OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.1

[-ck]

Yes that's right, the "donated" power so far is amassing to between 100-200 Mhash at the moment. The bithopper app which uses a 1% donation model is getting ~2.4Ghash from "donations".

[Wepp]

Hi there. I'm running a Windows 7 x64 machine and dual 6990's. I just switched to cgminer after having used poclbm directly launched from the command line for about 3 months without issues. Since switching, I have three problems:

1) Without fail, every time I [Q]uit the miner, it causes a BSOD. The driver reporting a failure is atikmdag.sys. I was trying Catalyst 11.5 when it started happening. Updated to Catalyst 11.9 and the behavior is the same. The only way for me to safely exit the program is to KILL the process, and then I use another program to bring the fan speeds down to normal, or reboot.

2) Attempting to set the memory clock speed to anything below 950Mhz causes instability. Anything below 900Mhz and the system locks up instantly. I've found the best course of action is to leave it at the default setting.

3) Fan speed. NO MATTER WHAT I DO, one of the fans eventually reaches 100%. Ambient temperature in the room is 68º. I never had this problem using poclbm directly. Here are my settings:
--auto-fan --auto-gpu --gpu-engine 800-900 --gpu-fan 60-75 --temp-target 85 --temp-overheat 90

I believe what is happening with the fan speed problem is that cgminer is attributing each fan to the wrong GPUs. Invariably, GPU 0 and GPU 2 are cool, at around 75º, while GPU1 and GPU3 are hot at around 89-91º. My guess is that 0+2 is one card, and 1+3 is the second card, but cgminer is thinks that the cards are 0+1 and 2+3 respectively. Therefore, it is not associating the GPU temperatures with the correct GPU fan speed settings. In fact, 0+2 used to run hotter than 1+3, because the card is physically on top of the other card in the machine. It is only so cool now because the fan is blowing so hard.

[kano]

Yay, finally got a real one (pool of course, not solo) Block 147825 (Ars)
[2011-10-03 10:12:03] Accepted 00000000.00000943.a6f1622c BLOCK! GPU 0 thread 0 pool 1
(That's Aus EDT 35 minutes ago)

[The LT]

Switched on 0.5% donation on my 4gh/s farm, hardly felt the difference. Hope that makes the difference for you, Con! Keep up the good work. Thanks!

[-ck]

Hi there. I'm running a Windows 7 x64 machine and dual 6990's. I just switched to cgminer after having used poclbm directly launched from the command line for about 3 months without issues. Since switching, I have three problems:

1) Without fail, every time I [Q]uit the miner, it causes a BSOD. The driver reporting a failure is atikmdag.sys. I was trying Catalyst 11.5 when it started happening. Updated to Catalyst 11.9 and the behavior is the same. The only way for me to safely exit the program is to KILL the process, and then I use another program to bring the fan speeds down to normal, or reboot.

2) Attempting to set the memory clock speed to anything below 950Mhz causes instability. Anything below 900Mhz and the system locks up instantly. I've found the best course of action is to leave it at the default setting.

3) Fan speed. NO MATTER WHAT I DO, one of the fans eventually reaches 100%. Ambient temperature in the room is 68º. I never had this problem using poclbm directly. Here are my settings:
--auto-fan --auto-gpu --gpu-engine 800-900 --gpu-fan 60-75 --temp-target 85 --temp-overheat 90

I believe what is happening with the fan speed problem is that cgminer is attributing each fan to the wrong GPUs. Invariably, GPU 0 and GPU 2 are cool, at around 75º, while GPU1 and GPU3 are hot at around 89-91º. My guess is that 0+2 is one card, and 1+3 is the second card, but cgminer is thinks that the cards are 0+1 and 2+3 respectively. Therefore, it is not associating the GPU temperatures with the correct GPU fan speed settings. In fact, 0+2 used to run hotter than 1+3, because the card is physically on top of the other card in the machine. It is only so cool now because the fan is blowing so hard.

1. Well that's very unlucky. Not sure what I can do about it though, software shouldn't be able to take out your entire machine (unless you're doing dangerous clocks and shit).
2. Some hardware really doesn't like the memory being much slower than the gpu clock speed. 69x0 cards in particular usually ignore lower settings to avoid that from happening.
3. The 6990 has dissociated sensors on the 2nd core so it really doesn't know what the fan speed is. Furthermore, the ATI drivers and the ATI display library do NOT have a way of agreeing on what the device numbers are, so we're left to *hoping* they coincide in order. However, the temperature and fan speeds reported by ADL are coming from the same device so even if your GPU number listed doesn't correspond with the fanspeed you think, the temperature/fanspeed combination listed for one card is definitely the temperature and fanspeed for that core. Bear in mind, though, that the 2nd core of 6990s doesnt have a fanspeed and is just sharing the fanspeed from the 1st core. Thus any fanspeed it's reporting will likely be a percentage value that is being ignored and the fan of its corresponding partner core in RPM is the real fanspeed.

