GPU load when pooling?

[carriun]

I just set up my new miner following the ubuntu guide.  I have 2 x 6970s. 

I'm running headless
I'm running poclbm.py like this:

./poclbm.py -d0 --host=btcguild.com --port=8332  -v -w 128 -r 10 > /tmp/gpu0&
./poclbm.py -d1 --host=btcguild.com --port=8332  -v -w 128 -r 10 > /tmp/gpu1&

(username & password removed).

I'm getting between 150M - 250M H/s on each gpu.  seems to be bouncing around quite a bit.

My GPUs are running about 50%, is that due to the pool?  Is there anything I should do to bump the GPU usage closer to 100%?

Adapter 0 - AMD Radeon HD 6900 Series
                            Core (MHz)    Memory (MHz)
           Current Clocks :    925           1375
             Current Peak :    925           1375
  Configurable Peak Range : [500-950]     [1375-1450]
                 GPU load :    38%

Adapter 1 - AMD Radeon HD 6900 Series
                            Core (MHz)    Memory (MHz)
           Current Clocks :    925           1375
             Current Peak :    925           1375
  Configurable Peak Range : [500-950]     [1375-1450]
                 GPU load :    59%

Adapter 0 - AMD Radeon HD 6900 Series
            Sensor 0: Temperature - 59.50 C

Adapter 1 - AMD Radeon HD 6900 Series
            Sensor 0: Temperature - 60.00 C


                                                                                   

[1715483500]

1715483500

[bcpokey]

I'm not familiar with linux so I can't help you from that standpoint but I will say it is not due to the pool. GPU load for most pool miners is 95-100%.

Try to make sure whatever the equivalent is to an "aggression" flag is set as high as possible.

[exahash]

What do your /tmp/gpu0 and /tmp/gpu1 logs say?  It could be your miner or your pool or your card(s).

I'm testing out btcguild right now (using phoenix) and keep getting disconnect and work queue empty messages.  That doesn't happen to me on deepbit or slush.  Lack of work could account for your gpu idle-ness.

You might get better performance by trying phoenix instead of poclbm. If it's a dedicated mining rig, try the parameters BFI_INT VECTORS AGGRESSION=11 FASTLOOP=false -k phatk   If it's not dedicated, lower the aggression to 5 (or whatever makes your UI responsive enough).

You want aticonfig to report 99% most of the time (I've never seen it show 100%).

[carriun]

I upgraded to SDK2.4,  I don't think that helped with mining but it fixed some aticonfig stuff.

I updated my poclbm command line

./poclbm.py -d0 --host=btcguild.com --port=8332  -v -w 256 -r 10 AGGRESSION=12 BFI_INT > /tmp/gpu0&
./poclbm.py -d1 --host=btcguild.com --port=8332 -v -w 256 -r 10 AGGRESSION=12 BFI_INT > /tmp/gpu1&

I'm now overclocked to 950Mhz and running 333MH/s/GPU @ 97% GPU

One 6970 was DOA so that will be swapped out on Monday, hopefully my power supply can handle all 3

Adapter 0 - AMD Radeon HD 6900 Series
                            Core (MHz)    Memory (MHz)
           Current Clocks :    950           1375
             Current Peak :    950           1375
  Configurable Peak Range : [500-950]     [1375-1450]
                 GPU load :    98%

Adapter 1 - AMD Radeon HD 6900 Series
                            Core (MHz)    Memory (MHz)
           Current Clocks :    950           1375
             Current Peak :    950           1375
  Configurable Peak Range : [500-950]     [1375-1450]
                 GPU load :    97%

Adapter 0 - AMD Radeon HD 6900 Series
            Sensor 0: Temperature - 69.00 C

Adapter 1 - AMD Radeon HD 6900 Series
            Sensor 0: Temperature - 67.00 C


still nice and cool @ 100% fan speed