Anyone getting this: [CGMiner] Error -6 (Command Queue) on (2) 7970?

[Dalkore]

Has anyone gotten - Error -6 (trying to establish Command Queue), what happens it that is drops GPU1 and it says DEAD under the mega-hash performance and I only get output from GPU0.   Even it I try, restart GPU, it doesn't work.  I have to quit and then open CGminer again and that gets it back up.  It seems to happen every 36 hours or so.

[Dalkore]

Yeah, I did down clock quite a bit.   I'll research and try and other people settings that have claim stability.     Regardless I just remote access my box and reboot CG.    Thanks

[-ck]

Has anyone gotten - Error -6 (trying to establish Command Queue), what happens it that is drops GPU1 and it says DEAD under the mega-hash performance and I only get output from GPU0.   Even it I try, restart GPU, it doesn't work.  I have to quit and then open CGminer again and that gets it back up.  It seems to happen every 36 hours or so.

#define CL_OUT_OF_HOST_MEMORY                       -6

Looks like you're overloading the memory bandwidth for that hashrate.

[bitlane]

I saw this error when downclocking memory too far..

How far down is too far ?

I am running 300Mhz Mem Clocks on 15 7970s and have not had a problem...
A small performance loss, but the power (and heat) savings more than makes up for it.

CGMiner 2.4.2 (for GPU Clock control and mining, of course), using Afterburner for MEM Clocks and Voltage Control.

Code:

GPU 0:	67.0C	3260 RPM 55% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 564.9 / 565.4 Mh/s | 99% Load  | A:3077 R:27 HW:0 U:7.87/m I: 9
GPU 1:	67.0C	3300 RPM 55% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 567.8 / 568.4 Mh/s | 99% Load  | A:3021 R:19 HW:0 U:7.72/m I: 9
GPU 2:	68.0C	2905 RPM 53% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.7 / 566.3 Mh/s | 99% Load  | A:3124 R:32 HW:0 U:7.99/m I: 9
GPU 3:	67.0C	3118 RPM 49% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 567.9 / 568.4 Mh/s | 99% Load  | A:3022 R:29 HW:0 U:7.73/m I: 9
GPU 4:	67.0C	3345 RPM 56% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 568.0 / 568.5 Mh/s | 99% Load  | A:3174 R:20 HW:0 U:8.12/m I: 9
----------------------------------------------------------------------------
GPU 0:	67.0C	3542 RPM 58% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 566.0 / 566.3 Mh/s | 99% Load  | A:3144 R:23 HW:0 U:8.03/m I: 9
GPU 1:	67.0C	3158 RPM 55% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 566.5 / 567.0 Mh/s | 99% Load  | A:3109 R:25 HW:0 U:7.94/m I: 9
GPU 2:	67.0C	3298 RPM 54% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 569.0 / 569.2 Mh/s | 99% Load  | A:3096 R:19 HW:0 U:7.91/m I: 9
GPU 3:	67.0C	3483 RPM 57% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 568.9 / 569.2 Mh/s | 99% Load  | A:3066 R:18 HW:0 U:7.83/m I: 9
GPU 4:	67.0C	3511 RPM 58% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 568.5 / 569.2 Mh/s | 99% Load  | A:3109 R:20 HW:0 U:7.94/m I: 9
----------------------------------------------------------------------------
GPU 0:	67.0C	3180 RPM 51% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.5 / 566.2 Mh/s | 99% Load  | A:3077 R:17 HW:0 U:7.87/m I: 9
GPU 1:	67.0C	3219 RPM 55% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.3 / 565.7 Mh/s | 99% Load  | A:3136 R:21 HW:0 U:8.03/m I: 9
GPU 2:	68.0C	3351 RPM 56% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.8 / 566.2 Mh/s | 99% Load  | A:3090 R:16 HW:0 U:7.91/m I: 9
GPU 3:	67.0C	3433 RPM 57% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.7 / 566.2 Mh/s | 99% Load  | A:3112 R:21 HW:0 U:7.96/m I: 9
GPU 4:	67.0C	3185 RPM 52% Fan | 960 Mhz GPU	1.02 V  300 Mhz MEM | 565.4 / 566.2 Mh/s | 99% Load  | A:3011 R:20 HW:0 U:7.71/m I: 9

This is simply a copy/paste from my monitoring app, with proper voltage levels added for reference.

Today (when I took this copy/paste selection above) has been the hottest day this year, for my miners...I am sweating, just standing in front of them yet they are hardly breaking a sweat.....

[-ck]

I saw this error when downclocking memory too far..

How far down is too far ?

I am running 300Mhz Mem Clocks on 15 7970s and have not had a problem...
A small performance loss, but the power (and heat) savings more than makes up for it.

Difference is you're running stock engine clock speeds which he may not be. At higher engine speeds voltage will be dynamically adjusted within the card and perhaps the power balance is shifted without enough memory speed. They're not really "tuned" internally to run these weird combinations of low voltage, high engine speed and low memory speeds.

[bitlane]

Moved here....
https://bitcointalk.org/index.php?topic=87504.0