Topic: 6950 HD rig restart problem (0.02 BTC bounty ) (Read 871 times)

i have tried many settings and it seems like heating problem...as soon as the temperature goes more then 75 degree ...the system gets restart after every 1 hour ..if i keep increasing the intensity to get more hash ..it will get restart more frequent .....any idea/suggestion how to stop restarts...

 0.02 bounty is still valid

Stock volts for a 6970 is 1.175, I run at 1.225 all day every day.

The issue sounds like it is directly related to voltage, giving it a slight bump over stock should confirm whether or not it is.

DONT overvolt... stock is 1.1v so that's the highest value you should ever set.
Remove 'shaders' as it has no affect when you set thread-concurrency.

Looks good to me. Fire it up and see how it goes.

@aulonocara does this look good
--gpu-engine 920 --gpu-memclock 1300 --lookup-gap 2 --thread-concurrency 8192 -g 1 -I 17 --temp-target 75 --gpu-powertune 20 --gpu-vddc 1.2 --shaders 1536 --auto-fan

anything to add

Oh and set powertune to 20, that is most important!

Try flashing it to a 6970 and try the following mild overclock: 1.2v, 920Mhz core, 1300MHz memory. Then try with -I 17 -g 1 --shaders 1536 --thread-concurrency 8192.

Also, I suggest 13.9 drivers for best results. You should be looking at around 475KH/s with those settings.

@ tyfeonas temp are doing fine,,no overheat

I tried with your settings (deleted bin files ) just now and after 15 min the rig restarted...not sure what is the problem..

i work with 6950 and on smos and windows no problems.

in the smos it have a autostart option?
maybe you start the card manually and then the os autostarts a second cgminer and then crashes?

what temps your card is working?


this is the setup that i am using on windows
same on smos (i replace the values from this lines to the conf file on smos)
also very important on windows try to erase all the .bin files from the cgminer folder
and then start your new config.

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
.\cgminer-3.7.2-windows\cgminer.exe --scrypt -o stratum+tcp://*******:3333 -u user**** -p pass*** -I 17 -g 1 -w 256 --shaders 1536 --thread-concurrency 7042 --gpu-engine 850 --gpu-memclock 1300 --gpu-powertune 10 --gpu-fan 40-85

--gpu-engine 570 --gpu-memclock 800 --lookup-gap 2 --thread-concurrency 5632 -g 1 -w 256 --intensity 18 --temp-target 75 --auto-fan --shaders 1408

Running with above settings now and its running at 280 KH/s without restarting for 2 hours now and hope will continue.

I tried to change the OS to SMOS but no luck.

@ nrg_wolf I was on 13.12 facing same issue,So i uninstalled 13.12 and installed the drivers which came with cd with no luck
now i have 14.2 beta installed and i see no diff with 13.12.I am assuming now 13.12 and 14.2 will not cause any issue

@ nrg_wolf i have tried tat settings but it restarts with in 15 min///its powercolour 2 gb version//using cgminer with arguments...

@Starscream i have removed vddc now and i can see voltage is set at 1.1,when i set to 800 and 1250 it restart very fast

Remove that gpu-vddc from the config and set your core to 800 and memclock to 1250.

If your PC still crashes than download and install MSI Afterburner (ver 3.xx - beta) and let it detect your clock speeds and voltage (you might need to restart after installing).

Stock voltage for the HD 6950 is 1.1v, so if MSI Afterburner reads anything else you might have a problem there (I assume you didn't flash any bios) and you would probably need to set it manually back to 1.1v and than change the clock speeds back to their stock settings (if you haven't already - 800/1250) and work from there.

change TC to 8000 at I19 for best results i have found, and have you checked the VRM temps while its mining? is the card reference design with a tunnel ram cooler on it or aftermarket with better cooling? also are you just useing a bat file with your settings or useing a config file?

