Need help keeping R9 280x Cool

[zak9494]

I have 5 Diamond 280x's running on one mother board with powered risers. They are in an open case with 4 inches apart from each other.  I have 2 large Vornado fans on them.  But temps run too hot and sometimes the cards get turned off.
Ive tried at least 30 different variations of configurations to cool down the cards without losing speed.
Its about 71 F in my apartment. Im getting around 700KH on each card.

GPU 0: 79 c @ 3651RPM
GPU 1: 73 c @ 3530RPM
GPU 2: 79 c @ 3671RPM
GPU 3: 88 c @ 3734RPM -- This ones keeps turning off
GPU 4: 74 c @ 2125RPM

If someone could please list configs that are working well for them or if any suggestions to be made to my current conf.

Thanks!

-------------------------------
Pool stuff….

]
,
"intensity" : "13,13,13,13,13",
"vectors" : "1,1,1,1,1",
"worksize" : "256,256,256,256,256",
"kernel" : "scrypt,scrypt,scrypt,scrypt,scrypt",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "8192,8192,8192,8192,8192",
"shaders" : "2048,2048,2048,2048,2048",
"gpu-engine" : "1026-1026,1026-1026,1026-1026,1026-1026,1026-1026",
"gpu-fan" : "0-95,0-95,0-95,0-95,0-95",
"gpu-memclock" : "1498,1498,1498,1498,1498",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "5,5,5,5,5",
"gpu-vddc" : "1.000,1.000,1.000,1.000",
"temp-cutoff" : "86,86,86,86,86",
"temp-overheat" : "87,87,87,87,87",
"temp-target" : "75,75,75,75,75",
"api-mcast-port" : "2048",
"api-port" : "4028",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "30",
"scrypt" : true,
"temp-hysteresis" : "1",
"shares" : "0",
"auto-fan" : true,
"no-submit-stale" : true,
"kernel-path" : "/usr/local/bin"
}

[1714771460]

1714771460

[1714771460]

1714771460

[1714771460]

1714771460

[rudyo]

Move the cards around and see if the problem follows.  Swap both the pci riser and the power connectors around and keep track. 

Try the system with just 4 GPU's, remove a good one and see what the bad one does.  I have seen strange things with power connections.

Oh and BTW, if all values on a single line in your config are the same, you can just list the value once instead of 5 times. You only need to list it multiple times if you want a different value for a card or cards.

Good Luck!

[Gazza1]

It could be turning off because the voltage is too low.  1.0 isn't stable.  Raise the voltage a bit.  Start at 1130 (1.13) and work your way down.  I have my 280x's set at 1100 which shows up as 1.019ish in GPUZ.

[crazyates]

CGMiner shouldn't be able to change the voltage of your cards. To change the voltage, you'll need a program like MSI AB or TRIXX.

GPU-z has two areas where it will display your voltages, and one can sometimes be false. The main one you see on your settings page is what the driver is "set" to. CGMiner will attempt to use the driver to set the voltage, which is what GPUz is reporting, but the card won't actually do it. Instead, you have to move over the "sensors" tab in GPUz and see what the card is reporting it's actual VDDC voltage at.

I would suggest downloading and running VBE7. You can use GPUz to save a copy of your cards BIOS, and then use VBE7 to tweak it and lower the voltage. This is a guaranteed method of forcing your cards to ALWAYS be undervolted.

Try going progressive lower and lower on one card, mine for a solid ~20 minutes, and then go lower to see when it fails. Then bump it up a little bit, and set the voltage for all the GPUs.

All of our Gigabyte 280x run at 1.019V, 1025/1500, 711Kh/s, 80% fan speed, and are usually in the 60C range. We keep 5 of them crammed into a 19" rack, so there's less than 4" space between them. Actually, there's more like 2" between each card.

[railzand]

the cards may be identical, but the sweet spots won't be.