DiabloMiner GPU Miner

[DiabloD3]

Use Debian unstable, use whatever is packaged newest by Debian (do NOT use the .run installer provided by AMD), and use AMDOverdriveCtrl to control clocks and voltage.

Correct me if I'm wrong, but AMDOverdriveCtrl cannot change the voltage or lower the memory clock substantially (unless it's been updated since I last used it).

It can set the memory clock to whatever your GPU's firmware is limited to, this is usually far below the minimum displayed in CCC.

It can set the voltage on any card that uses stock VRMs.

[1713999115]

1713999115

[1713999115]

1713999115

[sveetsnelda]

It can set the memory clock to whatever your GPU's firmware is limited to, this is usually far below the minimum displayed in CCC.

It can set the voltage on any card that uses stock VRMs.

The 6000 and 7000 series cards have always been limited to 125Mhz below engine clock (IIRC, my 7000's went 150Mhz below engine clock).  I've had to flash every 6000/7000 card to drop the memory clock by a significant amount and to drop the voltage at all (in Linux).  It's been about 4 months since I've tried AMDOverdriveCtrl though.  Apparently I need to give it another go?  I have 84 video cards with modified BIOSes.  It'd be a godsend to be able to hit low memory clocks and adjust voltages without flashing cards anymore.

The 5000's are a different story...  a reference board will clock to just about whatever engine/memory/voltage is entered.

[DiabloD3]

It can set the memory clock to whatever your GPU's firmware is limited to, this is usually far below the minimum displayed in CCC.

It can set the voltage on any card that uses stock VRMs.

The 6000 and 7000 series cards have always been limited to 125Mhz below engine clock (IIRC, my 7000's went 150Mhz below engine clock).  I've had to flash every 6000/7000 card to drop the memory clock by a significant amount and to drop the voltage at all (in Linux).  It's been about 4 months since I've tried AMDOverdriveCtrl though.  Apparently I need to give it another go?  I have 84 video cards with modified BIOSes.  It'd be a godsend to be able to hit low memory clocks and adjust voltages without flashing cards anymore.

The 5000's are a different story...  a reference board will clock to just about whatever engine/memory/voltage is entered.

Yes, as long as you've modified the firmwares, you're fine.

[ensign_lee]

Gotta say Diablo, your miner kinda rocks. Thanks so much for making it available for everyone!

[Prattler]

I've had to flash every 6000/7000 card to drop the memory clock by a significant amount and to drop the voltage at all (in Linux).

Would you please share the knowledge how to flash (or modify the bios of) 7000 cards?

[ShadesOfMarble]

Is there a way to use optimal settings for both VLIW and GCN without running two instances of DiabloMiner?

[DiabloD3]

Is there a way to use optimal settings for both VLIW and GCN without running two instances of DiabloMiner?

No, not yet.

[tinman951]

I've been using DiabloMiner for months on Windows, but I wanted to try it on mac (10.7) with a 6750M.  I downloaded the newest version and cd'd the directory.  Using slush,
So I type    ./DiabloMiner-OSX.sh -u username -p password -0 api.bitcoin.cz -r 8332 -w 64

It opens and starts to run
Started
Connecting to :http..api2.bitcoin.cz:8332
Using Apple OpenCL 1.1 (Jul 25 2011 time)
Added ATI Radeon HD 6750M.......

But then it will say ERROR: Cannot connect to api2.bitcoin.cz connect time out

I have checked the password and user name several times, even created a new worker.
Is the OpenCL up to date? All I can find are .exe files for OpenCL and SDK upgrades.

Any help would be greatly appreciated.

[DiabloD3]

I've been using DiabloMiner for months on Windows, but I wanted to try it on mac (10.7) with a 6750M.  I downloaded the newest version and cd'd the directory.  Using slush,
So I type    ./DiabloMiner-OSX.sh -u username -p password -0 api.bitcoin.cz -r 8332 -w 64

It opens and starts to run
Started
Connecting to :http..api2.bitcoin.cz:8332
Using Apple OpenCL 1.1 (Jul 25 2011 time)
Added ATI Radeon HD 6750M.......

But then it will say ERROR: Cannot connect to api2.bitcoin.cz connect time out

I have checked the password and user name several times, even created a new worker.
Is the OpenCL up to date? All I can find are .exe files for OpenCL and SDK upgrades.

Any help would be greatly appreciated.

This isn't a miner bug, are you sure slush's pool isn't actually down?

[tinman951]

I've been using DiabloMiner for months on Windows, but I wanted to try it on mac (10.7) with a 6750M.  I downloaded the newest version and cd'd the directory.  Using slush,
So I type    ./DiabloMiner-OSX.sh -u username -p password -0 api.bitcoin.cz -r 8332 -w 64

It opens and starts to run
Started
Connecting to :http..api2.bitcoin.cz:8332
Using Apple OpenCL 1.1 (Jul 25 2011 time)
Added ATI Radeon HD 6750M.......

But then it will say ERROR: Cannot connect to api2.bitcoin.cz connect time out

I have checked the password and user name several times, even created a new worker.
Is the OpenCL up to date? All I can find are .exe files for OpenCL and SDK upgrades.

Any help would be greatly appreciated.

This isn't a miner bug, are you sure slush's pool isn't actually down?

I know the miner is working, and Slush was not down.  But how do I know the SDK or OpenCL is up to date for it to run correctly.  The only other option is that somehow the port is blocked.