I have 600 watts cooler master silent pro supply...i have running now at 297 kh/s for 1 hour now without any restart on 14.2 beta with below settings:

--gpu-engine 600 --gpu-memclock 800 --lookup-gap 2 --thread-concurrency 7000 -g 1 -w 256 --intensity 18 --temp-target 80 --auto-fan --gpu-vddc 1.030

it has flashed to 6970 but i switched back to the old one....both have same problem.

what is your powersupply rated at?

also go back to 13.12 drivers, current beta drivers are not ideal for mining. can you show me your config file, and have you flashed the bios to 6970 specs  or still useing the standard one, also reset all settings on the core and memory to defult and work them up via 5mhz at a time.

i have tried all combination and trying out from last 5 days...have used gpu power tune and also decreasing the mem clock  and memory....

--gpu-engine 900 --gpu-memclock 1250 --lookup-gap 2 --thread-concurrency 7000 -g 1 -w 256 --intensity 14 --temp-target 80 --auto-fan --gpu-vddc 1.030

trying added vddc and work size ..i have installed 14.2 beta drivers...still testing the settings..

but the problem still exist..and the time before it restarts increases with increase in hash

try my config, i get 500khs of HD 6950 with a WU: of about 450-460.
make sure that you update the bios to 6970 as that gives more play room on the core and memory.

],
"api-allow" : "W:127.0.0.1",
"api-listen" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "120",
"hotplug" : "5",
"kernel-path" : "/usr/local/bin",
"log" : "5",
"no-submit-stale" : true,
"queue" : "1",
"scan-time" : "30",
"scrypt" : true,
"shares" : "0",
"gpu-threads" : "1",
"gpu-dyninterval" : "7",
"gpu-engine" : "0-925",
"gpu-fan" : "0-85",
"gpu-platform" : "0",
"gpu-memclock" : "1425",
"gpu-memdiff" : "0",
"gpu-powertune" : "20",
"gpu-vddc" : "1.030",
"intensity" : "19",
"temp-target" : "85",
"temp-overheat" : "90",
"temp-cutoff" : "95",
"temp-hysteresis" : "3",
"vectors" : "1",
"worksize" : "256",
"lookup-gap" : "2",
"shaders" : "0",
"thread-concurrency" : "8000",
"no-pool-disable" : true
}

What are your temps?

As in the previous post, will more than likely be a combination of TC Mem and GPU clock ... Your TC seems like an odd figure to me, but I have 7950s.

Windows or linux?

Cheers!

neurocis
Author of https://github.com/neurocis/coinvert

Well, try this: Change clock speeds to --gpu-engine 800 --gpu-memclock 1250 and than remove all undervolting you may have done (i.e. return to stock settings).
If it's running fine than your next step would be to up the intensity and see if it's still stable.
After that point it's all fine tuning (playing with the TC value, undervolting, small adjustments to core/mem clocks, etc).


If I had to guess why your PC restarts it's because you got a mismatch between clock speeds and voltage.

I have 6950 HD radeon card.When mining i get maximum 300 kh/s being stable,,if i increase the hashing power the system restart frequently .I am using below settings:

**** 297KH/s stable average with 6950 ******
--gpu-engine 700 --gpu-memclock 900 --lookup-gap 2 --thread-concurrency 6000 -g 1 --intensity 14 --temp-target 65 --auto-fan

i know this card is capable of 450 KH/s but the problem is stability.Any suggestion??

I am ready to give 0.02 BTC to the person who helps me to solves this problem as i will be able to cover the profit .

*************Update 1*********************
I have 600 watts cooler master silent pro supply...i have running now at 297 kh/s for 1 hour now without any restart on 14.2 beta with below settings:

--gpu-engine 600 --gpu-memclock 800 --lookup-gap 2 --thread-concurrency 7000 -g 1 -w 256 --intensity 18 --temp-target 80 --auto-fan --gpu-vddc 1.030

it has flashed to 6970 but i switched back to the old one....both bios have same problem.


*************Update 2*********************
--gpu-engine 570 --gpu-memclock 800 --lookup-gap 2 --thread-concurrency 5632 -g 1 -w 256 --intensity 18 --temp-target 75 --auto-fan --shaders 1408

Running with above settings now and its running at 280 KH/s without restarting for 2 hours now and hope will continue.

I tried to change the OS to SMOS but no luck.