iXBT Labs - Computer Hardware in Detail

Platform

Video

Multimedia

Mobile

Other

Processors for Socket AM2: AMD Catches Up in Technology and Shoots Out in Performance

May 23, 2006



New platform in brief

The new platform from AMD is actually a well-known old one with the only difference - a new memory controller in the processor: it now supports DDR2 (from DDR2-400 to DDR2-800). As you can see on the photo, it has absolutely no effect on the CPU exterior (reverting to 940 pins, like in the first Athlon 64 FX models for the server socket, is hardly a major modification).

  

The first motherboard to arrive at our lab was MSI K9N SLI Platinum based on the latest NVIDIA nForce 570 SLI chipset. It looks quite usual, but for DDR2-800 memory used together with a typically AMD socket :).



The socket itself has a slightly modified CPU retention mechanism — there disappeared side stops, which did not let the heatsink to shift sideways, and the number of cooler hooks is reduced from three to one (per each side). Thus, if we do not take into account that the hooks are attached not to the socket, but to a special binding screwed to the motherboard, the new retention mechanism resembles... Socket A or Socket 370. I don't know what inspired AMD engineers to "Back to the future". However, coolers for Socket 754/939/940 can be installed on the new socket (even if with some effort) and work well. We have also seen for the first time a connector for a 4-pin PWM cooler on a motherboard for AMD. To all appearances, AMD decided to "catch up with Intel in all respects" in the new platform :).



There is nothing sensational in other respects. Just a usual Athlon 64 processor, but with a new socket and support for another memory type. At the same time, the qualitative advance is quite noticeable: two DDR2-800 channels versus two DDR-400 channels — that's (in pure theory) a twofold advantage in bandwidth! Well, we'll see how our tests will react to it...

Hardware and Software

Testbed configurations

CPU
Motherboard
Memory
Video
Athlon 64 FX-62
MSI K9N SLI Platinum
Corsair CM2X1024-6400
GeForce 7800GTX 256 MB
Athlon 64 X2 4000+
MSI K9N SLI Platinum
Corsair CM2X1024-6400
GeForce 7800GTX 256 MB
Athlon 64 FX-60
EPoX EP-9NPA3
Corsair CMX1024-3500LLPRO
GeForce 7800GTX 256 MB
  • Memory — 2 GB (2 modules)
  • HDD: Samsung SP1614C (SATA)
  • Coolers: standard models that come shipped with processors
  • PSU: Thermaltake PurePower 680 APD
Processor
Athlon 64 FX-62
Athlon 64 X2 4000+
Athlon 64 FX-60
Process Technology
90 nm
90 nm
90 nm
Core Clock, GHz
2.8
2.0
2.6
Number of Cores
2
2
2
L2 Cache*, KB
2x1024
2x1024
2x1024
FSB clock**, MHz
400 DDR2
400 DDR2
200 DDR
Multiplier
14
10
13
Socket
Socket AM2
Socket AM2
Socket 939
Typical thermal emission***
125 W
89 W
110 W
AMD64/EM64T
+
+
+
Hyper-Threading
-
-
-
Virtualization Technology
-
-
-

* — "2x..." means "per each core"
** — in AMD processors it's frequency of the memory controller bus
*** — it's measured differently in Intel and AMD processors, so a direct comparison is not correct

As Athlon 64 X2 4000+ has no direct opponent (Athlon 64 X2 3800+ has half as much cache memory and Athlon 64 X2 4400+ operates at a higher clock), we decided to decrease the multiplier to ten in a regular Athlon 64 FX-60. Thus, we get exactly the same dual-core processor as Athlon 64 X2 4000+ for Socket AM2, only for Socket 939. In our opinion, it will be the best posed comparison from the point of view of evaluating advantages of the new platform.

Besides, you might have already noticed that the list of test participants does not include Intel processors. It seems logical, as you may read, for example, Testing Athlon 64 FX-60 to learn performance results of various Intel NetBurst-based CPU families and AMD flagships. This comparison should be enough. It would be really interesting to compare the new platform from AMD with Intel "Conroe". But alas, we still haven't got it in our lab.

