Topic: 6 GPU Sapphire Tri-X R9 290 OC (non-x) Rig (Read 5969 times)
I did some more multimeter testing on the PSUs. The 1300W tested fine last week when it arrived, but last night I tested it again and heard a pop followed by the unmistakable smell of charred PCB. So I have an RMA started with EVGA. The 1050W Seasonic now shows 12.75V which from what I've read is out of spec, so I'm planning to call Seasonic tomorrow. The 850W Corsair still tested perfect. I tried it on the new MB and it booted to uefi. I then installed the hard drive, but kept getting Windows errors during the boot up so I'm assuming that it was fried by the faulty 1300w PSU. I guess to look at things "glass half full", I didn't miss out on huge profits last week. Hope things improve soon.
Well I installed a replacement motherboard today and I still just get a click from the power supplies. The CPU fan doesn't budge. Should I get another CPU of are there any other tricks I should try? I've tested other sticks of RAM and three different PSUs. I've tried booting with just the 8 pin CPU power and ATX plugged in. It's really frustrating because it was mining fine and I shut it down to install a new larger PSU and it hasn't booted up since.
Bummer, tht really is a shitty situation man. CPUs almost never fail especially since you weren't overclocking, the motherboard or PSU is almost 100X more likely to. One of my other miners wouldn't boot with the PSU daisychained So I was only able to get one card installed on tht one. Good luck with figuring out what it is
Well I installed a replacement motherboard today and I still just get a click from the power supplies. The CPU fan doesn't budge. Should I get another CPU of are there any other tricks I should try? I've tested other sticks of RAM and three different PSUs. I've tried booting with just the 8 pin CPU power and ATX plugged in. It's really frustrating because it was mining fine and I shut it down to install a new larger PSU and it hasn't booted up since.
I made some headway last weekend. I gave up on Linux after being unsuccessful with BAMT and SMOS. I installed Windows 8.1 and was mining in 30 minutes. Amazon still hasn't shipped my 1300W PSU so I'm currently only running 5 GPUs with an 850W and 1050W. Out of the box they were all running around 860 kh/s. I've spent the last couple evenings tuning and things are looking promising.
Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
..
Here are my current settings.
Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
..
Here are my current settings.
Code:
"intensity" : "20,20,20,20,20",
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
Yesterday I bought 4x Sapphire 290 TRI-x after few tests with basic core and mem clocks I ended up with 700kh/s.
So I have decided to test your shared configuration,which ended up with very nice 900-910 kh/s
I had to delete these rows in config:
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
and had to lower gpu-engine to 1020. In case I leave it at 1040, temperature rocket sky to 85, which is too high for me in long term.
So I want to say BIG THANK YOU. Thank you dloganbill for sharing your tips..
Sending greetings from Czech republic.
I'm pretty sure the motherboard is toast, especially after reading all the bad reviews it has.
So I'm thinking of getting one of these. Any preferences?
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128592
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128681
So I'm thinking of getting one of these. Any preferences?
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128592
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128681
I haven't touched voltages.
Have you tried this bios https://litecointalk.org/index.php?topic=12830.390
and -I 22 --thread-concurrency 32768 -w 256
I get better hash rate with it but occasional driver crashs.
btw:Are you running your rigs on win 8.1 or 7?
You will get better performance from your cards if you revert the display to Windows basic high contrast, So the GPU's don't have to waste processing power on windows 7-8 themes which will make one or two of your cards hash slower.
I'm pretty frustrated at the moment. I finally received the 1300W PSU and installed it. When I went to power on, I would only get a single click out of the PSUs. It wouldn't boot to bios or anything. I started removing components and even with just ATX 24-pin and 8-pin CPU it won't boot. Cleared CMOS nothing. I'm guessing the motherboard has failed or CPU. Any ideas? It had been running pretty good for the last while. It had a few hangs in Windows that required a hard reset.
Ram reseat the ram. Also remove sticks to check if one has failed. Other than tht it must be ur motherboard.. never heard of a bad cpu.
I'm pretty frustrated at the moment. I finally received the 1300W PSU and installed it. When I went to power on, I would only get a single click out of the PSUs. It wouldn't boot to bios or anything. I started removing components and even with just ATX 24-pin and 8-pin CPU it won't boot. Cleared CMOS nothing. I'm guessing the motherboard has failed or CPU. Any ideas? It had been running pretty good for the last while. It had a few hangs in Windows that required a hard reset.
To add multiple PSUs, you'll need one of these for each additional PSU. I'm using 2 power supplies so I only needed one adapter. You would need two adapters to run 3.
I am using this one:
http://www.amazon.com/gp/aw/d/B00DL3L2J6/ref=mp_s_a_1_2?qid=1394485633&sr=8-2
Many people are using this one:
http://www.amazon.com/gp/aw/d/B009P98Q8U
Or you can go cheap and use this method:
http://www.overclock.net/t/14455/multiple-power-supply-guide
I am using this one:
http://www.amazon.com/gp/aw/d/B00DL3L2J6/ref=mp_s_a_1_2?qid=1394485633&sr=8-2
Many people are using this one:
http://www.amazon.com/gp/aw/d/B009P98Q8U
Or you can go cheap and use this method:
http://www.overclock.net/t/14455/multiple-power-supply-guide
Hey guys
6 cards are coming my way tomorrow and I still do not know one key information.
