Reference R9 290 suitable for Crossfire?

[Morgan Everett]

I'm thinking of crossfiring my current 290 (non-reference) with a reference 290. I'm aware the reference card is a loud beast. However, it would be the secondary card, and I would undervolt both cards. Would that be enough to save my eardrums/sanity, or not?

 

Note I'm not looking for general advice beyond this. I'm just interested in this question. 

[manikyath]

seems a bit counter productive to get a second card for CF, and then make both cards slower to be honest...

[Morgan Everett]

Not sure I follow. Sorry.

[ventus0907]

seems a bit counter productive to get a second card for CF, and then make both cards slower to be honest...

 

 

Not sure I follow. Sorry.

 

he's just saying why crossfire then undervolt the cards

[Morgan Everett]

he's just saying why crossfire then undervolt the cards

 

Why not? Good way of keeping temps down. It needn't make the cards "slower", either. 

[Guest Generallee]

Why not? Good way of keeping temps down. It needn't make the cards "slower", either. 

 

If you plan on CF 290's i'd suggest to watercool them 

[Morgan Everett]

If you plan on CF 290's i'd suggest to watercool them 

 

No thanks.

[Guest Generallee]

No thanks.

 

Then you're gonna cope with a LOT of heat

[boarder2k7]

My reference 290 crossfire rig does fine keeping the temps in spec on 45% fan, and I'm over clocked to 1000 MHz, stock voltages. I'm not sure how much thermal throttling is in play, but they behave fine.

If you're under volted on stock clock I think you'll be fine as long as your case airflow is good enough to keep them from sucking in their own hot air.

[Morgan Everett]

My reference 290 crossfire rig does fine keeping the temps in spec on 45% fan, and I'm over clocked to 1000 MHz, stock voltages. I'm not sure how much thermal throttling is in play, but they behave fine.

If you're under volted on stock clock I think you'll be fine as long as your case airflow is good enough to keep them from sucking in their own hot air.

 

Is the noise level tolerable? 

[boarder2k7]

Is the noise level tolerable? 

 

Just ran a ~30 minute test for you. 72 F room, Valley Bench, 1920x1200, Ultra Settings, 2x AA, came in at around 44 FPS

This card is running 1000 MHz core, 1350 MHz Memory, stock voltage, 150% max power limit.

 

Stock fan profile gave me this:

 

At 20% fan speed, I can't hear the card over the PSU and CPU fans on this open bench. (2200 RPM on the CPU fan, and both are 90mm PSU is a Sunon, the other is whatever came on this stock Athlon II X4 heatsink I swiped from an HP tower)

 

For comparison the idle card fan is about 1100 RPM. At 42% its up to 2290 RPM and is certainly noticeable, comparable to the CPU fan at about 3400 RPM, but a bit higher pitched, almost a hiss type note rather than the lower pitched woosh of the CPU fan. When moving around the bench, much of this sound seems to be intake noise, rather than exhaust noise, so I'd expect it to be much more muted inside a case. The headphones I have here (ATH M30) will block all the GPU noise when set at 10% volume on my mobo.

 

I hate the arbitrary comparison type of measurements I'm giving you here, but all I have available for a dB meter right now is my phone, and it doesn't seem to pick up the fan sounds very well at all for some reason. I can crank the fans up and down and barely see a difference in the meter readings.

 

For the sake of being thorough though, from my seating position (about 2 feet from the back of the card, one foot sideways from the intake fan) the meter reads about 29 dB with the fan at 20%

30% is also about 30 dB

40% is about 34 dB

45% also 34 dB but its definitely louder

50% its still unchanged

75% sees a sudden jump to 48 dB as if the phone is just now hearing it (there are more present low tones which is what I think the phone is picking up)

100% which is insanity jumps to around 54 dB from the phone on the table. Sounds a bit like a small hair dryer running in the next room

 

The CPU fan running at 4500 RPM is reading in at about 49 dB.

 

For additional comparison the phone is picking up this Dell mechanical keyboard at about 56 dB as well, and if I clear my throat it jumps to 73.... So like I said, the frequency response of the phone dB meter is extremely suspect.

 

For a better idea of what I mean by the description of the sounds, here is a spectrogram (again from a probably inaccurate phone) of a few data points. Also keep in mind that I think the 0 dB reference point here is actually 90 dB which is the max volume any dB meter on the Galaxy S5 ever displays, but again, I'm not sure. (I miss having my real equipment set up and available.... soon I hope)

 

Room background (includes 20% GPU Fan, CPU fan at 2200 RPM, PSU Fan and some wind noise from the winter blowing on the house)

 

GPU Fan at 50%

 

GPU Fan at 100%

 

And the CPU fan at 4500 RPM (100%)

 

With default fan profiles, the thing never spins up over 45% which I really don't find bothersome. As with all noise related things though, its very open to interpretation.

 

 

I personally got the reference style card because the system I'm going to put it in has a 95 Watt thermal stability limit in its PCI Express bay, so I needed something that can offload the majority of the heat outside the case. (Putting this into a server chassis. They're meant to cool the HDD, processor, and memory, but not really the PCI bay, its bypass air only)

 

Any other questions please ask.