[-ck]

Switched on 0.5% donation on my 4gh/s farm, hardly felt the difference. Hope that makes the difference for you, Con! Keep up the good work. Thanks!

Thanking you muchly!

[Remember remember the 5th of November]

So, i tried compiling on mingw32 and

In file included from sigaction.c:21:0:
./signal.h:616:3: error: expected specifier-qualifier-list before 'uid_t'
./signal.h:659:1: error: conflicting types for 'restrict'
./signal.h:659:1: note: previous definition of 'restrict' was here
sigaction.c:130:58: error: expected ';', ',' or ')' before 'act'
sigaction.c:84:1: warning: 'sigaction_handler' defined but not used

This on windows. The configure i use is

CFLAGS="-g -O3 -Wall -msse2 -march=amdfam10 -msse4a -mtune=amdfam10 -mabm -co
mbine -funroll-all-loops -ffast-math -fprefetch-loop-arrays -ftree-parallelize-
loops=4 -IC:/APP/include -IC:\MinGW\msys\1.0\home\curl\include\curl" LDFLAGS="-
LC:/APP/lib/x86 -LC:\MinGW\msys\1.0\local\lib\libcurl.a" ./configure

and then just make and i get that error.

[exahash]

Can you copy and paste your AMDOverDriver configuration file, with overvolt, for your 5970?!

I've never been able to overvolt with AMDOC, sorry if that wasn't clear.  I was saying you could potentially do better if *you* have the ability to overvolt (using a different tool like Trixx or Afterburner).

I do know that running the fans at 100% is risky.  I've siezed up four of them on my 5850's (my 5970s are all fine so far).  I just remove the fan, lay the card down and put a std 80mm fan over the hole, problem solved.  If all my fans melt down it'll be a problem, but I needed the extra air flow over the summer.  With colder weather coming I'll probably try reducing them all in a few weeks.

Here is the AMDOverdriveCtrl file for one half of one 5970:

Code:

<?xml version="1.0" encoding="utf-8"?>
<OVERDRIVE_PROFILE>
  <PERFORMANCE_LEVEL level="2" gpu="87000" mem="18000" voltage="1050"/>
  <PERFORMANCE_LEVEL level="1" gpu="55000" mem="15000" voltage="1038"/>
  <PERFORMANCE_LEVEL level="0" gpu="15700" mem="15000" voltage="950"/>
  <FAN_SETTING percentage="100"/>
  <FAN_CTRL enabled="no"/>
  <FAN_CTRL_CURVE type="0"/>
  <FAN_CTRL_POINT nr="0" temperature="2000" percentage="0"/>
  <FAN_CTRL_POINT nr="1" temperature="4000" percentage="2500"/>
  <FAN_CTRL_POINT nr="2" temperature="6000" percentage="5000"/>
  <FAN_CTRL_POINT nr="3" temperature="8000" percentage="7500"/>
  <FAN_CTRL_POINT nr="4" temperature="10000" percentage="10000"/>
  <MONITOR_SAMPLE_TIME interval="10"/>
  <COLOR_PROFILE enabled="no" longitude="-13.000000" latitude="52.000000" color_temp_day="0" color_temp_night="0" transition="30"/>
  <POWERTUNE percentage="0"/>
</OVERDRIVE_PROFILE>

IIRC, I just took the default file and tweaked the gpu and mem speeds... when I tried the voltage AMDOC hung because it couldn't increase it.

[AniceInovation]

Hello again, i have 2 questions:

I have a mixed rig, with
-GPU0:5850
-GPU1:5870
-GPU2:6950
-GPU3:5850
-GPU4:6950

when running: cgminer -o url -u user -p pass -w 256 -v 2 -I 8,8,9,8,9 --gpu-engine 930,930,930,930,930 --gpu-memclock 300 --donation 1

Only the first 2 mine correcly, the other 3 kind of tries to mine from gpu0 and gpu1. i think gpu2, gpu3 and gp4 are getting wrongly detected.
Maybe it's something to do with the 5800 series being detected 2 devices, and the 6900 series being detected 3? (bad english, hope you understand)

EDIT: removed -w 256 and -v 2 and it looks fine now.

second, can you help me or point me up to instructions to install CGMINER on Ubuntu 11 x64?

[kano]

...
second, can you help me or point me up to instructions to install CGMINER on Ubuntu 11 x64?

Try my xubuntu 11.04 script in the main directory ... linux-usb-cgminer
Start from step 6)

[c_k]

2 request for features please, since we've been donating a while:

1.) SSL support (to the extent of what bitcoind offers)
2.) SOCKS proxy support (to the same extent bitcoind supports)

Cheers! 

[sirky]

I just want to say thanks for all the hard work! I just donated a bit more than 4 btc. I don't have a request, only thanks.

[Red Emerald]

I wish I had found this program weeks ago.  I get more hash/s than phoenix and all my GPUs are setup with one command!

Thanks for your great work!

[Diapolo]

I tried

Code:

-d 0 -v 2 -w 128 --auto-fan --auto-gpu --gpu-fan 25-50 --gpu-engine 600-900 --gpu-memclock 350 --temp-target 60

for my HD5870, which is my 1st card, which works, but the 2nd card gets the fan ramped up to 85% even if it is disabled. What about this? Bug or a stupid config error from me?

By the way, is there a chance to reset the cards values to default if clicking the close button on Windows? CGMINER only resets the values, if I press Q.

Thanks,
Dia

[Meatball]

I tried

Code:

-d 0 -v 2 -w 128 --auto-fan --auto-gpu --gpu-fan 25-50 --gpu-engine 600-900 --gpu-memclock 350 --temp-target 60

for my HD5870, which is my 1st card, which works, but the 2nd card gets the fan ramped up to 85% even if it is disabled. What about this? Bug or a stupid config error from me?

By the way, is there a chance to reset the cards values to default if clicking the close button on Windows? CGMINER only resets the values, if I press Q.

Thanks,
Dia

I don't think that's possible honestly.  Clicking the X on a Dos window just flat out kills the process, so there's no way the program will be able to reset things.

You can create a second batch file to launch your miner again with all the defaults or just make sure you press Q

[m3ta]

Can not compile cgminer-2.0.5:

Code:

make[2]: Entering directory `/home/administrativo/cgminer-2.0.5'
  CC     cgminer-main.o
main.c: In function ‘main’:
main.c:5418:12: error: ‘struct cgpu_info’ has no member named ‘gpu_engine’
main.c:5419:12: error: ‘struct cgpu_info’ has no member named ‘gpu_memclock’
main.c:5420:12: error: ‘struct cgpu_info’ has no member named ‘gpu_vddc’
main.c:5421:12: error: ‘struct cgpu_info’ has no member named ‘gpu_fan’
main.c:5422:12: error: ‘struct cgpu_info’ has no member named ‘gpu_powertune’
make[2]: *** [cgminer-main.o] Error 1
make[2]: Leaving directory `/home/administrativo/cgminer-2.0.5'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/administrativo/cgminer-2.0.5'
make: *** [all] Error 2

Ubuntu 11.04 32 bits.

CGMiner 2.0.4 compiles just fine.

What can I do?!

Thanks!
Thiago

Still no solution?

[Luke-Jr]

Can not compile cgminer-2.0.5:

Code:

make[2]: Entering directory `/home/administrativo/cgminer-2.0.5'
  CC     cgminer-main.o
main.c: In function ‘main’:
main.c:5418:12: error: ‘struct cgpu_info’ has no member named ‘gpu_engine’
main.c:5419:12: error: ‘struct cgpu_info’ has no member named ‘gpu_memclock’
main.c:5420:12: error: ‘struct cgpu_info’ has no member named ‘gpu_vddc’
main.c:5421:12: error: ‘struct cgpu_info’ has no member named ‘gpu_fan’
main.c:5422:12: error: ‘struct cgpu_info’ has no member named ‘gpu_powertune’
make[2]: *** [cgminer-main.o] Error 1
make[2]: Leaving directory `/home/administrativo/cgminer-2.0.5'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/administrativo/cgminer-2.0.5'
make: *** [all] Error 2

Ubuntu 11.04 32 bits.

CGMiner 2.0.4 compiles just fine.

What can I do?!

Thanks!
Thiago

Still no solution?

It's fixed in git.

[bronan]

This far i think its a nice program but the auto-engine seems to fail
Manually had my 5870 set to its basic clocks at 850 core 1200 memory
When i add the parameters for core 850-950 and memory 450 nothing happens when i watch over time with gpu-z it keeps running at the speed it started with so no change at all
Now i certain this card can run at 890 Mhz but i have not seen any increase... maybe i am missing something here

[Sekioh]

This far i think its a nice program but the auto-engine seems to fail
Manually had my 5870 set to its basic clocks at 850 core 1200 memory
When i add the parameters for core 850-950 and memory 450 nothing happens when i watch over time with gpu-z it keeps running at the speed it started with so no change at all
Now i certain this card can run at 890 Mhz but i have not seen any increase... maybe i am missing something here

I had issues getting my cards to clock at all until a little bit ago so haven't tested, but from what I read you manually set the cards to the HIGH side, and the range will clock DOWN based on throttling.

[-ck]

Thanks to continued donations  Much appreciated.

New version: 2.0.6
- Expire shares as stale with a separate timeout from the scantime, defaulting
to 120 seconds.
- Retry pools after a delay of 15 seconds if none can be contacted on startup
unless a key is pressed.
- Don't try to build adl features without having adl.
- Properly check shares against target difficulty - This will no longer show
shares when solo mining at all unless they're considered to be a block solve.
- Add altivec 4 way (cpu mining) support courtesy of Gilles Risch.
- Try to use SSL if the server supports it.
- Display the total solved blocks on exit (LOL if you're lucky).
- Use ADL activity report to tell us if a sick GPU is still busy suggesting it
is hard hung and do not attempt to restart it.
- Don't make donation work interfere with block change detection allowing
donation to work regardless of the block chain we're mining on.