I would like to buy 3x 850w power supplies, but I do not know how to connect them when all 6 cards will be on one motherboard. Can you give me a hint?
Do I just connect one power supply to motherboard and with 2cards, second power supply just with another two cards,third power supply to last two cards?
In that case, how do I turn the additional power supplies on, when the computer is starting up?
thanks
M.
6 cards are coming my way tomorrow and I still do not know one key information.
I would like to buy 3x 850w power supplies, but I do not know how to connect them when all 6 cards will be on one motherboard. Can you give me a hint?
Do I just connect one power supply to motherboard and with 2cards, second power supply just with another two cards,third power supply to last two cards?
In that case, how do I turn the additional power supplies on, when the computer is starting up?
thanks
M.
I built my case from about $20 worth of stuff from Lowes. The rest of your list looks good. As for power, mine is currently pulling about 15 amps so I have it on two separate breakers. I'll get new amp readings once I get number 6 running. I recommend getting a couple $20 killawatts for setting things up.
As for ROI, altcoins have taken a big dive. I was originally planning on a 3-4 month ROI, but now I'm not sure how long it will take. $8k will take a while at current rates.
As for ROI, altcoins have taken a big dive. I was originally planning on a 3-4 month ROI, but now I'm not sure how long it will take. $8k will take a while at current rates.
Looks nice! My plan was to buy something almost identical to yours. For 2 rigs:
GPU
R9 290 (x12)....6 per rig
MOBO
ASRock H81 Pro BTC (x2)
CPU
Intel Pen G3220 Haswell LGA 1150 Dualcore (x2)
PSU
Silverstone 1000W 80+ Gold (x2)
1300W 80+ Gold PSU (x2)
Dual PSU Adapter Cable (x2)
MEMORY
Kingston 4GB DDR3 1333 (x2)
SSD/HDD
SanDisk 32GB (x2)
OS
MS Windows 8.1 64 bit OEM (x2)
MISC
PCIe 1x to 16x Riser (x2)
Linksys Wireless USB Adp (x2)
Mining Case (x2)
~$8,000
For hopefully around 10-11 Mh (using wafflepool hopefully it pays for itself in 5-9 months).
How do you think this looks? Am I missing anything? Is it better to go with 3 or 4 GPUs per mobo...or is going with 6 per mobo alright?
I'm unsure of what kind of mining case I should get. And would I need to plug in one rig into one wall outlet and another rig to another outlet I take it? I was worried that perhaps one rig with 6 GPUs would be too much power for my apartment outlets.
Thanks for any input.
GPU
R9 290 (x12)....6 per rig
MOBO
ASRock H81 Pro BTC (x2)
CPU
Intel Pen G3220 Haswell LGA 1150 Dualcore (x2)
PSU
Silverstone 1000W 80+ Gold (x2)
1300W 80+ Gold PSU (x2)
Dual PSU Adapter Cable (x2)
MEMORY
Kingston 4GB DDR3 1333 (x2)
SSD/HDD
SanDisk 32GB (x2)
OS
MS Windows 8.1 64 bit OEM (x2)
MISC
PCIe 1x to 16x Riser (x2)
Linksys Wireless USB Adp (x2)
Mining Case (x2)
~$8,000
For hopefully around 10-11 Mh (using wafflepool hopefully it pays for itself in 5-9 months).
How do you think this looks? Am I missing anything? Is it better to go with 3 or 4 GPUs per mobo...or is going with 6 per mobo alright?
I'm unsure of what kind of mining case I should get. And would I need to plug in one rig into one wall outlet and another rig to another outlet I take it? I was worried that perhaps one rig with 6 GPUs would be too much power for my apartment outlets.
Thanks for any input.
I have the rig installed in my basement, so noise was never a big concern. According to a decibel meter app it is currently 63-68 dB holding the phone about 1.5 feet away. The screenshot below was taken with fans at
Good news! I got an email from Amazon today that says my 1300W PSU should arrive on Monday. I'll finally get that 6th card mining and be able to tune harder. As you can see below, the 1050W PSU has been overloaded for the last week. My goal is a stable 945kh per card. I'm also planning on trying out Stilt's BIOS soon.
Code:
"gpu-fan" : "65,35,57,37,72",
Good news! I got an email from Amazon today that says my 1300W PSU should arrive on Monday. I'll finally get that 6th card mining and be able to tune harder. As you can see below, the 1050W PSU has been overloaded for the last week. My goal is a stable 945kh per card. I'm also planning on trying out Stilt's BIOS soon.
How is the noise of the rig?
when fans over 70%
at those clock speeds how much the fan % go normally?
when fans over 70%
at those clock speeds how much the fan % go normally?
I haven't touched voltages.
