Hi,

    Testing the 19.2-1 beta image - coin set to RVN - In 1bash line 460 mentions SUPRminer, but suprminers not in miners update script.   I cloned/built suprminer into the miners directory, the directory is suprminer all lowecase, the comment has SUPRminer.  Changed it to match the directory name and it works fine.  Can you add it to the miners update script & fix the comment.

I want to say thanks to pap and LuKe for the awesome re-write.

Don

Hi Pap,

I saw the MSFT ccminer for RVN in the miners update, I tried the code and it segmentation faults pretty quickly - I am running suprminer(no fee), it runs without issue - Tried it on two different frames, its been stable.

https://github.com/ocminer/suprminer.git

./suprminer/build.sh

Don




Test frame.   

ID,VENDOR,MODEL,PSTATE,TEMP,FAN,UTILIZATION,POWER,POWERLIMIT,MAXPOWER,GPUCLOCK,MEMCLOCK
--------------------------------------------------------------------------------
0, GIGABYTE, P106-100, P0, 49, 50, 100, 67.39, 85.00, 140.00, 1493, 3905
1, ASUS, P106-100, P0, 54, 50, 100, 66.95, 85.00, 140.00, 1594, 3905
2, ASUS, P106-100, P0, 41, 50, 99, 55.56, 85.00, 140.00, 1493, 3905
3, ASUS, P106-100, P0, 54, 50, 100, 51.85, 85.00, 140.00, 1657, 3905
4, ASUS, P106-100, P0, 56, 50, 99, 96.83, 85.00, 140.00, 1721, 3905
5, ASUS, P106-100, P0, 54, 50, 100, 57.54, 85.00, 140.00, 1493, 3905
6, ASUS, P106-100, P0, 57, 50, 100, 69.79, 85.00, 140.00, 1620, 3905
7, ASUS, P106-100, P0, 55, 50, 99, 61.98, 85.00, 140.00, 1493, 3905
8, MSI, P106-100, P0, 52, 50, 100, 51.64, 85.00, 140.00, 1468, 3905
9, MSI, P106-100, P0, 51, 50, 100, 54.11, 85.00, 140.00, 1493, 3905
10, MSI, P106-100, P0, 49, 50, 100, 52.74, 85.00, 140.00, 1468, 3905

Remember to change your coin to RVN in 1bash -

This works on mine - switched a frame from ETH to RVN - compiled suprminer following papa's instructions for adding a new coin in 19.2-1

1bash

0miner

Hi,


This might be a recurring topic but I have some 1060 and 1070's, all in P2 mode.  Getting these to run with mixed cards that are in P0 like P106 seems more difficult than separating them into two systems, it feels like some type of a race condition.  So, I have separated the p2 cards into different systems them but lose 1-2MH on the 1070 running in P2.  I can run on Windows and use profile inspector to clear a driver flag called CudaForceP2Mode to off.  Does anyone know how to pass this parameter in Nvidia-settings or put binary name/value pair into their xorg.conf?


Pap, any thoughts on who to talk with?



Thanks

Hi papampi,

This issue was with a mix of cards like the P106 and 1070 in a frame setting the OC speeds requires that the nvidia command to set the GPU or memory OC needs to be the [2] for some and [3] for others.  The loop looks like it was designed to try both [2] [3] if NORMAL=YES above, for each gpu but the gpu variable was incremented in the inner loop so you would get [2] on gpu[0] then [3] on gpu[1],  [2] on gpu [2] and so on. By moving the gpu increment out of the for loop it and into the outer while loop I believe it works as intended.  I added the two echo's in here to show the behavior. 

Original loop:


TI is 2 3, J is 2 and GPU is 0
nvidia-settings -a [gpu:0]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:0]/GPUMemoryTransferRateOffset[2]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:0]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:0]) assigned value 1500.

TI is 2 3, J is 3 and GPU is 1
nvidia-settings -a [gpu:1]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:1]/GPUMemoryTransferRateOffset[3]=1300




TI is 2 3, J is 2 and GPU is 2
nvidia-settings -a [gpu:2]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:2]/GPUMemoryTransferRateOffset[2]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:2]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:2]) assigned value 1500.

TI is 2 3, J is 3 and GPU is 3
nvidia-settings -a [gpu:3]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:3]/GPUMemoryTransferRateOffset[3]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:3]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:3]) assigned value 1500.

TI is 2 3, J is 2 and GPU is 4
nvidia-settings -a [gpu:4]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:4]/GPUMemoryTransferRateOffset[2]=1300

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:4]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:4]) assigned value 1300.

TI is 2 3, J is 3 and GPU is 5
nvidia-settings -a [gpu:5]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:5]/GPUMemoryTransferRateOffset[3]=1300


===============================================================
By moving the gpu increment to the while loop the execution looks like:


TI is 2 3, J is 2 and GPU is 0
nvidia-settings -a [gpu:0]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:0]/GPUMemoryTransferRateOffset[2]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:0]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:0]) assigned value 1500.

TI is 2 3, J is 3 and GPU is 0
nvidia-settings -a [gpu:0]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:0]/GPUMemoryTransferRateOffset[3]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:0]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:0]) assigned value 1500.

TI is 2 3, J is 2 and GPU is 1
nvidia-settings -a [gpu:1]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:1]/GPUMemoryTransferRateOffset[2]=1300

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:1]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:1]) assigned value 1300.

TI is 2 3, J is 3 and GPU is 1
nvidia-settings -a [gpu:1]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:1]/GPUMemoryTransferRateOffset[3]=1300




TI is 2 3, J is 2 and GPU is 2
nvidia-settings -a [gpu:2]/GPUGraphicsClockOffset[2]=-200
nvidia-settings -a [gpu:2]/GPUMemoryTransferRateOffset[2]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:2]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:2]) assigned value 1500.

TI is 2 3, J is 3 and GPU is 2
nvidia-settings -a [gpu:2]/GPUGraphicsClockOffset[3]=-200
nvidia-settings -a [gpu:2]/GPUMemoryTransferRateOffset[3]=1500

  Attribute 'GPUGraphicsClockOffset' (19_1_4:0[gpu:2]) assigned value -200.


  Attribute 'GPUMemoryTransferRateOffset' (19_1_4:0[gpu:2]) assigned value 1500.







if [ $P106_100_FULL_HEADLESS_MODE == "NO" ]
then

gpu=0
while [ $gpu -lt $GPUS ]
do
for j in $TI
do
CORE=${__CORE_OVERCLOCK[${gpu}]}
MEM=${MEMORY_OVERCLOCK[${gpu}]}
echo "TI is $TI, J is $j and GPU is $gpu"
echo "${NVD} -a [gpu:$gpu]/GPUGraphicsClockOffset[${j}]=$CORE"
echo "${NVD} -a [gpu:$gpu]/GPUMemoryTransferRateOffset[${j}]=$MEM"
${NVD} -a [gpu:$gpu]/GPUGraphicsClockOffset[${j}]=$CORE
${NVD} -a [gpu:$gpu]/GPUMemoryTransferRateOffset[${j}]=$MEM
done
gpu=$(($gpu+1))
done

Hi,

I was looking at the maxximus007_auto_temperature_control script and I think you may want to make the following change to allow for 3 digit power levels, it was truncating a board at 110w to 10 and trying to reset it to 110 when it was already there.

Don


POWERLIMIT=$PWRLIMIT
POWERLIMIT=$(echo -n $PWRLIMIT | tail -c -6 | head -c -3 ) # changed tail -c -5 to -c -6

Hi,

Hi All,

I absolutely love nvOC.


I was looking at the loop code for setting individual clock limits and I may have found a bug, I have a mix of 1070's and P106 boards in this mining frame. 

When I looked at the code in 3main, the gpu=$(($gpu+1)) was up one line in the for loop, if I am reading the for loop correctly, you would want that to iterate from 2 to 3 for each GPU, do I moved it down one line as seen below:

gpu=0
while [ $gpu -lt $GPUS ]
do
for j in $TI
do
CORE=${__CORE_OVERCLOCK[${gpu}]}
MEM=${MEMORY_OVERCLOCK[${gpu}]}
${NVD} -a [gpu:$gpu]/GPUGraphicsClockOffset[${j}]=$CORE
${NVD} -a [gpu:$gpu]/GPUMemoryTransferRateOffset[${j}]=$MEM
done
gpu=$(($gpu+1))
done

Really like the telegraph functions in this release - setting time to a couple of intervals daily after getting a mining frame stable is a nice production feature.

Don