Digit-Life Hardware News
17.05.2008
[01:09] Razer Launches Goliathus Speed and Control Edition Soft Mats
[00:53] Plextor Shrinks New MediaX PC-to-TV Media Player
[00:42] Daily Mailbox
[00:30] Transcend Unveils New V90P USB Flash Drive
[00:23] Microsoft Windows Now Available on One Laptop per Child XO Laptops
[00:18] IBM Research Unveils Breakthrough In Solar Farm Technology
[00:10] Creative Unleashes Sound Blaster X-Fi Titanium Fatal1ty Professional and Champion Cards
15.05.2008
[16:16] Sharp Achieves the Highest Power Density for Direct Methanol Fuel Cells (DMFC)
[16:10] Elpida Offers First DDR2 SDRAM With x32-bit I/O Configuration
14.05.2008
[21:13] Daily Mailbox
Your link here
Home Home
Latest News | Platform | Coolers | HDD/DVD | Video | Sound | Network | Imaging | Mobile
Monthly | Rightmark Tools | Search | Forum | Mailing | Links | Advertise | About Us 		Digit-Life Articles Feed    Digit-Life News Feed

Latest Articles:

Gigabyte MA78GM-S2H Motherboard on AMD 780G Chipset (Socket AM2+)

NVIDIA GeForce 8800 GT Performance Scaling in Modern Games

Sapphire PURE CrossFireX 790FX Motherboard on AMD 790FX Chipset (Socket AM2+)

Passively Cooled Gigabyte GeForce 9600 GT 512MB

i3DSpeed, April 2008




Motherboards and Memory Modules Compatibility (Socket AM2)

Reviews of motherboards traditionally pay much attention to functionality and performance. They sometimes include overclocking test results. And only when implementations of interfaces have certain problems, we plug devices with known problems and run additional tests. It's quite explainable. Even the biggest web sites cannot test each motherboard for months (their load is always pressing). Besides, our readers will hardly be pleased to get reviews with such a delay. Forum posts supplement the picture, where real owners of such motherboards share their impressions. However, this approach also has its drawbacks. Having read a forum topic almost about any motherboard, your first impression will most certainly be that it's a bad model, because many people complain about it. We recommend that you should introduce some reliability threshold. For example, if a single person complains that his motherboard does not work with Samsung memory modules, it does not mean anything (the problem may be in memory modules, perhaps both modules are installed in the same channel, or BIOS version is old, and there is an updated version offering improved compatibility - there may be a lot of options). But if a given problem is reported by many users, it most likely exists for real.

However, from the point of view of a reviewer, it's not fair to send readers to forums for details. And it's absolutely not an option, if you choose a motherboard for several computers with definite hardware requirements. In this case additional tests should be run in equal conditions with all possible contenders according to the same test procedure. We now have an opportunity to publish results of such tests, carried out by RTLab, an independent testlab. This shootout includes relatively inexpensive Socket AM2 motherboards and seven memory kits. We'd like to point out that this analysis will not draw glovbal conclusions, like any private research. But in out opinion, it will be interesting to a wide audience (not only to the customer, who ordered this reaserch). Why choose DDR2-667? When we speak of computers on inexpensive motherboards (and configurations not intended to break any records in performance tests), but with high requirements to reliability, it makes sense to choose components with slightly better specifications than you really need.

Test procedure

We ran the following programs and tests to evaluate stability.

  • RightMark Memory Analyzer 3.65 (all tests) + Memory Stability Test (1 hour)
  • CPU RightMark 2004B (1 hour)
  • A lengthy demo in Doom 3, FarCry, and F.E.A.R. in 1024x768 and 1600x1200 modes with maximum graphics quality (2 hours)
  • S&M 1.8.2 b in the memory test mode, all tests, long + 100% load (4 hours)

RightMark tests can be downloaded from this web site.

The testbed was assembled on motherboards under review and the following components:

All the tests were carried out with standard frequencies and voltages, timings were detected automatically by motherboards, so they might not agree with SPD data and values recommended by module manufacturers. Automatic detection of timings (instead of manual values) was used because correct operation of this procedure is important for system integrators. Users often don't know how to specify correct timings. If CMOS is cleared and timings are not detected correctly, users may end with an unstable system, and they will probably address their tech support.

