DiabloMiner GPU Miner

[Gebastos]

Update: Fixed -g

Thanks

[DiabloD3]

Update: Fixed issue #48, fixed possible overflow in outputBuffer

[DiabloD3]

Update: I didn't fix issue #48. Everyone who was getting -56 errors, update to this release and tell me the number

Code:

Spurious CL_INVALID_GLOBAL_OFFSET error

says after it. At least the -56 error won't halt DM anymore.

[DiabloD3]

Update: Forcefully disconnect HTTP on any error

Should fix issues with some pools not properly killing the connection and spamming errors until DM restart.

[DiabloD3]

Update: Force OSX to do -w 64 by default, improve -56 test

[DiabloD3]

Update: When I modularized the code, I forgot to readd the code that sends the miner back to the first pool every hour. Now it does.

[DiabloD3]

Update: Believe it or not, I made it faster.

[.m.]

wow - it really is faster but I can not overclock my 7850 in linux with this miner - can I ?

[.m.]

In case somebody needs overclocking under Linux :

git clone git://git.code.sf.net/p/amdovdrvctrl/code amdovdrvctrl-code
yum -y install wxGTK-devel
copy ADL(amd adl sdk) headers to ./src
make
./amdovdrvctrl-code/Release/AMDOverdriveCtrl 1050.ovdr # nice - 315Mh/s
... 1120.ovdr nice - 337Mh/s

[DiabloD3]

In case somebody needs overclocking under Linux :

git clone git://git.code.sf.net/p/amdovdrvctrl/code amdovdrvctrl-code
yum -y install wxGTK-devel
copy ADL(amd adl sdk) headers to ./src
make
./amdovdrvctrl-code/Release/AMDOverdriveCtrl 1050.ovdr # nice - 315Mh/s
... 1120.ovdr nice - 337Mh/s

This is what I was going to suggest to you. Its the only way to really handle this problem. I'm aware cgminer can overclock automatically, but I don't think thats the right solution.

[PaulC2K]

I've recently started getting 100% hardware errors, and ive no idea what to do.
Ive used driver sweeper to remove the AMD drivers and reinstalled them with reboots each time, and the hw errors tally is the only thing moving.

I've been using the same settings for quite some time, but ive removed them and just used the username/password/port and still get the same thing. I have 2x 6870s and mine each individually, and both are doing the same thing.

Can anyone offer any insight as to what the cause may be?

[DiabloD3]

I've recently started getting 100% hardware errors, and ive no idea what to do.
Ive used driver sweeper to remove the AMD drivers and reinstalled them with reboots each time, and the hw errors tally is the only thing moving.

I've been using the same settings for quite some time, but ive removed them and just used the username/password/port and still get the same thing. I have 2x 6870s and mine each individually, and both are doing the same thing.

Can anyone offer any insight as to what the cause may be?

Sounds like theres something radically wrong with your hardware. Take the video cards completely out, clean the heatsinks, and then firmly reinsert them and make the PCI-E power plugs are inserted firmly and correctly.

[PaulC2K]

Figured i'd give it a try, put each one in separately and powered it up, as well as alternating the PCI-E plugs, and always get hardware errors.
I guess this is their way of saying its over for me and GPU mining.

They've been fine during gaming, so ive no reason to believe they've had a significant failure, but clearly something isnt happy in there. Oh well, thanks for the suggestions anyway.

[DiabloD3]

Figured i'd give it a try, put each one in separately and powered it up, as well as alternating the PCI-E plugs, and always get hardware errors.
I guess this is their way of saying its over for me and GPU mining.

They've been fine during gaming, so ive no reason to believe they've had a significant failure, but clearly something isnt happy in there. Oh well, thanks for the suggestions anyway.

Run DM with -d and see which GPU is giving errors. One might actually have gone faulty. Also, are both fans functioning correctly? And are they both consistently under 85f?

[Elmojo]

I'm sorry to ask what I'm certain has been covered before, but how do I configure Diablo to switch to another pool server if the "main" one fails?
I understand that I can specify multiple pools with commas, but doesn't that mine concurrently?

[DiabloD3]

I'm sorry to ask what I'm certain has been covered before, but how do I configure Diablo to switch to another pool server if the "main" one fails?
I understand that I can specify multiple pools with commas, but doesn't that mine concurrently?

No, it switches on failure.

[Elmojo]

No, it switches on failure.

Oh? Perfect, thanks!

[.m.]

Hi all, after some successfull 1 GPU mining I tried 2 GPUs setup, but DiabloMiner does not recognize the second GPU.
My primary is Gigabyte 7950 now, but the secondary MSI 7850 can not be started.
I used "-D " from 0 to 21. but only "3" gives me access to my primary GPU.
It may be because I installed IntelOCL and Nvidia drivers before ATI drivers.
(Nvidia's oclDeviceQuery gives me "no nvidia" - defaulting to first platform - intel i5)
How can I force "platform" to AMD ?
How can I enable the second GPU (lspci sees it)
Thanks a lot, happy mining !
----
Fedora Core 18 x64, kernel 3.8.4, amd app 2.8, diablo probably 3 weeks old from git (but compiled on older fedora 16 - can this be a problem ?)

[.m.]

...probably something wrong AMD drivers according to "atitweak" thread.

[test@localhost adl3]$ aticonfig --lsa
* 0. 01:00.0 AMD Radeon HD 7900 Series
  1. 02:00.0 AMD Radeon HD 7800 Series

* - Default adapter
[test@localhost adl3]$ aticonfig --adapter=0 --od-enable
AMD Overdrive(TM) enabled
[test@localhost adl3]$ aticonfig --adapter=1 --od-enable
AMD Overdrive(TM) enabled
[test@localhost adl3]$ aticonfig --adapter=0 --odgt

Adapter 0 - AMD Radeon HD 7900 Series
            Sensor 0: Temperature - 42.00 C
[test@localhost adl3]$ aticonfig --adapter=1 --odgt
ERROR - Get temperature failed for Adapter 1 - AMD Radeon HD 7800 Series
[test@localhost adl3]$ ./atitweak -l
ADL_Adapter_Active_Get failed.

[DiabloD3]

Hi all, after some successfull 1 GPU mining I tried 2 GPUs setup, but DiabloMiner does not recognize the second GPU.
My primary is Gigabyte 7950 now, but the secondary MSI 7850 can not be started.
I used "-D " from 0 to 21. but only "3" gives me access to my primary GPU.
It may be because I installed IntelOCL and Nvidia drivers before ATI drivers.
(Nvidia's oclDeviceQuery gives me "no nvidia" - defaulting to first platform - intel i5)
How can I force "platform" to AMD ?
How can I enable the second GPU (lspci sees it)
Thanks a lot, happy mining !
----
Fedora Core 18 x64, kernel 3.8.4, amd app 2.8, diablo probably 3 weeks old from git (but compiled on older fedora 16 - can this be a problem ?)

You're using DM wrong. -D is a list, and the default is all devices that can run OCL.