Topic: 3 Card Rig submit 50% less shares than 2 card rig (Read 1102 times)
hi, I was mining terracoins. Now its all resolved as rig died on its ass last night. Mobo is dead so I have to build a new one. It was old mobo so I was not surprised. Many thanks for all the input guys. Im going to get a new mobo with 5 pcie and another 2 7950 on Friday and build a new miner.
What are you mining with a 23.5k difficulty? Litecoin?
Your 3 card box looks like it has a config problem. It'd be helpful if you posted the full configs from both boxes, not just one and say they're the same. They're not.
Your 3 card box looks like it has a config problem. It'd be helpful if you posted the full configs from both boxes, not just one and say they're the same. They're not.
Terracoin or Freicoin.
What are you mining with a 23.5k difficulty? Litecoin?
Your 3 card box looks like it has a config problem. It'd be helpful if you posted the full configs from both boxes, not just one and say they're the same. They're not.
Your 3 card box looks like it has a config problem. It'd be helpful if you posted the full configs from both boxes, not just one and say they're the same. They're not.
Back again with the same issue. Last night Ive started cgminer on both rigs at the same time and after about 12 hours I get below results. Looks like my 2 card rig submitted over 2000 shares more than 3 card rig. No idea how is that possible.
Block: f77f0b2538197bb6... Diff:23.8K Started: [10:26:12] Best share: 22.1K
------------------------------------------------------------------------------
[P]ool management [G]PU managementettings [D]isplay options [Q]uit
GPU 0: 74.0C 3300RPM | 535.2M/527.3Mh/s | A:2563 R:1 HW:0 U: 3.71/m I: 9
GPU 1: 78.0C 3233RPM | 534.0M/526.8Mh/s | A:2504 R:1 HW:0 U: 3.63/m I: 9
GPU 2: 75.0C 3225RPM | 605.5M/605.4Mh/s | A:2936 R:5 HW:0 U: 4.26/m I: 9
------------------------------------------------------------------------------
Thread 3: 318.0 Mh/s Enabled ALIVE
Block: f77f0b2538197bb6... Diff:23.8K Started: [10:26:12] Best share: 24.3K
-------------------------------------------------------------------------------
[P]ool management [G]PU managementettings [D]isplay options [Q]uit
GPU 0: 71.0C 3123RPM | 551.5M/548.7Mh/s | A:5059 R:4 HW:0 U: 7.33/m I: 9
GPU 1: 76.0C 3271RPM | 551.2M/548.8Mh/s | A:5194 R:6 HW:0 U: 7.53/m I: 9
-------------------------------------------------------------------------------
------------------------------------------------------------------------------
[P]ool management [G]PU management
GPU 0: 74.0C 3300RPM | 535.2M/527.3Mh/s | A:2563 R:1 HW:0 U: 3.71/m I: 9
GPU 1: 78.0C 3233RPM | 534.0M/526.8Mh/s | A:2504 R:1 HW:0 U: 3.63/m I: 9
GPU 2: 75.0C 3225RPM | 605.5M/605.4Mh/s | A:2936 R:5 HW:0 U: 4.26/m I: 9
------------------------------------------------------------------------------
Thread 3: 318.0 Mh/s Enabled ALIVE
Block: f77f0b2538197bb6... Diff:23.8K Started: [10:26:12] Best share: 24.3K
-------------------------------------------------------------------------------
[P]ool management [G]PU management
GPU 0: 71.0C 3123RPM | 551.5M/548.7Mh/s | A:5059 R:4 HW:0 U: 7.33/m I: 9
GPU 1: 76.0C 3271RPM | 551.2M/548.8Mh/s | A:5194 R:6 HW:0 U: 7.53/m I: 9
-------------------------------------------------------------------------------
Issue resolved!!! 
I have dropped gpu clock slightly to 1075 and now all 5 cards seems to have similar submission rate and speed. I used to run all cards at 1200mhz with no problems at all so i think one of my psu's is burning out. many thanks for help
I have dropped gpu clock slightly to 1075 and now all 5 cards seems to have similar submission rate and speed. I used to run all cards at 1200mhz with no problems at all so i think one of my psu's is burning out. many thanks for helpMaybe it's time to reapply thermal paste?
Issue resolved!!! I have dropped gpu clock slightly to 1075 and now all 5 cards seems to have similar submission rate and speed. I used to run all cards at 1200mhz with no problems at all so i think one of my psu's is burning out. many thanks for help
I have dropped gpu clock slightly to 1075 and now all 5 cards seems to have similar submission rate and speed. I used to run all cards at 1200mhz with no problems at all so i think one of my psu's is burning out. many thanks for help 
Make a copy of your config and try this... or compare the parameters...
I'm not sure how the gpu-engine and auto-gpu settings conflict in your config... but at least give it a low floor limit on the autotuning.
I also prefer the memdiff statement rather than memclock. Seems more intuitive and if it tunes the engine it'll tune the memory along with it.
I also prefer using diablo, it's faster I think on these GPUs than poclbm. Play with the worksize too, 256 seems to work best for me but I'm on a 7870xt (same processor, less shaders).
"intensity" : "9,9,9",
"vectors" : "1,1,1",
"worksize" : "256,256,256",
"kernel" : "diablo,diablo,diablo",
"gpu-engine" : "925-1100,925-1100,925-1100",
"gpu-memdiff" : "-50",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "0",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "55",
"temp-hysteresis" : "4",
"shares" : "0",
"no-submit-stale" : true,
"kernel-path" : "/usr/local/bin"
I'm not sure how the gpu-engine and auto-gpu settings conflict in your config... but at least give it a low floor limit on the autotuning.
I also prefer the memdiff statement rather than memclock. Seems more intuitive and if it tunes the engine it'll tune the memory along with it.
I also prefer using diablo, it's faster I think on these GPUs than poclbm. Play with the worksize too, 256 seems to work best for me but I'm on a 7870xt (same processor, less shaders).
"intensity" : "9,9,9",
"vectors" : "1,1,1",
"worksize" : "256,256,256",
"kernel" : "diablo,diablo,diablo",
"gpu-engine" : "925-1100,925-1100,925-1100",
"gpu-memdiff" : "-50",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "0",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "55",
"temp-hysteresis" : "4",
"shares" : "0",
"no-submit-stale" : true,
"kernel-path" : "/usr/local/bin"
auto-gpu needs a range, so he should have his range just like you have yours, instead of specific gpu setting. The whole point is that will adjust the gpu setting, along with the fan setting, to within the range you set in order to hit and keep your target temp as best it can.
Make a copy of your config and try this... or compare the parameters...
I'm not sure how the gpu-engine and auto-gpu settings conflict in your config... but at least give it a low floor limit on the autotuning.
I also prefer the memdiff statement rather than memclock. Seems more intuitive and if it tunes the engine it'll tune the memory along with it.
I also prefer using diablo, it's faster I think on these GPUs than poclbm. Play with the worksize too, 256 seems to work best for me but I'm on a 7870xt (same processor, less shaders).
"intensity" : "9,9,9",
"vectors" : "1,1,1",
"worksize" : "256,256,256",
"kernel" : "diablo,diablo,diablo",
"gpu-engine" : "925-1100,925-1100,925-1100",
"gpu-memdiff" : "-50",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "0",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "55",
"temp-hysteresis" : "4",
"shares" : "0",
"no-submit-stale" : true,
"kernel-path" : "/usr/local/bin"
I'm not sure how the gpu-engine and auto-gpu settings conflict in your config... but at least give it a low floor limit on the autotuning.
I also prefer the memdiff statement rather than memclock. Seems more intuitive and if it tunes the engine it'll tune the memory along with it.
I also prefer using diablo, it's faster I think on these GPUs than poclbm. Play with the worksize too, 256 seems to work best for me but I'm on a 7870xt (same processor, less shaders).
"intensity" : "9,9,9",
"vectors" : "1,1,1",
"worksize" : "256,256,256",
"kernel" : "diablo,diablo,diablo",
"gpu-engine" : "925-1100,925-1100,925-1100",
"gpu-memdiff" : "-50",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "0",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "55",
"temp-hysteresis" : "4",
"shares" : "0",
"no-submit-stale" : true,
"kernel-path" : "/usr/local/bin"
Check how your pool allocates share difficulty, its likely automatically varied based on your hashing speed. Your 3 card rig is probably working at a higher share difficulty, meaning it takes longer to submit a share, but the share is worth more when accepted than a share from a lower difficulty level.
Could be nothing to worry about.
Could be nothing to worry about.
Im running cgminer 3.2.1 and windows 7 32bit on both rigs. I have started cgminer on both rigs at the same time and after few hours checked the it and found that 2 card rig submitted just over 4600 shares and 3 card rig only just over 2800. I don't get much stale or hw errors but I have noticed that on 2 card rig utility is at 7-8 s/m on each card but on 3 card rig utility is 3-4 s/m per card. both configs are the same. All cards on 3 card rig submit similar shares numbers but a lot less than cards on 2 card rig.
"intensity" : "9,9,9",
"gpu-engine" : "1100,1100,1100",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "1050,1050,1050",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "60",
"temp-hysteresis" : "3",
"kernel-path" : "/usr/local/bin"
}
"intensity" : "9,9,9",
"gpu-engine" : "1100,1100,1100",
"gpu-fan" : "0-100,0-100,0-100",
"gpu-memclock" : "1050,1050,1050",
"gpu-powertune" : "20,20,20",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"scan-time" : "60",
"temp-hysteresis" : "3",
"kernel-path" : "/usr/local/bin"
}
Can you post your configs from both machines? What miner software are you using? What OS?
If you're using CGMiner, you should be able to find out which card/s are slower in shares. How long do you wait until you make the determination one is slower than the other? most of the stats take some time (serveral hours to a day or two) to smooth out and become more accurate.
If you're using CGMiner, you should be able to find out which card/s are slower in shares. How long do you wait until you make the determination one is slower than the other? most of the stats take some time (serveral hours to a day or two) to smooth out and become more accurate.
Hi guys,
I have a small problem. I have 2 rigs with 2x and 3x 7950. My rig with 2 cards submit almost 50% more shares than my 3 card rig. How is that possible? All cards are pretty much the same configuration hashing at around 550mh/s at 67-72C. i have the same drivers 13.4 on both rigs and same OS version. Any idea what may be causing that?
thanks in advance.
I have a small problem. I have 2 rigs with 2x and 3x 7950. My rig with 2 cards submit almost 50% more shares than my 3 card rig. How is that possible? All cards are pretty much the same configuration hashing at around 550mh/s at 67-72C. i have the same drivers 13.4 on both rigs and same OS version. Any idea what may be causing that?
thanks in advance.
Jump to: