CCminer(SP-MOD) Modded GPU kernels.

[sp_]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

[1714705961]

1714705961

[1714705961]

1714705961

[1714705961]

1714705961

[pallas]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

is it different than using "-s"?

[antonio8]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

I'll try and find the version but none of the .60's have worked solo mining for me.

[sp_]

This is a bug T Nelson is good at finding..

Solomine in a testnet with low hashrate. Probobly a division by zero or something in the hashrate statistics.

[chrysophylax]

Hi miners.
Can you please advice, for 750Ti card what is the best Nvidia driver version for linux x64?
in Ubuntu official repo stable is this one -  NVIDIA binary driver - version 340.76


P.S. as i understand from this thread, CUDA must be 6.5.

The easiest, is probably to keep the drivers which get installed when CUDA 6.5 gets installed.

in fedora 20 x64 - the latest of the drivers are very stable ... so it is a situation of whether you want to muck around with settings to get the best one - or ( as suggested ) stay with the drivers that are installed ...

as for it being a 'must' with cuda - 6.5 is the most stable AND the highest hashrate ... technically it will compile and run using the 'other' cuda kits - but not well ... i also have a fedora 22 x64 with cuda 7.0.28 and that is running very stable now using x11 ( and a couple of other algos - though you do get errors with most algos ) but at the loss of hashrate ...

so it doesnt HAVE to be cuda 6.5 - but you are much better off with it ... as cuda 7.0 and 7.5 so far come out with lots of other issues ...

#crysx

[scryptr]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

I'll try and find the version but none of the .60's have worked solo mining for me.

SCANTIME--

Sorry for my ignorance, but I need an explanation of what "scantime" actually does.  I spent more than 24 hours solo-mining, and apparently found 3 or 4 blocks of VTC.  I never was credited the blocks because of the crashes.  So, about the time I should be getting an "accepted" message, something goes wrong.

The miner just crashed.  I didn't know why at the time, but it occured about the same frequency at which I should be finding blocks.

--scryptr

[qqqq]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

I'll try and find the version but none of the .60's have worked solo mining for me.

SCANTIME--

Sorry for my ignorance, but I need an explanation of what "scantime" actually does.  I spent more than 24 hours solo-mining, and apparently found 3 or 4 blocks of VTC.  I never was credited the blocks because of the crashes.  So, about the time I should be getting an "accepted" message, something goes wrong.

The miner just crashed.  I didn't know why at the time, but it occured about the same frequency at which I should be finding blocks.

--scryptr

Same case but one crash is enough for me to understand that has a bug.

From http://cryptomining-blog.com/ - We are experiencing some issues with this release for solo mining, so you might want to stick to 66 or earlier for local wallet mining, the miner seems to be crashing when it finds a block.

[scryptr]

Wich version is it working in?

You can try to set:

LP_SCANTIME   60

instead of 25 in ccminer.cpp

I'll try and find the version but none of the .60's have worked solo mining for me.

SCANTIME--

Sorry for my ignorance, but I need an explanation of what "scantime" actually does.  I spent more than 24 hours solo-mining, and apparently found 3 or 4 blocks of VTC.  I never was credited the blocks because of the crashes.  So, about the time I should be getting an "accepted" message, something goes wrong.

The miner just crashed.  I didn't know why at the time, but it occured about the same frequency at which I should be finding blocks.

--scryptr

Same case but one crash is enough for me to understand that has a bug.

From http://cryptomining-blog.com/ - We are experiencing some issues with this release for solo mining, so you might want to stick to 66 or earlier for local wallet mining, the miner seems to be crashing when it finds a block.

VERSION dot 66--

I was solo-mining with version 66.  There was a string of posts several pages back about solo-mining, that is why I attempted to solo-mine VTC.  Several miners were trying to find a version of CCminer that works solo-mining.  And, I did read the CryptoMining Blog article.

Release dot 66 hashes away at VTC, communicates with the wallet, and crashes on finding a block.  I will also say that even though I used a similar setup for FeatherCoin, I was not able to communicate properly with the wallet, and could not generate hash for FTC.  FTC is really a better candidate for solo-mining at its current difficulty.

--scryptr

[sp_]

Can I see your error.log file on the time of the crash?

[qqqq]

Can I see your error.log file on the time of the crash?

Do your miner has a logger system ? Cause i didn't find some flags like cgminer's 2>logfile.txt
And i can't find any error.log file in folder with miner.

[scryptr]

Can I see your error.log file on the time of the crash?

ERROR.LOG--

I wasn't logging the event.  I just took a quick look at CCminer "help" and I do not see a "log" command, although I do see a "-D" debug comand.  I never bothered to create an "error.log" file, actually.

There are several new command line switches not documented in the "help" output.

I could solo-mine again, but it means going without coin.  I still would like to understand the miner better.       --scryptr

[sp_]

Sorry. It was the minercontrol software I use that produce this file

but you can always run ccminer with >logfile.log in the end. Then all the output to the shell will be to file instead.

[serg_25]

Guys what i'm doing wrong...
git clone https://github.com/sp-hash/ccminer.git
./autogen.sh
./configure.sh
everything is fine here.
but when i try to make:

Code:

/usr/local/cuda/bin/nvcc -gencode=arch=compute_52,code=\"sm_52,compute_52\" -gencode=arch=compute_50,code=\"sm_50,compute_50\" -I/usr/local/cuda/include -I. -O3 -Xcompiler -Wall  --ptxas-options="-v" --maxrregcount=128 -o heavy/heavy.o -c heavy/heavy.cu
nvcc fatal   : Unsupported gpu architecture 'compute_52'
make[2]: *** [heavy/heavy.o] Error 1

i'm using GTX 750Ti card and

Code:

nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2014 NVIDIA Corporation
Built on Thu_Jul_17_21:41:27_CDT_2014
Cuda compilation tools, release 6.5, V6.5.12

Thanks

[sp_]

the 750ti is a compute 50 card

you should use compute_50

[djm34]

Guys what i'm doing wrong...
git clone https://github.com/sp-hash/ccminer.git
./autogen.sh
./configure.sh
everything is fine here.
but when i try to make:

Code:

/usr/local/cuda/bin/nvcc -gencode=arch=compute_52,code=\"sm_52,compute_52\" -gencode=arch=compute_50,code=\"sm_50,compute_50\" -I/usr/local/cuda/include -I. -O3 -Xcompiler -Wall  --ptxas-options="-v" --maxrregcount=128 -o heavy/heavy.o -c heavy/heavy.cu
nvcc fatal   : Unsupported gpu architecture 'compute_52'
make[2]: *** [heavy/heavy.o] Error 1

i'm using GTX 750Ti card and

Code:

nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2014 NVIDIA Corporation
Built on Thu_Jul_17_21:41:27_CDT_2014
Cuda compilation tools, release 6.5, V6.5.12

Thanks

you need to use a cuda 6.5 version which support compute 5.2 (there is another one released before the 900's were out which doesn't support it)

as said in the previous message you can also compile it only for compute 5.0

[t-nelson]

This is a bug T Nelson is good at finding..

Solomine in a testnet with low hashrate. Probobly a division by zero or something in the hashrate statistics.

Someone hook me up with a precise test case and I'll take a look.  Full command line should do it.  Yes, I'd rather mine for you than bother setting up accounts.

[scryptr]

This is a bug T Nelson is good at finding..

Solomine in a testnet with low hashrate. Probobly a division by zero or something in the hashrate statistics.

Someone hook me up with a precise test case and I'll take a look.  Full command line should do it.  Yes, I'd rather mine for you than bother setting up accounts.

WALLET--

You would be solo-mining with your own VTC wallet.  My wallet is not open to the web; lag would likely ruin the setup anyway.

My configuration file for a single GTX 960, saved as solo-vtc.conf:

{
   "url" : "192.168.1.111:5888",
   "user" : "yourname",
   "pass" : "x",
   "algo" : "lyra2v2",
   "devices" : "0",
   "intensity" : "19",
   "cpu-priority" : 5,
   "cpu-mining" : "true",
      
   "api-bind" : "0"
}

My batch file, saved as "solo-vtc.bat":

:loop

ccminer.exe -c solo-vtc.conf

pause
goto loop

My Vertcoin wallet configuration file, saved as "vertcoin.conf" in ...appdata/roaming/vertcoin under the login user's name:

server=1
rpcuser=yourname
rpcpassword=x
rpcallowip=192.168.1.*

The example IP of the server in the miner's configuration file must match the LAN IP address of the computer with the VTC wallet.   You don't need to assign a port in the wallet configuration file, "5888" is the default VTC RPC port.  You must name the wallet configuration file "vertcoin.conf" for it to be found by default.  This particular setup is for the VTC wallet computer, and the GTX 960 on the same computer.  I ponted my other rigs in my LAN at this same computer with no problem in communications.  Each rig crashed at least once.


--scryptr

[sp_]

Submitted a speedup in the Keccak algo.

quark is up 100KHASH on the 970.

[serg_25]

the 750ti is a compute 50 card

you should use compute_50

Thanks for reply
Sorry for noob question but where i have to set this configuration parametr? Or i have to use different drivers?

[sp_]

in the makefile.am comment out the compute_52 like this:


nvcc_ARCH  = -gencode=arch=compute_50,code=\"sm_50,compute_50\"
#nvcc_ARCH += -gencode=arch=compute_52,code=\"sm_52,compute_52\"
#nvcc_ARCH += -gencode=arch=compute_35,code=\"sm_35,compute_35\"
#nvcc_ARCH += -gencode=arch=compute_30,code=\"sm_30,compute_30\"