[BBR] Boolberry GPU Miner Discussion - AMD & Nvidia - OpenCL & CUDA

[hahaerer01]

Also, this does a little bit better on AMD cards. Well, the 7950 I tested on, at least.

https://ottrbutt.com/tmp/wild_keccak_amd.cl

Hello wolf0. How to use wild_keccak_amd.cl ?

How to optimize the config of solo ? I have solo but found on block for 2 days, 5*270 kards.

The network Difficulty is 166962161187 ,but the default difficult config is 16777215. Should I chang it ?

Thank you

[teknohog]

Hello wolf0. How to use wild_keccak_amd.cl ?

Copy the new kernel in the same directory with the binaries (and the old kernels). Change the kernel name in miner_conf.json to the new one.

How to optimize the config of solo ? I have solo but found on block for 2 days, 5*270 kards.

If you have a thread_delay of 2000 or so in miner_conf.json, there's probably not much more to do. It is typical to have a big variance when solo mining -- sometimes you find two blocks in a short time, sometimes there's a long dry season.

Also, you can probably increase the GPU frequencies from normal. This algorithm is very easy on GPUs, so there is less danger of overheating. RAM speeds are not very important for this.

The network Difficulty is 166962161187 ,but the default difficult config is 16777215. Should I chang it ?

No need to change it. The GPU miner will find shares at the low difficulty, but the network only counts blocks with its own difficulty. Having those low-diff shares is good for checking your GPU works, and calculating the hash rate.

[hahaerer01]

 I have changed the cong,amd.cl is used. I got speed more fasted than old ,from 3.29 to 3.3  .

I set donate to 2%.I hope you release new version to get more fast speed !

Thank you!

[koko2530]

solo 2rig   5xr9280x/rig   adsl internet
get only 1 block from 30hr ago
have something wrong ?
thanks

[kingscrown]

so can we mine now any pool or not yet?

[whyyk7]

I tested amd driver must be 14.6

what version Nvidia driver will be ok?

[tuwenxin]

Now GPU mining ore pool now?

[whyyk7]

Thanks~,I made it ok now.speed up with your cl,~  Thanks !!!!!

[clintar]

You can use my repo with simpleminer updated with cryptozoidberg's fixes so it runs better.  https://github.com/clintar/boolberry-opencl until mbk pulls those changes in.

If you can't get it working following mbk's instructions in the README.md, try

Code:

cmake -DOPENCL_LIBRARY=/usr/lib/libOpenCL.so
make

or

Code:

cmake -DOPENCL_LIBRARY=/usr/lib64/libOpenCL.so
make

that's how ive had to get it compiling on some nvidia machines.

then, after it's started compiling i break with ctrl-c and do

Code:

build/release/src
make -j4 simpleminer

then to get the wild_kecak.cl file for nvidia, do:

Code:

wget https://ottrbutt.com/tmp/wild_keccak_nv.cl -O wild_keccak.cl

or for amd:

Code:

wget https://ottrbutt.com/tmp/wild_keccak_amd.cl -O wild_keccak.cl

then to run it

Code:

./simpleminer --pool-addr cncoin.farm:1111 --login 1GU1ZLjjkeAYyGDkq1NebLcSq4BQA37ZTSqoXG5zfDkQdTjb71Tma8vNsqJ3uxRdrKe89b6josHBF1oxaxovPPYy47vN4CD --pass x --mining-threads 1 --device 0 --remote_scratchpad http://cncoin.farm/scratchpad.bin --local_scratchpad scratchpad.bin

replacing your address, pool info, and paths of course.

[naimc3243]

How does one get this miner to work with multiple GPU's ?.

I have two 7970 cards, --mining-threads=1 or 2 or 3 or 4 gives the same results. Only one card is showing GPU Core activity in GPU-Z.

in the %appdata%\boold\miner_conf.json, the    "device_index": 1, or   "device_index": 0, lets me use either cards but I did not find the syntax that would allow both to work.

Also my log files always contain these two errors,  Do I have any errors in my config file ?
14-Aug-05 20:41:52.995227 [miner 1][OCL 0/1] error:  CL_INVALID_VALUE
14-Aug-05 20:41:52.995227 [miner 1][OCL 0/1] platform index must me between 0
d 4294967295

I get this errors all the time with a miner_conf.json file or without.

Running on Windows 8.1 with AMD 13.2 drivers

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

I did, but it's REALLY dull work. I gave clintar the half-finished code, but he hasn't been able to figure it out yet...

Oh. So, the scratchpad part of BBR really is an annoying issue with implementing a stratum version of a gpu miner for it?

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

I did, but it's REALLY dull work. I gave clintar the half-finished code, but he hasn't been able to figure it out yet...

Oh. So, the scratchpad part of BBR really is an annoying issue with implementing a stratum version of a gpu miner for it?

Oh, no, see the screenshot: https://ottrbutt.com/tmp/bbrcudaminer5.png

Stratum. It's just OpenCL that's the major pain.

Oh ok.

Has anyone try to use Mantle or is that exclusive to gaming coding only?

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

I did, but it's REALLY dull work. I gave clintar the half-finished code, but he hasn't been able to figure it out yet...

Oh. So, the scratchpad part of BBR really is an annoying issue with implementing a stratum version of a gpu miner for it?