[DiabloD3]

I've been using DiabloMiner for months on Windows, but I wanted to try it on mac (10.7) with a 6750M.  I downloaded the newest version and cd'd the directory.  Using slush,
So I type    ./DiabloMiner-OSX.sh -u username -p password -0 api.bitcoin.cz -r 8332 -w 64

It opens and starts to run
Started
Connecting to :http..api2.bitcoin.cz:8332
Using Apple OpenCL 1.1 (Jul 25 2011 time)
Added ATI Radeon HD 6750M.......

But then it will say ERROR: Cannot connect to api2.bitcoin.cz connect time out

I have checked the password and user name several times, even created a new worker.
Is the OpenCL up to date? All I can find are .exe files for OpenCL and SDK upgrades.

Any help would be greatly appreciated.

This isn't a miner bug, are you sure slush's pool isn't actually down?

I know the miner is working, and Slush was not down.  But how do I know the SDK or OpenCL is up to date for it to run correctly.  The only other option is that somehow the port is blocked.

You cant control OpenCL on OSX, its up to Apple there.

However, the error you pasted only happens when it cant connect. It has nothing to do with the rest of the miner.

[ShadesOfMarble]

How many hw errors are ok for a 7970? I currently have 0,17% (52/29410) hw errors, card running overclocked (1125 Mhz) and undervolted (1100 mV).
Everything running stable for >2 days now.

[Elmojo]

Hi all,
I've searched the thread, but came up with nothing...
My desktop PC has 2 GPUs: an Nvidia GTX480 as my primary work and gaming card, and a brand new Radeon 7970, just for mining.
I'd like to just run diablo on the 7970, and have it ignore the 480.
Is there a command line switch for this?
Please keep in mind that I'm really new at mining and configuring miners, so please spell it out as though you were explaining it to your granny.
Thanks,
Elmojo

[Wordlet]

To specify a specific card you can use the -D option and add a number at the end. This number will always be the same afaik as long as you dont remove/install cards.

For example:

DiabloMiner-Windows.exe -u user -p password -o pool.address.com -r port(e.g. 8337) -v 2 -D 2

Where 2 would be the number for your mining card. Since you have 2 cards, it could be either 1 or 2. Diablo tells you what card it's found and is going to use so trial and error will get you what number you need to put there.

[Elmojo]

To specify a specific card you can use the -D option and add a number at the end. This number will always be the same afaik as long as you dont remove/install cards.

For example:

DiabloMiner-Windows.exe -u user -p password -o pool.address.com -r port(e.g. 8337) -v 2 -D 2

Where 2 would be the number for your mining card. Since you have 2 cards, it could be either 1 or 2. Diablo tells you what card it's found and is going to use so trial and error will get you what number you need to put there.

Awesome! Exactly what I was looking for.
-Elmojo

[DiabloD3]

Hi all,
I've searched the thread, but came up with nothing...
My desktop PC has 2 GPUs: an Nvidia GTX480 as my primary work and gaming card, and a brand new Radeon 7970, just for mining.
I'd like to just run diablo on the 7970, and have it ignore the 480.
Is there a command line switch for this?
Please keep in mind that I'm really new at mining and configuring miners, so please spell it out as though you were explaining it to your granny.
Thanks,
Elmojo

Use -D, it counts starting 1 not 0. So most likely you want -D 1 (assuming your 7970 is listed first at startup, -D 2 otherwise).

Edit: Heh, someone beat me to it

[DiabloD3]

How many hw errors are ok for a 7970? I currently have 0,17% (52/29410) hw errors, card running overclocked (1125 Mhz) and undervolted (1100 mV).
Everything running stable for >2 days now.

Significant errors like that is insane.  Set stock clocks and volts and see if it still does it, if it does its a driver bug, if it doesn't, you overclocked too far. You shouldn't see more than 1 or 2 HW errors per 10k shares give or take, if it starts skyrocketing you went too far.

[Joshwaa]

Mine are at 1.006v 1160/1000

If you get HW errors like that I think your volts might be too low. What is you ASIC quality %? That will give you a good idea how low you can go on the volts.

[Blackasaurus]

Mine are at 1.006v 1160/1000

If you get HW errors like that I think your volts might be too low.

I think this is correct, if you tweak it I'm pretty sure it should sit at 0 hw errors (I haven't had any)

[e21]

DiabloMiner-Linux noob question, sorry if it's been asked, but I couldn't find anything anywhere on this forum or google regarding this error. I downloaded python-jsonrpc and ran setup.py, installed the SDK and drivers, etc, when I run Diablo I get this:



user@HOST:~/DiabloMiner$ ./DiabloMiner-Linux.sh -u user.worker -p password -o api2.bitcoin.cz -r 8332
[2/23/12 11:12:09 PM] Started                                               
[2/23/12 11:12:09 PM] Connecting to: http://api2.bitcoin.cz:8332/           
[2/23/12 11:12:09 PM] Using AMD Accelerated Parallel Processing OpenCL 1.1 AMD-APP-SDK-v2.4 (595.10)
[2/23/12 11:12:09 PM] ERROR: Cannot connect to api2.bitcoin.cz: Bitcoin returned unparsable JSON
[2/23/12 11:12:10 PM] BFI_INT patching enabled, disabling hardware check errors
[2/23/12 11:12:10 PM] Added Turks (#1) (6 CU, local work size of 256)       
[2/23/12 11:12:11 PM] ERROR: Cannot connect to api2.bitcoin.cz: Bitcoin returned unparsable JSON

Did I screw up and forget to install something important?