Results

Test results are published in the following table.

Motherboard Chipset BIOS Memory modules Timings S&M CPU RM RMMA Doom 3 FarCry F.E. A.R.
Formoza F550-ADES nForce 550 1.2 Patriot
4-4-4-12 1T
+
+
+
+
+
+
Transcend
4-4-4-12 1T
+
+
+
+
+
+
Corsair
4-4-4-12 1T
+
+
+
+
+
+
Hynix
4-4-4-12 1T
+
+
+
+
+
+
Samsung
4-4-4-12 1T
+
+
+
+
+
+
Kingmax
4-4-4-12 1T
+
+
+
+
+
+
Kingston
4-4-4-12 1T
+
+
+
+
+
+
ECS AMD690GM-M2 AMD 690G 07/01/13 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-15 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-15 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
ECS GeForce6100SM-M NVIDIA GeForce 6100 + nForce 405 06/10/16 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-13 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
Biostar NF520-A2 NVIDIA nForce 520 LE 4.26 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-15 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-15 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
Biostar TA690G AM2 AMD 690G P¾C‚ 09.03.07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-13 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
Foxconn K8M890M2MB-KRS2H VIA K8M890 P26 Patriot
-
-
-
-
-
-
-
Transcend
5-5-5-13 2T
+
+
+
+
+
+
Corsair
-
-
-
-
-
-
-
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-13 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
-
-
-
-
-
-
-
Sapphire Pure Element (PE-AM2RS485M) ATI Radeon Xpress 1150 2K070617 Patriot
5-5-5-15 1T
+
+
+
+
+
+
Transcend
5-4-4-12 1T
+
+
+
+
+
+
Corsair
5-4-4-12 1T
-
+
+
+
-
+
5-5-5-15 1T
+
+
+
+
+
+
Hynix
5-5-5-15 1T
+
+
+
+
+
+
Samsung
5-5-5-13 1T
+
+
+
+
+
+
Kingmax
5-5-5-15 1T
+
+
+
+
+
+
Kingston
4-4-4-12 1T
+
+
+
+
+
+
Gigabyte MA69VM-S2 AMD 690V F4 Patriot
4-4-4-12 1T
+
+
+
+
+
+
Transcend
4-4-4-12 1T
+
+
+
+
+
+
Corsair
4-4-4-12 1T
+
+
+
+
+
+
Hynix
4-4-4-12 1T
+
+
+
+
+
+
Samsung
4-4-4-12 1T
+
+
+
+
+
+
Kingmax
4-4-4-12 1T
+
+
+
+
+
+
Kingston
4-4-4-12 1T
+
+
+
+
+
+
MSI K9AG Neo2-Digital AMD 690G 1B13 Patriot
4-4-4-12 2T
+
+
+
+
+
+
Transcend
4-4-4-12 2T
+
+
+
+
+
+
Corsair
4-4-4-12 2T
+
+
+
+
+
+
Hynix
4-4-4-12 2T
+
+
+
+
+
+
Samsung
4-4-4-12 2T
+
+
+
+
+
+
Kingmax
4-4-4-12 2T
+
+
+
+
+
+
Kingston
4-4-4-12 2T
+
+
+
+
+
+
MSI K9NGM3 NVIDIA GeForce 7050 2.0B15 Patriot
4-4-4-12 1T
+
+
+
+
+
+
Transcend
4-4-4-12 1T
+
+
+
+
+
+
Corsair
4-4-4-12 1T
+
+
+
+
+
+
Hynix
4-4-4-12 1T
+
+
+
+
+
+
Samsung
4-4-4-12 1T
+
+
+
+
+
+
Kingmax
4-4-4-12 1T
+
+
+
+
+
+
Kingston
4-4-4-12 1T
+
+
+
+
+
+
PCChips A13G+ NVIDIA GeForce 6100 + nForce 405 19/03/07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-13 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
ECS 761GXM-M SiS761GX 27/03/07 Patriot
5-5-5-15 1T
+
+
+
+
+
+
Transcend
5-5-5-13 1T
+
+
+
+
+
+
Corsair
5-5-5-15 1T
+
+
+
+
+
+
Hynix
5-5-5-15 1T
+
+
+
+
+
+
Samsung
5-5-5-13 1T
+
+
+
+
+
+
Kingmax
5-5-5-15 1T
+
+
+
+
+
+
Kingston
-
-
-
-
-
-
-
Formoza FM2A692-GHG AMD 690G 17/07/07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 2T
+
+
+
+
+
+
Corsair
5-5-5-15 2T
+
+
+
+
+
+
Hynix
5-5-5-15 2T
+
+
+
+
+
+
Samsung
5-5-5-13 2T
+
+
+
+
+
+
Kingmax
5-5-5-15 2T
+
+
+
+
+
+
Kingston
5-5-5-15 2T
+
+
+
+
+
+
ECS NFORCE6M-A NVIDIA nForce 520 LE 05/10/07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 1T
+
+
+
+
+
+
Corsair
5-5-5-15 1T
+
+
+
+
+
+
Hynix
5-5-5-15 1T
+
+
+
+
+
+
Samsung
5-5-5-13 1T
+
+
+
+
+
+
Kingmax
5-5-5-15 1T
+
+
+
+
+
+
Kingston
5-5-5-15 1T
+
+
+
+
+
+
ECS GeForce7050M-M NVIDIA GeForce 7050 28/08/07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend
5-5-5-13 1T
+
+
+
+
+
+
Corsair
5-5-5-15 1T
+
+
+
+
+
+
Hynix
5-5-5-15 1T
+
+
+
+
+
+
Samsung
5-5-5-13 1T
+
+
+
+
+
+
Kingmax
5-5-5-15 1T
+
+
+
+
+
+
Kingston
5-5-5-15 1T
+
+
+
+
+
+
Palit N61S NVIDIA GeForce 6100 + nForce 405 31/05/07 Patriot
5-5-5-15 2T
+
+
+
+
+
+
Transcend 5-5-5-13 1T + + + + + +
Corsair 5-5-5-15 1T + + + + + +
Hynix 5-5-5-15 1T + + + + + +
Samsung 5-5-5-13 1T + + + + + +
Kingmax 5-5-5-15 1T + + + + + +
Kingston 5-5-5-15 1T + + + + + +