Don't go with what the graphics cards usually draw. You will be hashing, so peak loads is what you will be needing to look at. A list of peak loads can be found here: http://forum-en.msi.com/faq/article/power-requirements-for-graphics-cards
Multiple the A (Ampere) with the V (Volts) to get the W (watts).
In your case, the 290 hogs up 31 Amperes through the 12 Volt rails. So at peak load, it would consume 31 x 12 = 372 Watts each.
Also Don't forget to allocate some power for the motherboard, RAM and HDD... about 150 Watts should be fine.
Most decent PSU are 80% efficiency. That means you will draw up 1200W from the wall if you are using a 1000W PSU. You ca use that as a rough calculation to estimate your electricity costs.
Multiple the A (Ampere) with the V (Volts) to get the W (watts).
In your case, the 290 hogs up 31 Amperes through the 12 Volt rails. So at peak load, it would consume 31 x 12 = 372 Watts each.
Also Don't forget to allocate some power for the motherboard, RAM and HDD... about 150 Watts should be fine.
Most decent PSU are 80% efficiency. That means you will draw up 1200W from the wall if you are using a 1000W PSU. You ca use that as a rough calculation to estimate your electricity costs.
This is incorrect! Hashing does not use peak power consumption. Please use Kill A Watt metter and my 280x's undervolted use about 235-250 watts while hashing. If you plan 300 for each 290= 1800w and 250 for mobo = 2050w. Maybe give some more headroom above that if you want but I doubt you will need it especially if you are not overclocking your CPU.
I made some headway last weekend. I gave up on Linux after being unsuccessful with BAMT and SMOS. I installed Windows 8.1 and was mining in 30 minutes. Amazon still hasn't shipped my 1300W PSU so I'm currently only running 5 GPUs with an 850W and 1050W. Out of the box they were all running around 860 kh/s. I've spent the last couple evenings tuning and things are looking promising.
Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
I think I can push them a little harder, but am currently limited by power. The 1050W is essentially maxed out, so I have the engines turned down on some. Here is my screen after 6 hours of mining. I would have had a longer run time, except the power blinked last night and for some reason it didn't autostart like it should have.
Here are my current settings.
What 's the vcc? Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
I think I can push them a little harder, but am currently limited by power. The 1050W is essentially maxed out, so I have the engines turned down on some. Here is my screen after 6 hours of mining. I would have had a longer run time, except the power blinked last night and for some reason it didn't autostart like it should have.
Here are my current settings.
Code:
"intensity" : "20,20,20,20,20",
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
MMMMM. The sweet smell of a brand new mining rig running at full capacity.
I made some headway last weekend. I gave up on Linux after being unsuccessful with BAMT and SMOS. I installed Windows 8.1 and was mining in 30 minutes. Amazon still hasn't shipped my 1300W PSU so I'm currently only running 5 GPUs with an 850W and 1050W. Out of the box they were all running around 860 kh/s. I've spent the last couple evenings tuning and things are looking promising.
Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
I think I can push them a little harder, but am currently limited by power. The 1050W is essentially maxed out, so I have the engines turned down on some. Here is my screen after 6 hours of mining. I would have had a longer run time, except the power blinked last night and for some reason it didn't autostart like it should have.
Here are my current settings.
Things I learned along the way that seem to have helped:
1. Use Catalyst 13.12. The 14.2 beta drivers were very unstable.
2. Don't bother with Trixx. Install MSI Afterburner and use the settings shown on this post (link).
3. I don't know if this helped or not, but I'm currently running with the bios switch toggled up (toward the end with DVI ports)
I think I can push them a little harder, but am currently limited by power. The 1050W is essentially maxed out, so I have the engines turned down on some. Here is my screen after 6 hours of mining. I would have had a longer run time, except the power blinked last night and for some reason it didn't autostart like it should have.
Here are my current settings.
Code:
"intensity" : "20,20,20,20,20",
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
"xintensity" : "0,0,0,0,0",
"rawintensity" : "0,0,0,0,0",
"worksize" : "512,512,512,512,512",
"kernel" : "ckolivas,ckolivas,ckolivas,ckolivas,ckolivas",
"lookup-gap" : "2,2,2,2,2",
"thread-concurrency" : "25601,25601,27401,25601,27401",
"shaders" : "0,0,0,0,0",
"gpu-threads" : "1,1,1,1,1",
"gpu-engine" : "1040,1020,1020,1040,1020",
"gpu-memclock" : "1500,1500,1500,1500,1500",
"gpu-memdiff" : "0,0,0,0,0",
"gpu-powertune" : "20,20,20,20,20",
"gpu-vddc" : "0.000,0.000,0.000,0.000,0.000",
"temp-cutoff" : "90,90,90,90,90",
"temp-overheat" : "85,85,85,85,85",
"temp-target" : "78,78,78,78,78",
"auto-fan" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "1",
"failover-only" : true,
"failover-switch-delay" : "60",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"log" : "10",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "1",
"tcp-keepalive" : "30",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"device" : "0,1,2,3,4"
}
Jump to: