Hi,
I switched from GUIminer to cgminer on my win7 based rig two days ago. I have two issues though that I can't seem to find the answers using the board search or google...
- The guiminer output does not show GPU temps and fanspeeds. I watch it via RDP. I noticed that I saw temps & fanspeeds once when initially configuring the box having a display connected to one of the cards. This is not a cgminer specific problem, as I have the same issue with other programs like GPUz. No clock, temp, fanspeed displayed.
- Tweaking GPU/mem clocks does not make any difference. Is this due to having CCC running in the background? Might that be related to my first problem as well?
thx
Manfred
Topic: OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.0 - page 376. (Read 5805537 times)
April 12, 2013, 10:53:35 AM
April 12, 2013, 10:15:25 AM
You didn't read linux-usb-cgminer ...
April 12, 2013, 08:56:46 AM
Got my first Debian Linux (squeeze) based rig online and working, so I have a completely Microsoft-free system now. :)
Materials I found helpful:
- the starting post of this thread
- this post: https://bitcointalksearch.org/topic/m.47174
The hardest part for me personally, as it was mostly undocumented, was understanding that I need to start "cgminer" as a regular user (not root) and issue the commands from a local X session (not SSH).
Before that, I was vainly trying to understand why the following error occurred:
As soon as I changed to user instead of root, and local console instead of SSH, all the things which should have, started working.
Materials I found helpful:
- the starting post of this thread
- this post: https://bitcointalksearch.org/topic/m.47174
The hardest part for me personally, as it was mostly undocumented, was understanding that I need to start "cgminer" as a regular user (not root) and issue the commands from a local X session (not SSH).
Before that, I was vainly trying to understand why the following error occurred:
Quote
Error -1: Getting Device IDs (num)
clDevicesNum returned error, no GPUs usable
All devices disabled, cannot mine!
clDevicesNum returned error, no GPUs usable
All devices disabled, cannot mine!
As soon as I changed to user instead of root, and local console instead of SSH, all the things which should have, started working.
April 12, 2013, 07:42:23 AM
Hi Everyone
I have tried quite a few settings, but I cannot get cgminer to stop dropping the clock speed on one of the cores on my ATI 5970 dual GPU card.
[...]
I have tried quite a few settings, but I cannot get cgminer to stop dropping the clock speed on one of the cores on my ATI 5970 dual GPU card.
[...]
From what you describe later on it's probably the driver doing it, not cgminer. If you see that, you better start clocking your core down: if the driver is protecting the card by clocking its core down automatically it means you are slowly damaging the hardware: you'll fry the VRM eventually.
April 12, 2013, 02:23:56 AM
Here's an odd one I'm running into today after re-burning my USB stick's linux image from a clean copy, pulling in the latest git, compiling and firing up cgminer...
cgminer's interface ignores all normal menu keystrokes (p, g, and so on), UNLESS I neglect to specify a pool in the config file, in which case it works normally because it had to accept keyboard input to allow me to set the pool credentials. I can use ctl-c to kill it and get back to a prompt, no problem. I just can't adjust settings while cgminer is running.
This is latest git pull as of "Thu Apr 11 13:31:17 2013 +1000".
cgminer's interface ignores all normal menu keystrokes (p, g, and so on), UNLESS I neglect to specify a pool in the config file, in which case it works normally because it had to accept keyboard input to allow me to set the pool credentials. I can use ctl-c to kill it and get back to a prompt, no problem. I just can't adjust settings while cgminer is running.
This is latest git pull as of "Thu Apr 11 13:31:17 2013 +1000".
April 11, 2013, 07:58:14 PM
Changing SDKs there is no need for a recompile, but you must delete any .bin files generated so they're regenerated. Note there is no clean way to uninstall though so be careful when removing the old one to make sure it really has gone. AMD seems to have taken its ideas for uninstall from windows where you just say it is uninstalled but leave half the files there.
Ok awesome, thank you for your help. I'll report back with how it goes.
April 11, 2013, 07:55:19 PM
Changing SDKs there is no need for a recompile, but you must delete any .bin files generated so they're regenerated. Note there is no clean way to uninstall though so be careful when removing the old one to make sure it really has gone. AMD seems to have taken its ideas for uninstall from windows where you just say it is uninstalled but leave half the files there.
April 11, 2013, 07:52:54 PM
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Any ideas? To update I changed TC to --shaders 1792 (from memory, whatever the 7950 shaders are) and it worked - but was only hashing at about 8 kh/s a card. I'm more than happy to try anything suggested. Cheers.
Ok. I've read the README again. I mean this respectfully because I know how frustrating it would be when seemingly someone is asking questions that are answered previously or in the README but as far as I can tell my specific problem isn't addressed. I've spent about 4-5 hours trying to get it to work, configuring with multiple APP SDK and Catalyst versions, read through the README and SCRYPT-README multiple times, but still can't seem to solve this issue.
I'm usuing Xubuntu 12.10, Catalyst 13.1, AMD ADL SDK 5.0, AMD APP SDK 2.8. I will provide the full debug of the mining when I get home (at work). I have 8GB of ram as I know LTC mining is RAM intensive.
Ok, so going off that you're saying 2.8 is crap, roll back to 2.7 (given I need 2.6+ for 7xxx cards) and recompile cgminer? As I mentioned with the Xubuntu guide I did try 2.7 before, but I was having issues then with the 13.1 drivers. I'll try it again since I've got the 13.1 working now. Thanks for at least pointing me in the right direction.
April 11, 2013, 07:48:25 PM
Hi Everyone
I have tried quite a few settings, but I cannot get cgminer to stop dropping the clock speed on one of the cores on my ATI 5970 dual GPU card.
Here is my startup string: (I know its scary but temps don't go over 80 degrees)
cgminer -o http://pool.com -u username -p password -I 9 --gpu-engine 930,930 --gpu-vddc 1.2 --gpu-memclock 300,300 --temp-hysteresis 10 --gpu-fan 85 --temp-target 100,100 --temp-cutoff 105,105
With these settings, cgminer seems to keep dropping the first core from 930 to 550 every few seconds for a second or two. From what I can tell, it is trying to maintain a temperature, but why when I have disabled all auto watch stuff from within the application?
Can anyone help or give suggestions? Thanks
I have tried quite a few settings, but I cannot get cgminer to stop dropping the clock speed on one of the cores on my ATI 5970 dual GPU card.
Here is my startup string: (I know its scary but temps don't go over 80 degrees)
cgminer -o http://pool.com -u username -p password -I 9 --gpu-engine 930,930 --gpu-vddc 1.2 --gpu-memclock 300,300 --temp-hysteresis 10 --gpu-fan 85 --temp-target 100,100 --temp-cutoff 105,105
With these settings, cgminer seems to keep dropping the first core from 930 to 550 every few seconds for a second or two. From what I can tell, it is trying to maintain a temperature, but why when I have disabled all auto watch stuff from within the application?
Can anyone help or give suggestions? Thanks
April 11, 2013, 07:46:07 PM
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Any ideas? To update I changed TC to --shaders 1792 (from memory, whatever the 7950 shaders are) and it worked - but was only hashing at about 8 kh/s a card. I'm more than happy to try anything suggested. Cheers.
Ok. I've read the README again. I mean this respectfully because I know how frustrating it would be when seemingly someone is asking questions that are answered previously or in the README but as far as I can tell my specific problem isn't addressed. I've spent about 4-5 hours trying to get it to work, configuring with multiple APP SDK and Catalyst versions, read through the README and SCRYPT-README multiple times, but still can't seem to solve this issue.
I'm usuing Xubuntu 12.10, Catalyst 13.1, AMD ADL SDK 5.0, AMD APP SDK 2.8. I will provide the full debug of the mining when I get home (at work). I have 8GB of ram as I know LTC mining is RAM intensive.
April 11, 2013, 07:40:51 PM
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Any ideas? To update I changed TC to --shaders 1792 (from memory, whatever the 7950 shaders are) and it worked - but was only hashing at about 8 kh/s a card. I'm more than happy to try anything suggested. Cheers.
Ok. I've read the README again. I mean this respectfully because I know how frustrating it would be when seemingly someone is asking questions that are answered previously or in the README but as far as I can tell my specific problem isn't addressed. I've spent about 4-5 hours trying to get it to work, configuring with multiple APP SDK and Catalyst versions, read through the README and SCRYPT-README multiple times, but still can't seem to solve this issue.
Yes, I could still be missing something, and if I am I'll look like an idiot, but I'm just saying it's not like I haven't been trying before pestering the forums and asking for help. I'm also trying to be useful and provide diagnostic information (hence the output of ./cgminer -n), and if I like I can run it further in debug mode with -T and -D flags to help out. I've also tried to make sure I'm using the right export settings before I run the cgminer (GPU_MAX_ALLOC_PERCENT, etc).
