New member
-
Hi all,
Im from Europe, Denmark.
Reading about bitcoin I figured I want a pice. But how can I get the lowest risk.
Mining a upcomming currency seems like the best option to me. ;)
Should it fail I can try another or sell the equipment. 8)Just setup a 4 x 7950 rig.
Its running fine on 3 GPU currently at ~ 550Kh/s on each GPU.But I tried overclocking and made it run at ~ 680Kh/s on each GPU - for a while.
Now the 4th GPU (GPU 2) seems dead. If active, the driver crash and cgminer mark it SICK or DEAD.And now for the strange part - and where i hope you can help.
If I remove the “dead” GPU 2 (I can feel the one thats cold) and restart, still GPU 2 run a ~ 80Kh/s until marked dead.
But now GPU 2 is another card. I then tried to move the card to another PCIe slot. Still dead. Switched riser cabel. Still dead. Switched CPU card. Still cgminer only get 80Kh/s until GPU 2 is marked dead. For now the system run fine on 3 GPU as long as “GPU 2” in gcminder are disabled. No driver crash. So if its not the PCIe slot on the MB, not the card, not the cable - what does then fail “GPU 2”? -
Don’t run them too hard straight away, I run 4 X 7950’s overclocked at 660K/hash each.
But you can start them off at 615K/hash each with no overclocking, set at intensity 20.I had those cards off for a two days last week while I was testing new equipment and had a problem of getting one of the cards working again, the ambient temperature was about freezing, so I assumed the card was too cold, heated the room a little (gentle and overnight to around 20C) and the card came to life next day when I fired it up.
Also if you have the monitor plugged into one of the cards try switching that around, just don’t hit the rig with a hammer. ;) Let me know how you get on.
& Welcome to feathercoin.
-
I have been trying different setting. With no overclock I can’t get 615.
Did try to switch cards around, and riser cabel, and use different PCIe slots. Always GPU 2 failed.
But I do not use --gpu-reorder that might show something else, I will try that and switch cards.
BTW: its Club 3D '13Series cards, and I have under voltage 1038 - after GPU failure - keep it on as it did not seems to change performance.
These are some of the settings I have tried.
Currently using: (550 Kh/s)
cgminer --scrypt -I 20 -g 1 -w 256 --thread-concurrency 22400 -o stratum+tcp://ltc.give-me-coins.com:3334 -u FDcoin.Worker1 -p [pass](and from within gcminer GPU 2 disabled).
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1rem 450
rem cgminer.exe --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --gpu-engine 1100 --gpu-memclock 1125 --gpu-powertune 20 --shaders 1792 --thread-concurrency 22400 -I 19 -w 256 --lookup-gap 2 -g 1 -u FDcoin.Worker1 -p [pass] --temp-target 70rem 470
rem cgminer --scrypt -o [url=http://stratum.give-me-ltc.com:80]http://stratum.give-me-ltc.com:80[/url] --thread-concurrency 21712 --intensity 16 --worksize 256 --lookup-gap 2 -g 1 --auto-gpu -u FDcoin.Worker1 -p [pass]rem 550Kh/s
cgminer --scrypt -I 20 -g 1 -w 256 --thread-concurrency 22400 -o stratum+tcp://ltc.give-me-coins.com:3334 -u FDcoin.Worker1 -p [pass]rem 550
rem cgminer --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --thread-concurrency 16384,16384 --gpu-engine 1100,1100 --gpu-memclock 1500,1500 --temp-target 70 -I 20,20 -w 256 -u FDcoin.Worker1 -p [pass]rem 580
rem cgminer.exe --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --gpu-engine 1100 --gpu-memclock 1250 --gpu-powertune 20 --shaders 1792 --thread-concurrency 22400 -I 20 -w 256 --lookup-gap 2 -g 1 -u FDcoin.Worker1 -p [pass] --temp-target 70rem 600
rem cgminer.exe --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --gpu-engine 1000 --gpu-memclock 1250 --gpu-powertune 20 --shaders 1792 --thread-concurrency 22400 -I 20 -w 256 --lookup-gap 2 -g 1 -u FDcoin.Worker1 -p [pass] --temp-target 70rem 640 HW
rem cgminer.exe --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --gpu-engine 1100 --gpu-memclock 1500 --gpu-powertune 20 --shaders 1792 --thread-concurrency 22400 -I 20 -w 256 --lookup-gap 2 -g 1 -u FDcoin.Worker1 -p [pass] --temp-target 70rem 650
rem cgminer.exe --scrypt -o stratum+tcp://ltc.give-me-coins.com:3334 --gpu-engine 1000,1100,1100,1100 --gpu-memclock 1500 --gpu-powertune 20 --shaders 1792 --thread-concurrency 22400 -I 20 -w 256 --lookup-gap 2 -g 1 -u FDcoin.Worker1 -p [pass] --temp-target 70 -
Do you run all cards without powered riser?
I think that might be my problem.
(But the problem do exists with only 3 cards also).MB: Gigabyte GA-990FXA-UD7 ATX AM3+ (GA-990FXA-UD7)
RAM: 8GB
OS: W7
Power: Seasonic X-1250 X-Series 80 Gold 1250W -
Hi and welcome.
-
[quote name=“Kirosoloks” post=“42524” timestamp=“1386691442”]
Hi and welcome.
[/quote]Thanks.
-
Switched two cards and restarted. Now all perform at 550Kh/s. Strange.
-
[quote name=“fdcoin” post=“42387” timestamp=“1386663877”]
Do you run all cards without powered riser?
I think that might be my problem.
(But the problem do exists with only 3 cards also).MB: Gigabyte GA-990FXA-UD7 ATX AM3+ (GA-990FXA-UD7)
RAM: 8GB
OS: W7
Power: Seasonic X-1250 X-Series 80 Gold 1250W
[/quote]Your MB & PSU specs are perfect, wish I had a seasonic, I run normal risers which is not recommended, but my rig is doing fine without.
[quote author=fdcoin link=topic=5751.msg42555#msg42555 date=1386694594]
Switched two cards and restarted. Now all perform at 550Kh/s. Strange.
[/quoteThat sounds like there was a poor connection between riser and card or riser and MB, the cool thing is if this ever happens again, you’ll know that you must go over everything 10 times.
-
Still running fine.
4 x ~550Kh/s at ~75 degree Celcius. (no overclock).Seems like a nights cool down have cured the malfunction. So I guess it can have been build up heat. But still strange.
Setting:
cgminer --scrypt -I 20 -g 1 -w 256 --thread-concurrency 22400 --auto-gpu --auto-fanI have ordered a powered riser cabel, but for now it seems like the MB can handle the load.
-
[quote name=“Pyxis” post=“42731” timestamp=“1386713050”]
Your MB & PSU specs are perfect, wish I had a seasonic, I run normal risers which is not recommended, but my rig is doing fine without.That sounds like there was a poor connection between riser and card or riser and MB, the cool thing is if this ever happens again, you’ll know that you must go over everything 10 times.
[/quote]Yes its a very nice PSU and MB - except I had to return the first PSU as it blown the house fuse! Sparks from behind the power cord plug. Seems like QC failed on that unit. But the replacement works.
I don’t think it was a bad connection as I did remove and replaced cards without luck, but your right, everything must be checked and rechecked.