Asus GTX 285 Matrix - Fastest Single GPU Card Out there?

Videocards/VGA Reviews by leeghoofd @ 2009-08-17

I was intrigued by this card and decided to put it head to head, not with all video cards out there, but against its reference design twin brother. I have been using a GTX 285 since its release and have been a pretty happy camper. But what extras this Asus card has to offer you want to check out!

  • prev
  • next

Headgear and sweatpants on

What's that noise ?

A lot of users asked me to include the noise levels generated by this card. As Asus claims it cools better, I suspected at first a higher spinning fan. Boy was I wrong. Besides the different cooler design (bigger heatpipes and the improved copper surface with the GPU's IHS), the Asus fan used is also a newer version then the one I have on my GTX 285 (fan nr MGT8012YR vs YB, both rated at 0.48A) I must say that I had my vanilla GTX 285 since it's first release and more recent versions might have got another fan version than my card (same counts for the PCB design which has changed too on later models)

What does that bring temp wise ? I ran 3 loops of 3D06 (without CPU test) to stress the cards. Monitoring was done via GPU-Z (which creates a nice log file) The results were averaged and then you get this:

Madshrimps (c)


At idle we only got a few degrees difference, but under load there's a gap of 10°C. This with the AUTO setting selected for both cards. The Asus Matrix runs at almost half the AUTO speed of the vanilla card (26 vs 40%). I have to warn you that due to the different fans used, that half the AUTO setting isn't the same as half the rpm. Secondly, keep in mind that the vanilla card was at stock reference clocks. These being a bit lower than the Matrix card, yet the Asus design is cooler all the way ! Good job there.

For the dBA testing we used our trustworthy Smart Sensor AR824 device. The device was put at 30cm from the video card, the microphone pointed versus the fan. First we measured the ambient noise with the PC off. Then we fired up the 3D06 benchmark, held our breath and after successful completion of the runs, looked at the max registered dBa level.

Madshrimps (c)


So we expect more silence then too ? Nope lads, it started nice at idle, where it was less noisy. This mainly due to the slower spinning fan. But once under load it topped at 61.2dBa. The noise was not generated by the fan. This Asus card has got a peculiar noise when being loaded. It's really pretty hard to explain. Most of us are used to a squealing noise when doing benches... The Matrix card does some electron gurgling act. It generates to my liking far too much noise and hope this was limited to my particular sample.


Power consumption :

As the card had several 2/3D profiles, we were obliged to verify the wattage consumption of this video card, comparing it with the stock eVGA card and the ATI Radeon 4890. For the Load test we used the Canyon flight of 3DMark 06 as it draws a lot of current from the mains... Idle mode was measured during 15 minutes idling in Windows.

Madshrimps (c)


As reported before the Optimized setting combines the best out of the two worlds and this also reflects from the graph. Gaming mode I wouldn't select as it has the same 3D clocks as the Optimized mode, but idles at far higher clocks making it draw a far higher wattage from the wall. Off topic, I was nicely surprised by the ATI under load.

Before we go to the verdict page I tried the card with the Folding at home GPU client. I usually use this application to burn in CPU's. So I used it too for this video card letting it run overnight, gradually increasing core clocks. Be aware that Folding@Home clocks can usually be set higher than for 3D marks, with no sign of instability...

Madshrimps (c)
(Click to enlarge)

  • prev
  • next
Comment from Massman @ 2009/08/17
Too bad it's not possible to put this one under LN2; I'm pretty sure I'd get the memory to bench at 1600MHz with added cold

 

reply