I've made sure I've run "sudo aticonfig --adapter=all -f --initial", and also checked that it can detect my cards and temps with "sudo aticonfig --adapter=all -odgt" which is successful. I've tried following this guide here - https://docs.google.com/document/d/1Gw7YPYgMgNNU42skibULbJJUx_suP_CpjSEdSi8_z9U/preview?pli=1&sle=true with no success (errors I'm talking about now).
I've rebuilt cgminer lots of times, ensure the configuration file is ok, including the ADL SDK and enabling scrypt as per the README. I've tried mirroring settings in the hardware wiki (https://github.com/litecoin-project/litecoin/wiki/Mining-hardware-comparison) and I'm still getting poor hashrates.
I'm usuing Xubuntu 12.10, Catalyst 13.1, AMD ADL SDK 5.0, AMD APP SDK 2.8. I will provide the full debug of the mining when I get home (at work). I have 8GB of ram as I know LTC mining is RAM intensive.
I've tried multiple intensities (I know -I 13 is popular here), workrates, lookup gaps, GPU threads and shaders. I've gone through the 7970 tuning example.
Please, if you can give some actual advice to try I'd be really appreciative. Again I know it seems frustrating when someone comes in and possibly hasn't looked at the basics, but as you can see I've tried quite a few things and spent a lot of time on it. I'm not pretending I haven't missed something basic, I could definitely be at fault, but just point it out to me please instead of asking me to read the README again.
April 11, 2013, 07:02:18 PM
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Any ideas? To update I changed TC to --shaders 1792 (from memory, whatever the 7950 shaders are) and it worked - but was only hashing at about 8 kh/s a card. I'm more than happy to try anything suggested. Cheers.
April 11, 2013, 07:00:23 PM
Having an issue, if anyone could help it would be greatly appreciated.
Output of ./cgminer -n
What I'm running to use cgminer...
Output
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Output of ./cgminer -n
Code:
root@miner-desktop:/home/miner/Downloads/cgminer-2.11.4# ./cgminer -n
[2013-04-11 22:47:53] CL Platform 0 vendor: Advanced Micro Devices, Inc.
[2013-04-11 22:47:53] CL Platform 0 name: AMD Accelerated Parallel Processing
[2013-04-11 22:47:53] CL Platform 0 version: OpenCL 1.2 AMD-APP (1016.4)
[2013-04-11 22:47:53] Platform 0 devices: 3
[2013-04-11 22:47:53] 0 Tahiti
[2013-04-11 22:47:53] 1 Tahiti
[2013-04-11 22:47:53] 2 Tahiti
[2013-04-11 22:47:53] GPU 0 AMD Radeon HD 7900 Series hardware monitoring enabled
[2013-04-11 22:47:53] GPU 1 AMD Radeon HD 7900 Series hardware monitoring enabled
[2013-04-11 22:47:53] GPU 2 AMD Radeon HD 7900 Series hardware monitoring enabled
[2013-04-11 22:47:53] 3 GPU devices max detected
What I'm running to use cgminer...
Code:
export DISPLAY=:0
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PERCENT=100
./cgminer --scrypt -o stratum+tcp://stratum.give-me-ltc.com:3333 -u-p -I 15 --thread-concurrency 8192 -g 1 -w 256
Output
Code:
[2013-04-11 22:59:15] Started cgminer 2.11.4
[2013-04-11 22:59:15] Started cgminer 2.11.4
[2013-04-11 22:59:16] Probing for an alive pool
[2013-04-11 22:59:26] Maximum buffer memory device 0 supports says 95420416
[2013-04-11 22:59:26] Your scrypt settings come to 1310720000
[2013-04-11 22:59:26] Error -61: clCreateBuffer (padbuffer8), decrease TC or in
crease LG
[2013-04-11 22:59:26] Failed to init GPU thread 0, disabling device 0
[2013-04-11 22:59:26] Restarting the GPU from the menu will not fix this.
[2013-04-11 22:59:26] Try restarting cgminer.
Press enter to continue:
As far as I'm aware I should be able to go up to ~20000 TC, what's up?
Any help greatly appreciated.
Any ideas? To update I changed TC to --shaders 1792 (from memory, whatever the 7950 shaders are) and it worked - but was only hashing at about 8 kh/s a card. I'm more than happy to try anything suggested. Cheers.
