CCminer(SP-MOD) Modded GPU kernels.

[scryptr]

 sp  could you please put the http protocol back in the source code. I can't wallet mine. And djm lyra2v2 is the same. thx    

Well I still can't wallet mine with latest versions.  http protocol failed.

Does it work in the 1.6.6 fork by tvpruvot?

I don't know sp . Do you mean 1.5.64   ?

my version works with solomining (and was tested with...) so it is probably on your hand (wrong rpc port, username or password)

I know someone who is using your older version because the new ones won't work solo mining.
maybe cuda build app files have changed ..because about the same time sp releases aren't working either.
I'm going to try an older version of lyra2v2  of yours djm.

older ? there are 4 releases which are mostly bug correction (and compatibility issues) and none of these changes are related to solo mining.
all are there, and all are working: https://github.com/djm34/ccminer-lyra/releases

regarding solo mining, I don't know what is the default vtc port, I use a custom config file to define which port are used and some other settings.
(make sure as well, that you are using the latest wallet...  )

also I am using --api-bind 0 (as it becomes problematic when there are several instances) and obviously don't use the same port for api-bind and the wallet

edit: there was a changed pushed by pallas to bmw256, however I don't think it would break solo mining, you can probably test that with testnet

Ok thank you djm. The other person had to use your version before ver4 to make it work.
But still something happened , related?, to solo mining about the time your oldest version was released I think. Will try your oldest tomorrow.
But to clarify ... sp r43 solo mining a coin works....now with exactly the same bat file and using r55 or r60 or r62 it won't work.
Will try again tomorrow. thx

I have tested solo mining Quarkcoin with release 64 now on my system.
Command line: ccminer -O name:password -o http://localhost:8372 -a quark --no-gbt
This happened:


ccminer actually crashed:

Unhandled exception at 0x00C99824 in ccminer.exe: Stack cookie instrumentation code detected a stack-based buffer overrun.

SOLO MINING VERTCOIN--

The question was asked a few pages back about solo mining VertCoin.  I was able to set-up and mine VTC with CCminer release dot 66 on Windows.  Here is what I learned:

     1) The VertCoin wallet looks for a file called "vertcoin.conf" BY DEFAULT on startup.
     2) The VertCoin configuration file "vertcoin.conf" only needs to be minimal.
     3) The folder to place "vertcoin.conf" is in ...appdata/roaming/vertcoin in windows, under the login user files.

Here is my vertcoin.conf file:
=======CUT LINE=======
server=1

daemon=1

rpcuser=username

rpcpassword=x

rpcallowip=127.0.0.1
=======CUT LINE=======

Notice that no ports are mentioned here!

Here is my CCminer command line:

ccminer.exe -c solo-vtc.conf

I use a configuration file to launch.  Here is my configuration file, solo-vtc.conf:

========CUT LINE=========
{
   "url" : "127.0.0.1:5888",
   "user" : "username",
   "pass" : "x",
   "algo" : "lyra2v2",
   "devices" : "0",
   "intensity" : "19",
   "cpu-priority" : 5,
   "cpu-mining" : "true",
      
   "api-bind" : "0",
   "statsavg" : 20,
   "quiet" : false,
   "debug" : false,
   "protocol" : false
}
========CUT LINE=========

Leave the cut lines out, but don't forget the curly brackets.  THERE IS A PORT HERE!  It is the default RPCport (5888) for VertCoin.  You may want to lower the cpu-priority or intensity for your hardware.

It works for me.       --scryptr

EDIT:  MINE ON A LAN--

For those of you that have a "work" or "home" computer with a wallet, and one or more mining rigs, the setup is a little different.  In "vertcoin.conf", set the "rpcallowip" line to "rpcallowip=192.168.1.*", or similar if your local net uses a different configuration.  The asterisk is a "wildcard" that will allow any rig on the local net to link to the wallet on the home computer.  On the mining rig(s), for the url use the physical, numerical address of the home computer, i.e. "192.168.1.111:5888".  If you are using a command line only, use "ccminer -a lyra2v2 -o http://192.168.1.111:5888 -u username -p x".  You can get the lan address of the home computer from your router; I set this example to "111".

There may be a looong pause on miner lanch, but if you don't see red, just wait.  Hash results for your cards should appear within 90 seconds.  Of course, you can tweak the command line with performance flags.  If you also mine with a card on the home computer, make sure you use the lan address for the wallet, just as in the previous paragraph, not "localhost" or "127.0.0.1".  There is still a long pause before hash results.


Local GTX 960 card on Win 7 work computer and Linux 750ti mining rig on LAN


The cards will hash until an "accept" is received.  Then your wallet will be 50 VTC fatter.       --scryptr

[1714688478]

1714688478

[1714688478]

1714688478

[squadbox]

Or maybe on windows it's different...

Yeps. On Windows it is extremely simple to set fixed clocks, fans, overclocking, so you can easily have a "benchmark platform".

With a headless Linux system I don't think there's any solution for fixed fans yet, others might know differently. I previously had fixed fans and clocks on Linux, but I perfectly recall that I specifically had to configure/attach a monitor in order to get that working at the time.

to set the fan on linux just test this

https://gist.github.com/squadbox/e5b5f7bcd86259d627ed

Thanks, but I fear it needs a monitor or the x session will not start... Or will it?

NOT ANY MORE--

If you have one of the latest driver packages, you can use nvidia-config to enable cards without monitors to be adjusted:

"sudo nvidia-xconfig -a --cool-bits=28 --allow-empty-initial-configuration" (Source: https://bitcointalk.org/index.php?topic=826901.msg12279696#msg12279696)

The numbeer "28" is a bitsum that signifies clocks, fanspeed, and power controls.

This was in this thread, earlier.  I think that "-a" and the longer "--allow-empty-initial-configuration" are equivalent.

--scryptr

-a is equivalent to --enable-all-gpus, which configures an X screen on every GPU in the system.

I've tested the sctipt on Ubuntu 14.04 with two 750 Ti.

To set fixed fan speeds

Code:

sudo ./cool_gpu2.sh 72

To reset to default

Code:

sudo ./cool_gpu2.sh stop

[scryptr]

Or maybe on windows it's different...

Yeps. On Windows it is extremely simple to set fixed clocks, fans, overclocking, so you can easily have a "benchmark platform".

With a headless Linux system I don't think there's any solution for fixed fans yet, others might know differently. I previously had fixed fans and clocks on Linux, but I perfectly recall that I specifically had to configure/attach a monitor in order to get that working at the time.

to set the fan on linux just test this

https://gist.github.com/squadbox/e5b5f7bcd86259d627ed

Thanks, but I fear it needs a monitor or the x session will not start... Or will it?

NOT ANY MORE--

If you have one of the latest driver packages, you can use nvidia-config to enable cards without monitors to be adjusted:

"sudo nvidia-xconfig -a --cool-bits=28 --allow-empty-initial-configuration" (Source: https://bitcointalk.org/index.php?topic=826901.msg12279696#msg12279696)

The numbeer "28" is a bitsum that signifies clocks, fanspeed, and power controls.

This was in this thread, earlier.  I think that "-a" and the longer "--allow-empty-initial-configuration" are equivalent.

--scryptr

-a is equivalent to --enable-all-gpus, which configures an X screen on every GPU in the system.

I've tested the sctipt on Ubuntu 14.04 with two 750 Ti.

To set fixed fan speeds

Code:

sudo ./cool_gpu2.sh 72

To reset to default

Code:

sudo ./cool_gpu2.sh stop

THANK YOU! --

For correcting me.  I should have read the manual first.         --scryptr

[hashbrown9000]

i've stayed with older nvidia drivers because my system is stable.  can anyone that is running the newest drivers under linux tell me if nvidia-settings works over SSH?  that was one big failure of the older drivers. all configuration of setup and overclocking had to be done locally with a monitor attached, and then the rig could be deployed headless.

[pallas]

tried:

sudo nvidia-xconfig -a --cool-bits=28 --allow-empty-initial-configuration

+ the cool_cpu2.sh script and it works! :-)

(latest drivers)

[sp_]

It may be time for me to use multiple work items on one Lyra2 execution... but I'm worried the increased memory accesses for the matrix (which I've worked to eliminate) will eat me alive.

In my latests fixes I added more memory accesses and speed..
Perhaps this is the opposite of what you are supposed to do when doing a modded kernal, but I think it was fun..
Only latency wankers working for NVIDIA it seems...
Still slow though. You know the opensource is slow, because the market makes it not profitable to mine.
Your optimized kernals might end up here wolf0, And I think they can afford to pay you enough:

Shit, that's massive.

This picture is taken from the inside of the largest bitcoinfarm in the world. (in china)

[sp_]

I will implement a benchmark mode wich will exit after a given interval.

--benchmark 1000

 (run for 1000 seconds and display the average rate in the end.)

This way I can compare small changes in the hashing algorithm with a higher precition.

Also useful for performance regression tests.

Cuda 7.5 seems to run a little bit slower than cuda 7.0.

It will convert your titanx to a 980ti card.  Performance wise..

[pallas]

I will implement a benchmark mode wich will exit after a given interval.

--benchmark 1000

 (run for 1000 seconds and display the average rate in the end.)

This way I can compare small changes in the hashing algorithm with a higher precition.

Also useful for performance regression tests.

Thanks, I was going to ask for something like this or do it myself.
This + fixed fan speed will help a lot.

[bathrobehero]

I will implement a benchmark mode wich will exit after a given interval.

--benchmark 1000

 (run for 1000 seconds and display the average rate in the end.)

This way I can compare small changes in the hashing algorithm with a higher precition.

Also useful for performance regression tests.

Cuda 7.5 seems to run a little bit slower than cuda 7.0.

It will convert your titanx to a 980ti card.  Performance wise..

If it's about precision, can I recommend the benchmark ignoring hashrate results in the first few minutes or so? Giving time for the cards to ramp up (temps/fan speed).

[sp_]

Testing out cuda 7.5 now. Managed to increase the hashrate in the lyra2v2 +50khash on the 750ti with some modification of the code..

But the other algos are down..

quark is bleeding. 5,5 MHASH vs 6.MHASH on the 750ti.

edit:

retuning x11 now. Managed 2860 in cuda7.5(x86)  vs (2940 in 6.5)  getting closer... Low profile 750ti 38w tdp

[impulse2000]

Testing out cuda 7.5 now. Managed to increase the hashrate in the lyra2v2 +50khash on the 750ti with some modification of the code..

But the other algos are down..

quark is bleeding. 5,5 MHASH vs 6.MHASH on the 750ti.

edit:

retuning x11 now. Managed 2860 in cuda7.5(x86)  vs (2940 in 6.5)  getting closer... Low profile 750ti 38w tdp

1.5.66 LAST
ASUS GTX750Ti 1421/2850 = 7Mh QUARK, 3.3Mh X11, 5Mh Lyra2v2

PS. And how about benchmark with all algo? Example: 5min QUARK, 5min X11, 5min Lyra and etc. And small table with result. Or maybe 2min... After some time your have bench in all algo for compare.

[squadbox]

i've stayed with older nvidia drivers because my system is stable.  can anyone that is running the newest drivers under linux tell me if nvidia-settings works over SSH?  that was one big failure of the older drivers. all configuration of setup and overclocking had to be done locally with a monitor attached, and then the rig could be deployed headless.

My rig was configured over ssh completely headless.

[bensam1231]

So I assume pretty much everyone here uses MSI afterburner for tweaking and monitoring (with GPUZ thrown in on the side). Is there a skin that makes the program more usable? I've been using it for years and it's always a PITA scrolling throw the graphs on the right and changing between cards for tweaking. Not to mention the graphs get messed up and aren't always organized by card.

[sp_]

I have merged flipclip's pull request.

Time-limit [seconds] option, and benchmark results with a time-limit. From the Tpruvot fork.

ccminer -a quark --benchmark --time-limit 30 2>result.txt

please test

[pallas]

I have merged flipclip's pull request.

Time-limit [seconds] option, and benchmark results with a time-limit. From the Tpruvot fork.

ccminer -a quark --benchmark --time-limit 30 2>result.txt

please test

something's wrong:

[2015-09-11 09:56:38] GPU #4: Found nonce $D350388E
[2015-09-11 09:56:38] GPU #4: GeForce GTX 970, 16657
[2015-09-11 09:56:38] CPU #4: 16657 kH/s
[2015-09-11 09:56:43] GPU #2: Found nonce $691B5546
[2015-09-11 09:56:45] GPU #4: GeForce GTX 970, 4511
[2015-09-11 09:56:45] CPU #4: 4511 kH/s
[2015-09-11 09:56:51] GPU #2: GeForce GTX 970, 4124
[2015-09-11 09:56:53] GPU #4: Found nonce $DB37B291
[2015-09-11 09:56:53] GPU #4: GeForce GTX 970, 4129
[2015-09-11 09:56:53] CPU #4: 4129 kH/s
[2015-09-11 09:56:54] GPU #0: GeForce GTX 970, 4086
[2015-09-11 09:57:00] GPU #2: GeForce GTX 970, 4088
[2015-09-11 09:57:02] GPU #4: GeForce GTX 970, 4143
[2015-09-11 09:57:02] CPU #4: 4143 kH/s
[2015-09-11 09:57:02] Benchmark: 11.58 MH/s
11576858

only the first gpu hashrate reported is correct, all the others (including the benchmark) are way lower than real.

this is the second run:

[2015-09-11 09:58:44] GPU #0: Found nonce $0BF9370A
[2015-09-11 09:58:44] GPU #0: GeForce GTX 970, 4078
[2015-09-11 09:58:45] GPU #1: GeForce GTX 970, 4165
[2015-09-11 09:58:50] GPU #2: Found nonce $691B5546
[2015-09-11 09:58:51] GPU #1: Found nonce $3BD0F7B9
[2015-09-11 09:58:51] GPU #1: GeForce GTX 970, 4104
[2015-09-11 09:58:52] GPU #0: GeForce GTX 970, 4121
[2015-09-11 09:58:52] Benchmark: 0 H/s
0

the hashrates are reported wrong in non-benchmark mode as well:

[2015-09-11 10:01:58] GPU #0: GeForce GTX 970, 4127
[2015-09-11 10:02:02] GPU #1: GeForce GTX 970, 4221
[2015-09-11 10:02:06] GPU #0: GeForce GTX 970, 4147
[2015-09-11 10:02:06] accepted: 1/1 (100.00%), 8368 kH/s yes!
[2015-09-11 10:02:10] GPU #1: GeForce GTX 970, 4214
[2015-09-11 10:02:10] accepted: 2/2 (100.00%), 8362 kH/s yes!
[2015-09-11 10:02:19] GPU #0: GeForce GTX 970, 4153

[sp_]

try to revert this commit:

https://github.com/sp-hash/ccminer/commit/9450917e60952fa64ba22469dca5017d8624e3e9

You might need to increase the benchmark time to get a bether result.

[pallas]

try to revert this commit:

https://github.com/sp-hash/ccminer/commit/9450917e60952fa64ba22469dca5017d8624e3e9

You might need to increase the benchmark time to get a bether result.

hashrates are reported low on pool side as well! O_o
gonna revert...

EDIT: reverted and hashrate is still low, gonna reboot and see...

EDIT2: rebooting fixed the hashrate, it was some kind of driver/hardware/something problem, never happened before.

[sp_]

But I thnk the last commit don't show the average rate for a period time, but the average rate after the miner has waited for n seconds.

ccminer -a quark --benchmark --time-limit 30 2>result.txt

Mine for 30 seconds and display the last calculated total hashrate..

I want the average rate over the whole interval.

[pallas]

But I thnk the last commit don't show the average rate for a period time, but the average rate after the miner has waited for n seconds.

ccminer -a quark --benchmark --time-limit 30 2>result.txt

Mine for 30 seconds and display the last calculated total hashrate..

I want the average rate over the whole interval.

what about always printing the total average rate when the program exits?

[myagui]

So I assume pretty much everyone here uses MSI afterburner for tweaking and monitoring (with GPUZ thrown in on the side). Is there a skin that makes the program more usable? I've been using it for years and it's always a PITA scrolling throw the graphs on the right and changing between cards for tweaking. Not to mention the graphs get messed up and aren't always organized by card.

bensam1231: Try out Nivida Inspector
You can even build simply batch files that tweak all your cards with different parameters in one go. To satisfy even the most OCD miners out there, you can also build custom batch files for every different algorithm (since different algos allow for different overclock settings).

Here's a sample batch file tweaking one 980 card, undervolted, high fan, high clocks, fixed pstate-2:
nvidiaInspector.exe -setBaseClockOffset:0,0,220 -setgpuclock:0,2,1390 -setMemoryClock:0,2,3505 -setpowertarget:0,120 -setfanspeed:0,80 -forcepstate:0,2

Cheers!