Software

  1. Windows XP Professional x64 Edition SP1.
  2. 3ds max 7.0
  3. Maya 6.5
  4. Lightwave 8.5 x64 Edition
  5. WinRAR 3.51
  6. 7-Zip 4.32 x64 Edition
  7. LAME 3.98
  8. Monkey Audio 4.01
  9. OGG Encoder 2.8 (Lancer)
  10. Windows Media Encoder 9 x64 Edition
  11. MATLAB 7.1
  12. Pro/ENGINEER Wildfire 2.0
  13. SolidWorks 2005
  14. Microsoft Visual C++ Professional 6.0
  15. CPU RightMark 2005 Lite x64 Edition
  16. F.E.A.R. 1.3
  17. Half-Life 2
  18. Unreal Tournament 2004 build 3339
  19. Quake 4 Point Release 1.1
  20. FineReader Professional 8.0
  21. Adobe Photoshop CS2 (9.0)
  22. Canopus ProCoder 2.01.30
  23. DivX 6.1.1
  24. Windows Media Video 9 VCM
  25. x264 v.438
  26. XviD 1.1.0 Release
  27. Apache 2.0.55 for Windows

Drivers

  1. NVIDIA ForceWare 81.98
  2. NVIDIA nForce SMBus Driver 4.50

Necessary preface to the diagrams

Our test procedure features two peculiarities of data representation: (1) all data types are reduced to one — integer relative score (performance of a given processor relative to Pentium D 805, if its performance is taken for 100 points), and (2) detailed results are published in a table in Microsoft Excel format, while the article contains only summary diagrams for benchmark classes.

3D Modeling and Rendering



The advantage of Athlon 64 X2 4000+ over its equally-clocked opponent for Socket 939 is not impressive, to say the least. Far from the 7% gains, mentioned in previews, you can see it with the naked eye. FX-62 results are inspiring, but nothing of the extraordinary.

CAD (computer-aided design)



This situation is similar to 3D modeling programs: the advantage of the new platform over the old one is practically unnoticeable in the bottom pair of processors and more pronounced on top of the list.

Compile



OK, let's try to calm down - compilation is not quite sensitive to memory bandwidth... But... What's that?! The first diagram, where Athlon 64 FX-62 shoots forward from the former performance leader — Athlon 64 FX-60. You cannot explain it only by the 200 MHz increase in frequency. So here is the first example of the advantage of the new platform.

RightMark Project



CPU RightMark has never been very sensitive to memory bandwidth, but it also demonstrates a contrast between the top and bottom pairs. The new platform practically didn't show itself in case of the bottom pair. The advantage of FX-62 over FX-60 in the top pair again cannot be explained by a frequency advantage only.

Processing Bitmap Images (Photos)



The same situation here: only FX-62 results on the diagram let us hope that dual-channel DDR2-800 can indeed breath a new life into the aged K8 core...

WEB server



So, the number of coincidences is too high to just be amazed at them. It's time to call in the analysis. Why does the new platform demonstrate excellent results in combination with Athlon 64 FX-62 and look so pale versus the previous platform in case of Athlon 64 X2 4000+? There may be two answers.

Answer One: the point is in CPU clocks. Huge memory bandwidth, provided by the dual-channel DDR2-800 controller for the Athlon 64 X2 4000+ processor is just... wasted. It cannot operate fast enough to load such high-speed memory. That's why we can see practically no difference between Athlon 64 FX-60 @ 2000 MHz + DDR-400 and Athlon 64 X2 4000+ + DDR2-800. Athlon 64 FX-62, which clock is higher by 800 MHz, is quite another matter. In this case, DDR-400 (for FX-60) becomes a bottleneck, it does not allow a processor to reveal its full performance potential. But the FX-62 with dual-channel DDR2-800 is not limited in this way — that's why it outperforms FX-60 so much that it cannot be explained by the clock advantage only.

Answer Two: a system with Athlon 64 X2 4000+ actually worked with DDR2-400 memory (533, 667...) instead of DDR2-800. As Dmitri Besedin already wrote in his article, devoted to a detailed analysis of the new Athlon 64 memory controller, there is no accurate way to find out at what frequency memory actually operates. We can only trust the value specified in BIOS Setup.