April 11, 2013, 06:47:32 PM
Hier noch mal mein Output mit cgminer.exe -n:
Can someone guess whats the problem? I described what i installed and tried here already: https://bitcointalksearch.org/topic/m.1799428
I checked the faq and so on but cant find the problem.
Quote
C:\Users\Manuel\Desktop\CGMiner>cgminer.exe -n
[2013-04-12 00:36:42] CL Platform 0 vendor: Advanced Micro Devices, Inc.
[2013-04-12 00:36:42] CL Platform 0 name: AMD Accelerated Parallel Processing
[2013-04-12 00:36:42] CL Platform 0 version: OpenCL 1.2 AMD-APP (1124.2)
[2013-04-12 00:36:42] Platform 0 devices: 1
[2013-04-12 00:36:42] 0 Pitcairn
[2013-04-12 00:36:42] CL Platform 1 vendor: Intel(R) Corporation
[2013-04-12 00:36:42] CL Platform 1 name: Intel(R) OpenCL
[2013-04-12 00:36:42] CL Platform 1 version: OpenCL 1.1
[2013-04-12 00:36:42] Error -1: Getting Device IDs (num)
[2013-04-12 00:36:42] GPU 0 AMD Radeon HD 7970M hardware monitoring enabled
[2013-04-12 00:36:42] 1 GPU devices max detected
[2013-04-12 00:36:42] USB all: found 6 devices - listing known devices
[2013-04-12 00:36:43] No known USB devices
[2013-04-12 00:36:42] CL Platform 0 vendor: Advanced Micro Devices, Inc.
[2013-04-12 00:36:42] CL Platform 0 name: AMD Accelerated Parallel Processing
[2013-04-12 00:36:42] CL Platform 0 version: OpenCL 1.2 AMD-APP (1124.2)
[2013-04-12 00:36:42] Platform 0 devices: 1
[2013-04-12 00:36:42] 0 Pitcairn
[2013-04-12 00:36:42] CL Platform 1 vendor: Intel(R) Corporation
[2013-04-12 00:36:42] CL Platform 1 name: Intel(R) OpenCL
[2013-04-12 00:36:42] CL Platform 1 version: OpenCL 1.1
[2013-04-12 00:36:42] Error -1: Getting Device IDs (num)
[2013-04-12 00:36:42] GPU 0 AMD Radeon HD 7970M hardware monitoring enabled
[2013-04-12 00:36:42] 1 GPU devices max detected
[2013-04-12 00:36:42] USB all: found 6 devices - listing known devices
[2013-04-12 00:36:43] No known USB devices
Can someone guess whats the problem? I described what i installed and tried here already: https://bitcointalksearch.org/topic/m.1799428
I checked the faq and so on but cant find the problem.
April 11, 2013, 06:32:35 PM
Quick question: is it possible to mine with the stratum protocol through an http proxy?
Theoretically yes, but only if the http proxy is very permissive, and none will by default. A socks proxy will definitely work though. 
April 11, 2013, 06:30:17 PM
Quick question: is it possible to mine with the stratum protocol through an http proxy?
April 11, 2013, 05:16:43 PM
I suspect that when switching users in windows 7 cgminer hangs. To p2pool it looks like it keeps on mining but with "100% local DOA" and the console interface is broken, nothing gets updated when I press Q or G etc.
Is this a known bug?
Is this a known bug?
From what I have read, it is my understanding that cgminer does not like user switching.
April 11, 2013, 04:45:04 PM
I suspect that when switching users in windows 7 cgminer hangs. To p2pool it looks like it keeps on mining but with "100% local DOA" and the console interface is broken, nothing gets updated when I press Q or G etc.
Is this a known bug?
Is this a known bug?
April 11, 2013, 04:12:02 PM
Hi,
i have 4 zTex FPGA running with cgminer - on startup i get the following message:
HS FPGA configuration failed: DONE pin does not go high
The boards are calculating - but what does this message say?
i have 4 zTex FPGA running with cgminer - on startup i get the following message:
HS FPGA configuration failed: DONE pin does not go high
The boards are calculating - but what does this message say?
April 11, 2013, 03:39:23 PM
Maybe not exactly a cgminer question...
I have to identical cards. How can i examine which one is gpu0 and which one is gpu1??
Greets
ImI
try cgminer.exe -n and check the gpu and the platform.
Jump to: