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

[ddd1]

I have good temps with water and wonder how I overclock more then AMD overdrive MAX values?

I use a program called CGMiner to do that.
Sam

yes I'm using cgminer but I can't get it overclock more then the AMD Overdrive limits.

I'm on the latest 2.4.1 and windows 7 64bit.

[DeathAndTaxes]

cgminer uses the driver library to set clocks.
cgminer can't set clocks higher than drivers & GPU BIOS will allow them to be set.

If you must have higher clocks your options are limited to:
a) utilities which bypass the drivers and force GPU clock change directly (MSI AB, Trixx, etc)
b) flashing a custom BIOS which allows higher settings

[check_status]

Using 2.4.0 there seemed to be a lot of rejected, greater than 10% of accepted. I switched to 2.4.1 and now the rejected rate is less than 5% of accepted. Efficiency is 30% on average, a 50% increase over 2.3.2 with identical settings while hashing on p2pool. So far so good.

[os2sam]

I have good temps with water and wonder how I overclock more then AMD overdrive MAX values?

I use a program called CGMiner to do that.
Sam

yes I'm using cgminer but I can't get it overclock more then the AMD Overdrive limits.

I'm on the latest 2.4.1 and windows 7 64bit.

Sorry for my somewhat sarcastic response.  I thought you were referring to the max limits that exist in the CCC program and that CG Miner can surpass.
Sam

[kano]

Using 2.4.0 there seemed to be a lot of rejected, greater than 10% of accepted. I switched to 2.4.1 and now the rejected rate is less than 5% of accepted. Efficiency is 30% on average, a 50% increase over 2.3.2 with identical settings while hashing on p2pool. So far so good.

Efficiency effectively means nothing to anything but the pool and your internet connection.
Efficiency is average how many shares are accepted per getwork request to the pool.
rollntime allows a getwork request to locally be expanded into many getwork requests (max of 10 in cgminer)

On p2pool you will regularly abort your share processing due to the 10s LP time (60 times as often as on a normal pool)
When an LP hits, the current work is stopped (and any shares found are submitted, by default)
Then new work related to the LP is started.
Thus e.g. if you only process half of each work item you will return (on average) 1 share per 2 work items.
i.e. Efficiency should be something like 50%

[dietwice]

Hi All,

I'm trying to run my 4-GPU rig and having something strange.

It runs perfectly with 3 GPUs, but when the 4th is being added it still runs only those 3 were running before.

winxp32, driver 11.11, cgminer 2.4.1

Here are some logs:

C:\ATI\04>cgminer.exe -n
 [2012-05-09 16:48:35] CL Platform 0 vendor: Advanced Micro Devices, Inc.
 [2012-05-09 16:48:35] CL Platform 0 name: AMD Accelerated Parallel Processing
 [2012-05-09 16:48:35] CL Platform 0 version: OpenCL 1.1 AMD-APP (831.4)
 [2012-05-09 16:48:35] Platform 0 devices: 4
 [2012-05-09 16:48:35]  0       Juniper
 [2012-05-09 16:48:35]  1       Juniper
 [2012-05-09 16:48:35]  2       Juniper
 [2012-05-09 16:48:35]  3       Cayman
 [2012-05-09 16:48:35] GPU 0 AMD Radeon HD 6900 Series  hardware monitoring enabled
 [2012-05-09 16:48:35] GPU 1 ATI Radeon HD 5700 Series hardware monitoring enabled
 [2012-05-09 16:48:35] GPU 2 AMD Radeon HD 6700 Series hardware monitoring enabled
 [2012-05-09 16:48:35] GPU 3 ATI Radeon HD 5700 Series hardware monitoring enabled
 [2012-05-09 16:48:35] 4 GPU devices max detected

So the 4nd GPU (AMD Radeon HD 6900 Series Cayman) is seen by both CL and ADL.


 cgminer version 2.4.1 - Started: [2012-05-09 16:36:00]
--------------------------------------------------------------------------------
 (5s):540.2 (avg):542.9 Mh/s | Q:26  A:5  R:0  HW:0  E:19%  U:10.8/m
 TQ: 8  ST: 8  SS: 0  DW: 15  NB: 2  LW: 16  GF: 0  RF: 0
 Connected to http://mining.eligius.st:8337 with LP as user 12pvJujLTCtUnJB5J9Ap4qUe6gthKy2TcG
 Block: 00000146cca09914093f182831cf3d76...  Started: [16:36:15]
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [ S]ettings [D]isplay options [Q]uit
 GPU 0:  42.0C  50%    | 187.1/221.1Mh/s | A:1 R:0 HW:0 U:  2.15/m I: 7
 GPU 1:  54.0C 2007RPM | 164.7/197.0Mh/s | A:3 R:0 HW:0 U:  6.46/m I: 7
 GPU 2:  56.0C 2033RPM | 181.7/214.8Mh/s | A:1 R:0 HW:0 U:  2.15/m I: 7
 GPU 3:  58.0C  50%    |   0.0/  0.0Mh/s | A:0 R:0 HW:0 U:  0.00/m I: 7

When started cgminer runs about a minute producing shares by GPUs 0-2, then says:
...
[2012-05-09 16:37:10] Device 3 idle for more than 60 seconds, GPU 3 declared SICK!
[2012-05-09 16:37:10] GPU still showing activity suggesting a hard hang.
[2012-05-09 16:37:10] Will not attempt to auto-restart it.

This GPU had been working perfectly about a year till today as a secong GPU at my home workstation under win7x64.
Now I want it to run in a rig.
Are there any suggestions how can I diagnose that case?

Thanks in advance.

[dietwice]

Heh, looks like I packed that rig tightly just to unpack it again.

Ok, going to unpack it.

p.s. I use no extensions and no displays, just resistors.

[ddd1]

the HW column which I believe stands for hardware error.

One of my cards has 0 while the other has 2.5% hardware errors ( HW/(A+R+HW)= hardwareerror divided by total accepted+rejected+HW. )

Should I downclock that card some because I think those errors are because of overclock?

[-ck]

the HW column which I believe stands for hardware error.

One of my cards has 0 while the other has 2.5% hardware errors ( HW/(A+R+HW)= hardwareerror divided by total accepted+rejected+HW. )

Should I downclock that card some because I think those errors are because of overclock?

If the rate is very low, then leaving it is fine. However 2.5% is fairly significant. Think of them like the glitches you see on screen when you overclock too much. I'd recommend just come backing a little on the overclock till they go away.

[DutchBrat]

Ckolivas,

As you know I am mining on Bonuspool, last night the pool was disabled by CGMiner after 3 mins of rejects and although it came back to  live,  not all of my miners went back to work there, some did

Code:

 cgminer version 2.4.1 - Started: [2012-05-06 15:22:54]
--------------------------------------------------------------------------------
 (5s):320.5 (avg):319.4 Mh/s | Q:38171  A:22639  R:849  HW:35  E:59%  U:4.3/m
 TQ: 2  ST: 6  SS: 181  DW: 5934  NB: 638  LW: 55164  GF: 84  RF: 215
 Connected to http://pit.deepbit.net:8332 with LP as user XXXX
 Block: 00000022fe8262b5f692521c7ada0720...  Started: [07:50:39]
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit
 GPU 0:  73.0C 2483RPM | 321.5/319.4Mh/s | A:22639 R:849 HW:35 U: 4.26/m I: 6
--------------------------------------------------------------------------------

0: Rejecting Alive Priority 0: http://pool.bonuspool.co.cc:80  User:XXXX
1: Enabled Dead Priority 1: http://pool.ABCPool.co:8332  User:XXX
2: Enabled Alive Priority 2: http://pit.deepbit.net:8332  User:XXX
3: Enabled Alive Priority 3: http://mine.btcguild.com:8332  User:XXX

Current pool management strategy: Failover
[A]dd pool [R]emove pool [D]isable pool [E]nable pool
[C]hange management strategy [S]witch pool [I]nformation
Or press any other key to continue

As you can see, it is still rejecting

Cheers,

Brat

p.s. Just by sheer luck I happen to see that on one of my other miners where Pool  0 also was not re-enabled, Pool 1, which is enabled but DEAD as you can see above, came back alive and that miner switched to Pool 1 immediately.... but Pool 0 stayed in Rejecting mode. After enabling it manually, the miners started delivering shares

[kano]

Go visit the ABCPool thread - ABCPool was dead last time I looked there a couple of hours ago.
Edit: OK someone posted that it's alive again (but it was dead for quite a while)

[DutchBrat]

Go visit the ABCPool thread - ABCPool was dead last time I looked there a couple of hours ago.
Edit: OK someone posted that it's alive again (but it was dead for quite a while)

Kano:

I know, it wasn't the issue with ABCPool I was talking about, when ABCPool came back to life and my miner switched back as it is supposed to

It is the new feature in where a pool gets disabled by CGMiner when there are only rejects over a period of 3 mins.
CGMiner then checks on every new LP if the rejected pool is back up and is supposed to switch back when the pool is alive again
It never came out of the disabled state on 3 of my miners, while it did go back to mining on a few others.... It could be that the pool was never disabled because of rejects on the other miners in the first place

All I know is on the miners where CGMiner disabled the pool due to rejects, the pool was never re-enabled when it started accepting shares again

I am talking about Pool 0 (Bonuspool), not Pool 1

[AniceInovation]

Is it possible to have an option like --computer-restart, to restart the computer when it detects a sick/dead card?
I a aware there is a gpu restart, but this would be a nice option for me too.

[-ck]

Is it possible to have an option like --computer-restart, to restart the computer when it detects a sick/dead card?
I a aware there is a gpu restart, but this would be a nice option for me too.

That's not possible in a cross platform way. It's not even possible to cleanly close cgminer once a driver has crashed (dead device). You need to use the API or log to monitor it happening and use an external program to do that.

[Askit2]

I hate to waste your time with stupid questions but I think I need to. Sorry in advance. On the per device line how often is the first value calculated the one that shows me the non averaged speed of my device?
Thank you for all your help and wonderful software

[The00Dustin]

I hate to waste your time with stupid questions but I think I need to. Sorry in advance. On the per device line how often is the first value calculated the one that shows me the non averaged speed of my device?
Thank you for all your help and wonderful software

CK/Kano, I'm not answering this one because I'm not 100% certain on a couple of things.  In case you aren't watching the BFL thread, you might read this post: https://bitcointalk.org/index.php?topic=60586.msg891966#msg891966 before replying to his question.  I am not 100% certain an incorrect assumption isn't being made there, and if so, the answer to this question may not help regarding its root cause (just trying to save everyone a couple steps).

[kano]

I hate to waste your time with stupid questions but I think I need to. Sorry in advance. On the per device line how often is the first value calculated the one that shows me the non averaged speed of my device?
Thank you for all your help and wonderful software

Default is 5seconds (as it says )
That's:
--log|-l <arg>      Interval in seconds between log output (default: 5)

(I'm just answering the actual question ...)

Edit: however on that other page linked above you have set it to 2s - so I'm not sure why you are asking.

[Askit2]

not the default log info up top. the per device info below it. You know generic all your craps info then a line
------------------------------------------------------------------------------------------------------------------------
then something like BFL 0 56.2 811.1/796(avg) blah blah blah......
How long of a time is the 811.1 in my example calculated over? not the uppper one I can read it. the single Unit one.

[PolymorphicAssasin]

...
Updated git tree:
I've added dynamic adjustment of intensity for usage on a normal desktop. This flag is enabled by default and tests to ensure the GPU is constantly available for desktop use and will scale intensity down when you are watching a movie, gaming or just about any other intense GPU usage, and scale it up when otherwise idle. This is best disabled on dedicated miners:
...
The difference can be quite dramatic in terms of how much smoother the desktop is, and how much higher the throughput is when it's idle.

I just switched over from Diablo on my work PC, and I have to say this is the one of the coolest things about cgminer.  I don't have to kill my miner whenever I'm doing graphic intense work (2x or 3x a day).  Kudos.

I'll send you some bitcents when next I access my wallet (from cold storage)

Thanks for some great software!

[-ck]

...
Updated git tree:
I've added dynamic adjustment of intensity for usage on a normal desktop. This flag is enabled by default and tests to ensure the GPU is constantly available for desktop use and will scale intensity down when you are watching a movie, gaming or just about any other intense GPU usage, and scale it up when otherwise idle. This is best disabled on dedicated miners:
...
The difference can be quite dramatic in terms of how much smoother the desktop is, and how much higher the throughput is when it's idle.

I just switched over from Diablo on my work PC, and I have to say this is the one of the coolest things about cgminer.  I don't have to kill my miner whenever I'm doing graphic intense work (2x or 3x a day).  Kudos.

I'll send you some bitcents when next I access my wallet (from cold storage)

Thanks for some great software!

Glad you appreciate it. I think it's one of the most underrated features myself too since most miners tend to just use dedicated rigs. Just remember you can also tune the intensity of the dynamic mode as well with the --gpu-dyninterval option.