Conclusions

Our overall impressions are certainly positive. Although we tested mostly inexpensive motherboards, all of them demonstrated absolute stability in the 8-hour stress test (our experience in 24-hour stress tests shows that the 8-hour cycle is quite trustworthy; in other words, it will detect all instabilities). We used a CPU with relatively high power consumption and a powerful graphics card (it was powered from an additional PSU cable), which also generated an increased load on a CPU voltage regulator, which could (theoretically) fail to cope with this load.

What concerns compatibility with various memory modules, only three motherboards had problems with this issue. Sapphire Pure Element (PE-AM2RS485M) on ATI Radeon Xpress 1150 (RS485+SB600) failed two tests with Corsair memory modules and default timings (all tests were passed after we adjusted the timings manually). Foxconn K8M890M2MB-KRS2H and ECS 761GXM-M failed to start up with Kingston modules (both motherboards), and the Foxconn motherboard also failed to start up with Corsair and Patriot modules (of course, it's a matter of individual samples used in the tests, not company products in general). So they have apparent problems with compatibility, because these motherboards work well with other memory modules. Interestingly, both motherboards are based on relatively low-abundance chipsets. However, it's not correct to apply these conclusions to other motherboards based on SiS and VIA chipsets. We have too little data. Besides, chipsets do not affect memory controllers in CPUs directly.

Automatically detected timings not always match SPD data. But computers are almost always stable (except for a single case we've already mentioned). Some motherboards automatically set the minimal value and 1T/2T, which helps raise performance. However, as this parameter is not documented in SPD specs, it should be set to maximum by default for higher stability (then it can be reduced by a user, or by a person who configures computers). So when you choose a motherboard for computers with high requirements to stability, you'd better choose products that use the least aggressive timings and 2T by default, and vice versa, minimum timings, if your tasks are not so critical, and you want to raise performance.

Dmitri Laptev (lpt@ixbt.com)
January 21, 2008



Latest News | Platform | Coolers | HDD/DVD | Video | Sound | Network | Imaging | Mobile
Monthly | Rightmark Tools | Search | Forum | Mailing | Links | Advertise | About Us

Copyright © by Digit-Life.com, 1997-2008. Produced by iXBT.com
Design by Explosion