AMD 790GX Motherboard Roundup
|
Overclocking
Phenom X4 9550 (2.2 GHz) |
Clock, MHz |
FSB Clock, MHz (multiplier) |
Core voltage, V |
HT/CPU NB bus frequency (multiplier), MHz |
Reboot to safe settings after failed overclocking attempt |
ASUS M3A78-T |
2925 |
265 (x11) |
1.35 |
2124 (x8) |
Second reset required |
Biostar TA790GX A2+ |
2728 |
248 (x11) |
td>1.22 |
2232 (x9) |
Automatically |
ECS A790GXM-A |
2695 |
245 (x11) |
1.26 |
2205 (x9) |
Automatically |
Foxconn A7DA-S |
2695 |
245 (x11) |
1.36 |
2205 (x9) |
Second reset required |
Gigabyte MA790GP-DS4H |
2750 |
250 (x11) |
1.26 |
2000 (x8) |
Automatically |
MSI DKA790GX |
3036 |
276 (x11) |
1.35 |
2208 (x8) |
Second reset required |
MSI demonstrates the best result. ASUS follows in its trail, and the other motherboards are behind. It must be noted that ASUS cannot raise its reference frequency above 240-250 MHz because of problems with asynchronous operation of the GPU (the graphics core itself has a good overclocking potential and can be overclocked using a corresponding option). The table contains the result obtained with a graphics card installed.
Phenom X4 9850 Black Edition (2.5 GHz) |
Clock (multiplier), MHz ACC OFF |
Clock (multiplier), MHz ACC=2% |
FSB Clock, MHz ACC=2% |
Core voltage, V ACC=2% |
HT/CPU NB bus frequency (multiplier), MHz ACC=2% |
ASUS M3A78-T |
2958 (x14.5) |
3100 (x15.5) |
200 |
1.50 |
2000 (x10) |
Biostar TA790GX A2+ |
3000 (x15) |
3000 (x15) |
200 |
1.47 |
2000 (x10) |
ECS A790GXM-A |
2900 (x14) |
2900 (x14) |
200 |
1.45 |
2000 (x10) |
Foxconn A7DA-S |
2900 (x14) |
3300 (x16.5) |
200 |
1.52 |
2000 (x10) |
Gigabyte MA790GP-DS4H |
2958 (x14.5) |
3297 (x16) |
206 |
1.54 |
2060 (x10) |
MSI DKA790GX |
2958 (x14.5) |
3100 (x15.5) |
200 |
1.55 |
2000 (x10) |
A motherboard can be blamed for limited CPU overclocking using a multiplier primarily if its voltage regulator fails to supply increased voltage to an overclocked CPU. That's exactly what happens in motherboard from Biostar and ECS. In both cases temperature of VRM elements grew above 90°C, when we tried to raise voltage, and these motherboards either turned off or our tests froze. Thus, further overclocking was impossible on these motherboards no matter whether Advanced Clock Calibration was enabled or disabled in BIOS. Yet, this technology expanded the overclocking potential of all other motherboards, although its effect varied: Gigabyte and Foxconn did better than MSI and ASUS. In case of MSI, it was insufficient VRM coolingthat acted as a constraining factor. What concerns ASUS, we cannot say for sure why it failed to overclock to a higher level -- we cannot complain about overheating, a heat sink on MOSFETs does not grow hot. Perhaps the problem is in some circuit restrictions.
