes)

sc.png (89919 bytes)

The scene does not change passing to other game titles, and the Core 2 Duo chips hold the best absolute performance in every situation. Increasing the resolution with AA and AF still enabled, the margin of difference between several processors increases. This is most evident at 1600x1200; in this case, though, the main bottle neck on the frame rate is not because of the processor but the video card itself.

The Athlon 64 X2 6000+ is the highest performing socket AM2 chip. The performance difference between the AMD processors, however, is mostly very slight, especially in the last two tests, and can be considered negligible.

Page 8 - Multimedia - 1

autogk.gif (2780 bytes)

autogk.png (67017 bytes)

Auto GK allows us to convert movies to Divx format, a task that is quite common for many enthusiasts. It is also a key test for us because it requires a large amount of CPU attention. The Athlon 64 X2 6000+ takes advantage of the 200 MHz it has over the Athlon X2 5600+, and succeeds in pulling ahead of the Intel Core 2 Duo E6400.

7zip.gif (3286 bytes)

7zip.png (90347 bytes)

Although 7-Zip can take advantage of dual-cores, it does not benefit from the additional cores the QX6700 has to offer. 7-Zip performs faster with Intel Core 2 Duo solutions; however, the margin of difference over Athlon 64 X2 processors is smaller than it was for the Auto GK test.

lame.gif (2393 bytes)

lame_mt.png (67379 bytes)

Lame MT is a tool that converts from Wav to MP3. Although the utility has a command line interface, it can still take advantage of multiple cores, but it still doesn't benefit from the added horsepower of the QX6700. The tests were run at a constant bit rate using the Intel compiler. Here, the Athlon 64 X2 6000+ performs on par with the Core 2 Duo E6400.

itunes.gif (2609 bytes)

itunes.png (67044 bytes)

In iTunes we converted a 613 MB audio file from WAV to MP3 192Kbps. Similar to the Lame MT test, the new AMD processor ends up with conversion times equal to that of the Core 2 Duo E6400.

Page 9 - Multimedia - 2

mainconcept.gif (2665 bytes)

mainconcept_1.png (69622 bytes)

mainconcept_h264.gif (2573 bytes)

mainconcept_2.png (67979 bytes)

The Mainconcept test results indicate that the program actually takes advantage of multiple cores, as the Intel Core 2 Extreme QX6700 performs the conversions in the least amount of time. The Athlon 64 X2 6000+ also has quite fast conversion times and ends up with performance similar to the Core 2 Duo E6600.

wme9.gif (2696 bytes)

wm9.png (69680 bytes)

The scene doesn't change in the DV to WMV conversion tests, and the Athlon 64 X2 6000+ has conversion times identical to the Core 2 Duo E6600.

dvd.gif (3067 bytes)

dvd.png (94655 bytes)

Our final multimedia application performance analysis ends with backing up a DVD film to a hard drive, setting DVD Shrink to compress the DVD to 2000MB. For the most part Core 2 Duo processors have the fastest compression times, while the Athlon 64 X2 6000+ is 8 seconds faster than the Core 2 Duo E6300. The Athlon 64 FX-62 is two seconds faster than the X2 6000+, primarily because it has a faster memory speed.

Page 10 - Rendering - Open GL

povray.gif (3022 bytes)

povray.png (66353 bytes)

cinebench.gif (2628 bytes)

cinebench.png (103968 bytes)

POV-Ray and Cinebench 9.5 both take advantage of the four cores the Core 2 Extreme QX6700 has, as is clearly shown in the POV-Ray benchmarks and the multi-threading tests of Cinebench. The new Athlon 64 X2 6000+ has performance very similar to the Core 2 Duo E6400, while in the Cinebench multi-threading tests it is faster than all Core 2 Duo processors except for the Core 2 Extreme X6800 and Core 2 Extreme QX6700.

Viewperf 9.03 is a synthetic benchmark developed by Spec.org that tests system performance in specific professional applications that use OpenGL. It is comprised of 9 viewsets, each one for a different application. It is important to remember that since this a graphical benchmark it also takes the video card used during the testing (ATI Radeon X1950XTX) into account.

vp1.png (63946 bytes)

vp2.png (63946 bytes)

vp3.png (63946 bytes)

vp4.png (63946 bytes)

vp5.png (63946 bytes)

vp6.png (63946 bytes)

vp7.png (63946 bytes)

vp9.png (63946 bytes)

The Viewperf 9.03 tests show the Core 2 Duo processors as the fastest performers. Viewperf is a benchmark that is very sensitive to the motherboard's chipset. The application's sensitivity can clearly be seen in the SW-01 viewset in which all AMD processors suffer in terms of performance, while the Intel Core 2 Duo line of processors perform quite well.

The Athlon 64 X2 6000+ performs better than all other AMD processors in the Viewperf tests, confirming that its faster operating frequencies gives it an advantage over other AMD processors.

Page 11 - Consumption

Intel has placed a large amount of emphasis on the Intel Core 2 Duo line of processors' low power consumption levels. As Intel explains in their introduction to the Core 2 Duo architecture, they believe the processors offer high performance at low levels of power consumption.