Oh, no, see the screenshot: https://ottrbutt.com/tmp/bbrcudaminer5.png

Stratum. It's just OpenCL that's the major pain.

Oh ok.

Has anyone try to use Mantle or is that exclusive to gaming coding only?

I'm no expert on Mantle, but as far as I can tell, it's pretty useless for compute.

From what I read about it, it's just another way to access the core technology for gaming. Probably only API from what I see.

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

I did, but it's REALLY dull work. I gave clintar the half-finished code, but he hasn't been able to figure it out yet...

Oh. So, the scratchpad part of BBR really is an annoying issue with implementing a stratum version of a gpu miner for it?

Oh, no, see the screenshot: https://ottrbutt.com/tmp/bbrcudaminer5.png

Stratum. It's just OpenCL that's the major pain.

Oh ok.

Has anyone try to use Mantle or is that exclusive to gaming coding only?

I'm no expert on Mantle, but as far as I can tell, it's pretty useless for compute.

From what I read about it, it's just another way to access the core technology for gaming. Probably only API from what I see.

Yeah, CUDA is a lot more fun, anyway.

CUDA is coded better than anything that AMD does, just that OpenCL is more universal and AMD compute core works with it better.

I'm in the process of migrating from AMD GPU's to Nvidia GTX 750ti's. Best bang for the buck via power/hash ratio.

[SpeedDemon13]

Hey Wolf0, does the CL patches hash better for pools? Or are they still hashing 50% or less than solo still?

Nothing in the CL will help the stales and crap caused by HTTP mining. Simpleminer is garbage.

Will you give a try to implement it in the cpu multiminer?

I did, but it's REALLY dull work. I gave clintar the half-finished code, but he hasn't been able to figure it out yet...

Oh. So, the scratchpad part of BBR really is an annoying issue with implementing a stratum version of a gpu miner for it?

Oh, no, see the screenshot: https://ottrbutt.com/tmp/bbrcudaminer5.png

Stratum. It's just OpenCL that's the major pain.

Oh ok.

Has anyone try to use Mantle or is that exclusive to gaming coding only?

I'm no expert on Mantle, but as far as I can tell, it's pretty useless for compute.

From what I read about it, it's just another way to access the core technology for gaming. Probably only API from what I see.

Yeah, CUDA is a lot more fun, anyway.

CUDA is coded better than anything that AMD does, just that OpenCL is more universal and AMD compute core works with it better.

I'm in the process of migrating from AMD GPU's to Nvidia GTX 750ti's. Best bang for the buck via power/hash ratio.

Agreed. I think when high end Maxwell comes out, AMD will be out of the mining game. It won't be a fight; it'll be an execution.

AMD suppose to release high destiny ram additions of cards with faster timings, but a little to late in my opinion. Also, they suppose to release something competitive to the Maxwell core in the next couple of months, but anything link to AMD is always a what if. Just like the AMD FX cpu series, what let down. The module concept is cost effective for them, but poorly implemented. It's because of the ATI brand that saved AMD, not the other way around. LOL

I like AMD, but from what is technology advance: GPU=Nvida and CPU=Intel. AMD does have the crown in the APU format, if there was a way to exploit that tech since the GPU and CPU are on the same waffer.

[whyyk7]

You can use my repo with simpleminer updated with cryptozoidberg's fixes so it runs better.  https://github.com/clintar/boolberry-opencl until mbk pulls those changes in.

If you can't get it working following mbk's instructions in the README.md, try

Code:

cmake -DOPENCL_LIBRARY=/usr/lib/libOpenCL.so
make

or

Code:

cmake -DOPENCL_LIBRARY=/usr/lib64/libOpenCL.so
make

that's how ive had to get it compiling on some nvidia machines.

then, after it's started compiling i break with ctrl-c and do

Code:

build/release/src
make -j4 simpleminer
cd ../../..

then to get the wild_kecak.cl file for nvidia, do:

Code:

wget https://ottrbutt.com/tmp/wild_keccak_nv.cl -O wild_keccak.cl

or for amd:

Code:

wget https://ottrbutt.com/tmp/wild_keccak_nv.cl -O wild_keccak.cl

then to run it

Code:

src/simpleminer --pool-addr cncoin.farm:1111 --login 1GU1ZLjjkeAYyGDkq1NebLcSq4BQA37ZTSqoXG5zfDkQdTjb71Tma8vNsqJ3uxRdrKe89b6josHBF1oxaxovPPYy47vN4CD --pass x --mining-threads 1 --device 0 --remote_scratchpad http://cncoin.farm/scratchpad.bin --local_scratchpad scratchpad.bin

replacing your address, pool info, and paths of course.

Can you share compiled miner..
I think share miner can improve much hashes  in pool  and let the price go higher

[clintar]

Can you share compiled miner..
I think share miner can improve much hashes  in pool  and let the price go higher

I've published a windows version. I can't test as I don't have windows. Let me know what you need a binary for.


https://github.com/clintar/boolberry-opencl/releases

[mbk]

Merged all updates from the main branch. Added binary release too. Update if you use pool miner. Included wolf's nvidia optimized kernel. I cannot measure it's performance on nvidia but it gives about +2% on AMD cards too so it's the default kernel now.
https://github.com/mbkuperman/boolberry-opencl