Integrated audio quality
The integrated audio quality of each motherboard was tested in the 16 bit 44 kHz mode using RightMark Audio Analyzer 6.0.5 and the ESI Juli@ sound card:
Motherboard |
ASUS M3A78-T |
Biostar TA790GX A2+ |
ECS A790GXM-A |
Foxconn A7DA-S |
Gigabyte MA790GP-DS4H |
MSI DKA790GX |
Audio codec |
Realtek ALC1200, HDA (7.1+2) |
Realtek ALC888, HDA (7.1+2) |
Realtek ALC888, HDA (7.1+2) |
Realtek ALC888, HDA (7.1+2) |
Realtek ALC889A, HDA (7.1+2) |
Realtek ALC888, HDA (7.1+2) |
Frequency response (from 40 Hz to 15 kHz), dB |
+0.01, -0.03 (Excellent) |
+0.04, -0.11 (Excellent) |
+0.02, -0.15 (Very good) |
+0.02, -0.14 (Very good) |
+0.04, -0.16 (Very good) |
+0.03, -0.20 (Very good) |
Noise level, dB (A) |
-89.9 (Good) |
-71.2 (Average) |
-86.3 (Good) |
-86.9 (Good) |
-76.6 (Average) |
-88.6 (Good) |
Dynamic range, dB (A) |
89.9 (Good) |
89.9 (Good) |
86.1 (Good) |
87.0 (Good) |
76.5 (Average) |
88.7 (Good) |
THD, % |
0.0042 (Very good) |
1.514 (Bad) |
0.0062 (Very good) |
0.0040 (Very good) |
0.018 (Good) |
0.0063 (Very good) |
Harmonic distortion + noise, dB(A) |
-82.3 (Good) |
-34.2 (Very bad) |
-78.7 (Average) |
-80.9 (Good) |
-70.4 (Average) |
-82.1 (Good) |
IMD + Noise, % |
0.012 (Very good) |
2.119 (Bad) |
0.017 (Very good) |
0.014 (Very good) |
0.045 (Good) |
0.012 (Very good) |
Channel crosstalk, dB |
-90.2 (Excellent) |
-72.5 (Good) |
-90.9 (Excellent) |
-87.9 (Excellent) |
-81.0 (Very good) |
-91.2 (Excellent) |
IMD at 10 kHz, % |
0.010 (Very good) |
1.908 (Bad) |
0.017 (Very good) |
0.014 (Very good) |
0.047 (Good) |
0.012 (Very good) |
General performance |
Very good |
Average |
Very good |
Very good |
Good |
Very good |
All manufacturers chose 8-channel HDA codecs from Realtek (with an extra stereo channel for the front panel). So we didn't expect to see significant differences in their test results. However, Biostar and Gigabyte (quite unexpectedly) screwed up. The analog section of Biostar motherboards often leaves much to be desired. Besides, this board lacks S/PDIF ports on the rear panel, so users will almost certainly have to install a sound card. What concerns Gigabyte, for one, this board has S/PDIF Out; for two, its subjective analog audio quality does not produce a negative impression. Its audio level is lower than in the other boards with the same audio mixer settings, so we had to increase the volume of the speakers. But we didn't notice any distinct distortions, although the dynamic range is quite naturally narrower in this situation. Nevertheless, it's strange to see these results, considering that the motherboard uses an advanced codec, which demonstrates excellent results in other motherboards from Gigabyte. So we took another motherboard of this model and tested it with the latest BIOS version (F3M) and Realtek drivers (2.10). However, we got almost the same results. The other motherboards offer good quality of analog audio, so you can use them with not very expensive speakers or headphones.
Conclusions
First of all we'd like to mention an excellent tendency of the past year: imposing motherboards with excellent expandability (owing to full-size design) with the integrated graphics core. Moreover, these motherboards are equipped with modern digital video outputs. And the graphics core itself can cope with 3D load of quiet casual games. Besides, it offloads a CPU, as it can decode HD video. This feature will come in handy for low-noise configurations, because it allows to do without a graphics card (even the weakest card with passive cooling raises temperature inside a PC enclosure and thus requires better ventilation, meaning more noise). As we can see in test results, such a system is really power efficient.
This conclusion applies to all these motherboards. Here is the most practical strategy how to choose such a motherboard: determine which model meets your key requirements (for example, whether you want a motherboard for overclocking or you will use nominal frequencies, whether the integrated GPU performance matters or you plan to install a graphics card or two, etc). What concerns differences in bundles, they are not very important, as full-size ATX motherboards are actually designed to let users assemble what they really need (for example, to add the second network port or eSATA, if your motherboard does not offer enough of them).
Write a comment below. No registration needed!
|
|
|
|
|