File Packing



And here the advantage of X2 4000+ has grown a little, FX-62 demonstrates lower results. I think in the first case it happens because archivers are critical to memory bandwidth. In the second case — because they are critical not only to memory bandwidth, but also to latency. (Of course, it's a tad better in DDR2-800 based systems, but not as better as their memory bandwidth advantage over DDR-400).

Audio Encoding



Amazingly, even audio encoders, unpretentious as they are to memory bandwidth, still welcome Athlon 64 FX-62 and the new platform from AMD.

Video Encoding



The situation in video encoding is redoubled compared to audio encoding. It's not surprising, considering the increased volumes of data to be processed.

OCR



Quite a standard behavior for this test.

3D Shooter Games



Here we can see that some programs, critical to memory bandwidth, can benefit from the new AMD platform, including relatively low-clocked processors. And the relatively low performance gain in FX-62 versus FX-60 can be explained by the fact that games depend more on a video card. Moreover, the faster a processor, the more noticeable this dependence, especially in high resolutions.

Total score







Conclusion

I repeat: nothing extraordinary has happened. Thank God :). AMD currently takes the "little and often fill the purse" stand: sometimes it slightly increases the core clock, sometimes it introduces SSE3 support or merges two cores into a single processor, or masters DDR2 (800 at that). Why should it bustle? Frankly speaking, it's the main competitor who must be bustling now. AMD is doing great anyway. At least in terms of CPU performance. Durability of the K8 core and its capacity to adapt to new market realia is admirable: having lived without major modifications through two process technologies, dual cores, and now a new memory controller, this core meticulously responds to each improvement with performance gains. We were very skeptic about future chances of the new AMD platform against the new processor core from Intel (Intel designed the new core nearly from scratch, while AMD K8 is rather old), but our tests warmed up our interest. The situation may turn out not that simple. At least, Intel Conroe will have a worthy opponent, especially as it also has some drawbacks to fix: according to our previous review, the DDR2 controller in new AMD processors is not perfect yet.



Memory modules for our testbeds are kindly provided by
Russian representatives of Corsair Memory

Stanislav Garmatiuk (nawhi@ixbt.com)
May 23, 2006

Write a comment below. No registration needed!


Article navigation:



blog comments powered by Disqus

  Most Popular Reviews More    RSS  

AMD Phenom II X4 955, Phenom II X4 960T, Phenom II X6 1075T, and Intel Pentium G2120, Core i3-3220, Core i5-3330 Processors

Comparing old, cheap solutions from AMD with new, budget offerings from Intel.
February 1, 2013 · Processor Roundups

Inno3D GeForce GTX 670 iChill, Inno3D GeForce GTX 660 Ti Graphics Cards

A couple of mid-range adapters with original cooling systems.
January 30, 2013 · Video cards: NVIDIA GPUs

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
September 9, 2008 · Sound Cards

AMD FX-8350 Processor

The first worthwhile Piledriver CPU.
September 11, 2012 · Processors: AMD

Consumed Power, Energy Consumption: Ivy Bridge vs. Sandy Bridge

Trying out the new method.
September 18, 2012 · Processors: Intel
  Latest Reviews More    RSS  

i3DSpeed, September 2013

Retested all graphics cards with the new drivers.
Oct 18, 2013 · 3Digests

i3DSpeed, August 2013

Added new benchmarks: BioShock Infinite and Metro: Last Light.
Sep 06, 2013 · 3Digests

i3DSpeed, July 2013

Added the test results of NVIDIA GeForce GTX 760 and AMD Radeon HD 7730.
Aug 05, 2013 · 3Digests

Gainward GeForce GTX 650 Ti BOOST 2GB Golden Sample Graphics Card

An excellent hybrid of GeForce GTX 650 Ti and GeForce GTX 660.
Jun 24, 2013 · Video cards: NVIDIA GPUs

i3DSpeed, May 2013

Added the test results of NVIDIA GeForce GTX 770/780.
Jun 03, 2013 · 3Digests
  Latest News More    RSS  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ||  About us & Privacy policy  ·  Twitter  ·  Facebook


[an error occurred while processing this directive]

Copyright © Byrds Research & Publishing, Ltd., 1997–2011. All rights reserved.