The following table compares the declared TDP values of Intel processors against AMD processors:

Processore Clock TDP Socket
Athlon 64 6000+ 3 GHz 125 Watt AM2
Athlon 64 FX62 2,8 GHz 125 Watt AM2
Athlon 64 X2 5000+ 65nm 2,6 GHz 65 Watt AM2
Athlon 64 X2 5000+ 2,6 GHz 89 Watt AM2
Athlon 64 X2 4600+ 2,4 GHz 89 Watt AM2
Athlon 64 X2 4200+ 2,2 GHz 89 Watt AM2
Athlon 64 X2 3800+ 2 GHz 89 Watt AM2

Core 2 Extreme QX6700

 2,67 GHz 130 Watt 775 LGA

Core 2 Extreme X6800

 2,93 GHz 75 Watt 775 LGA
Core 2 Duo E6700 2,67 GHz 65 Watt 775 LGA
Core 2 Duo E6600 2,4 GHz 65 Watt 775 LGA
Core 2 Duo E6400 2,13 GHz 65 Watt 775 LGA
Core 2 Duo E6300 1,86 GHz 65 Watt 775 LGA
Core 2 Duo E4300 1,8 GHz 65 Watt 775 LGA

Intel's declared TDP (Thermal Design Power) values are not the maximum consumption levels of the processors. Instead, they are the typical TDP the processors reach in what Intel believes is a typical situation. AMD's figures, on the other hand, are the maximum consumption the processors can have in any situation. This is something very important to keep in mind when reading TDP values of Intel and AMD processors.

Moreover, the TDP value supplied by AMD also includes the integrated memory controller. For Intel processors the memory controllers aren't integrated in the CPU. Instead, they are external and located on the north bridge. For this reason we have tested the entire power draw of the system, using the same parts where possible:

  • motherboard: Asus M2R32-MVP ( ATI Crossfire Xpress 3200 chipset)
  • motherboard: Intel D975XBX2KR ( Intel 975X chipset)
  • memory: Corsair CM2X1024 6400 (5-5-5-15) @ 800 MHz; 2x1 Gbyte
  • hard disk: Western Digital WD1600JS - Serial ATA - 7200 rpm, 160 Gbytes
  • video card : ATI Radeon X1950XTX ( 650 MHz GPU; 2,000 MHz video memory)
  • operating system : Windows XP Professional, Service Pack 2
  • video driver: ATI Catalyst 6.10

We tested at two power states; idle and full load. Idle consisted of an empty desktop loaded in Windows XP Service Pack 2, and "Full Load" consisted of all processors at 100% usage by running POV-Ray 3.7. In addition, all processors had power saving features enabled.

consumo_1.png (63679 bytes)

At idle, AMD's Cool'n'Quiet kicked in, which is why we see nearly identical power consumption for all AMD processors. Intel processors are nearly identical to each other, except for minor variations caused by the different L2 caches of higher-end processors. AMD's Cool'n'Quiet helped keep power consumption low for AMD processors by reducing operating frequencies to 1 GHz when the system was idle.

consumo_2.png (61970 bytes)

Under 100% the scene shifts to favor Intel Core 2 Duo processors. The Core 2 Extreme QX6700 consumes the most power, obviously because the presence of four cores increases its power needs. The 65nm Athlon 54 X2 5000+ consumes the least power out of all AMD processors, while the Athlon 64 X2 6000+ consumes the most.

Page 12 - Athlon 64 X2 6000+ Overclocked

Two factors make overclocking headroom for the new AMD CPU limited. For one, the processor is the highest clocked Athlon 64 X2 processor in commerce, meaning that AMD has already pushed the processor to its limit. In addition, the fact that the Athlon 64 X2 6000+ is manufactured using a 90nm makes it even tougher to have high overclocking hopes for the processor. Please also keep in mind that overclocking capabilities usually vary chip to chip. Although we might be able to reach high overclocks with some of our chips, there is no guarantee that another would perform exactly the same.

6000+_oc_cpuz.gif (16935 bytes)

6000+_oc_screen_s.jpg (43564 bytes)

The maximum frequency our engineering sample reached in the overclocking tests was 3.255 GHz. We reached this frequency by increasing the bus frequency to 217 MHz and keeping the multiplier at its default 15x. To maintain system stability at this level, the operation required a very slight overvolt. As our tests confirm, the overclocking headroom on Athlon 64 6000+ processors is rather limited, primarily because of the 90nm manufacturing process.

Page 13 - Conclusion

A clock increase of 200 MHz cannot radically change the performance of the Athlon 64 X2 series, and the debut of the 6000+ confirms this. Although the new processor is able to increase the performance levels of AMD Athlon processors, it doesn’t change performance levels to such an extent that they will completely counter Intel’s high-end Core 2 Duo processors. Logically, though, this should be expected as AMD is currently a processor generation behind.

Again AMD has released its newest processor at a lower price point than expected. Although it was originally expected to debut in the $600 range, AMD lowered the figure right before the official February 20 launch. As such, the new Athlon 64 X2 6000+ comes at 464 USD, a price that is cheaper than the Core 2 Duo E6700.

prezzi.png (93929 bytes)

Our Moneybench results also show the Athlon 64 X2 6000+ have a price to performance ratio very similar to the Core 2 Duo E6700.

moneybench.png (96600 bytes)

For the next six months the desktop processor market will remain nearly completely stagnant, save for a few minor additoins. Because of this, we can expect to see AMD further lower the price of its processors in the coming months. Only with the release of the K10 architecture can we really expect to see AMD start chugging out some real solid competition against Intel. The decision of buying the Athlon 64 X2 6000+ therefore lies solely on the consumer. AMD’s aggressive pricing is keeping the company in the game, and therefore you see many low priced socket AM2 processors.
Printed from Hardware Upgrade: http://www.hwupgrade.com
Copyright 1997 - 2010 Hardware Upgrade S.r.l