[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]

[BrewCrewFan]

I'll start with the CPU miner.
I'm getting the following error following his instructions, with my launch.bat saved as follows:

minerd -o http://coinotron.com:3334 -O myuser.worker1:workerpassword

"4 miner threads started, using 'scrypt' algorithm.
HTTP requrest failed: coudln't connect to host
json_rpc_call failed, retry afer 30 seconds."


Appreciate the help. This looks fun.

One point:

• Coinotron LTC is Stratum only. You will need to run a stratum proxy program and point your minerd.exe (and your cudaminer.exe when you get to that) to http://170.0.0.1:8332 or http://localhost:8332 (same user/pass info).

This would be my suggestion for the .bat file for your setup on coinotron:
start mining_proxy.exe -pa scrypt -o coinotron.com -p 3334
start cudaminer.exe -i 0 -C 2 -l 30x8 -o http://localhost:8332 -O worker1username:worker1password
start minerd.exe --threads 4 --url http://localhost:8332 --userpass worker2username:worker2password

You will see json_rpc_call failed errors until the mining proxy program is up and running. I suggest running minerd with --threads 3 if you're seeing significantly lower hashrate for your GPU when also running minerd.

Thanks, but I have no idea what Stratum is or want to get into that.

I have been able to get launched on WinCoin for CPU mining. here is my command line (I guessed on # of threads...)

minerd.exe --algo scrypt --s 6 --threads 8 --url http://94.23.199.14:9327/ --

I'm getting like 50 khash/s on a Intel core i-5 2320 @ 3.00 GHZ

What do I put on CudaMiner launch.bat? I was able to launch cgminer like this (didn't change a single setting, don't know what to put.):
cgminer --scrypt -o http://94.23.199.14:9327/ --intensity 10 --worksize 128 -g 1 but only got like 60 Khash a second and it sounded like my video card was going to explode.

I have a Geforce GTX 570. I removed the user/pass parts.

.01 BTC donation to whoever gets me over 150 gpu hash with this pool.
I've found some kind of gdoc but don't know what I'm looking for.

Thanks guys.

https://docs.google.com/spreadsheet/ccc?key=0AjMqJzI7_dCvdG9fZFN1Vjd0WkFOZmtlejltd0JXbmc#gid=2

First you make a folder for the cuda program to unzip to. Unpack it there. You then install all the other stuff that is needed listed in the first post that is also needed. I think there is one called prethreads that I also unpacked into the cuda folder.

Once that is done, open notepad, insert :  cudaminer -o "pooladdy" -O "username.worker":"userpassword". notice the : between the user and pw.
Once you got that done, you save it as Launch.bat....now make sure you that when you save it, you change it from saving as a text doc to All files. Save it to the folder you got your cuda stuff in.
Once you got that, then just double click on the .bat file you just made and it should launch. It will first do an auto tune. just wait and then note what is says... should day XXXkhash at some numberxsome number

[1713884327]

1713884327

[termhn]

I'll start with the CPU miner.
I'm getting the following error following his instructions, with my launch.bat saved as follows:

minerd -o http://coinotron.com:3334 -O myuser.worker1:workerpassword

"4 miner threads started, using 'scrypt' algorithm.
HTTP requrest failed: coudln't connect to host
json_rpc_call failed, retry afer 30 seconds."


Appreciate the help. This looks fun.

One point:

• Coinotron LTC is Stratum only. You will need to run a stratum proxy program and point your minerd.exe (and your cudaminer.exe when you get to that) to http://170.0.0.1:8332 or http://localhost:8332 (same user/pass info).

This would be my suggestion for the .bat file for your setup on coinotron:
start mining_proxy.exe -pa scrypt -o coinotron.com -p 3334
start cudaminer.exe -i 0 -C 2 -l 30x8 -o http://localhost:8332 -O worker1username:worker1password
start minerd.exe --threads 4 --url http://localhost:8332 --userpass worker2username:worker2password

You will see json_rpc_call failed errors until the mining proxy program is up and running. I suggest running minerd with --threads 3 if you're seeing significantly lower hashrate for your GPU when also running minerd.

Thanks, but I have no idea what Stratum is or want to get into that.

I have been able to get launched on WinCoin for CPU mining. here is my command line (I guessed on # of threads...)

minerd.exe --algo scrypt --s 6 --threads 8 --url http://94.23.199.14:9327/ --

I'm getting like 50 khash/s on a Intel core i-5 2320 @ 3.00 GHZ

What do I put on CudaMiner launch.bat? I was able to launch cgminer like this (didn't change a single setting, don't know what to put.):
cgminer --scrypt -o http://94.23.199.14:9327/ --intensity 10 --worksize 128 -g 1 but only got like 60 Khash a second and it sounded like my video card was going to explode.

I have a Geforce GTX 570. I removed the user/pass parts.

.01 BTC donation to whoever gets me over 150 gpu hash with this pool.
I've found some kind of gdoc but don't know what I'm looking for.

Thanks guys.

https://docs.google.com/spreadsheet/ccc?key=0AjMqJzI7_dCvdG9fZFN1Vjd0WkFOZmtlejltd0JXbmc#gid=2

Your threads (for CPU miner) should be 3 when mining with cudaminer on at the same time.

For Cudaminer, you should put this:
cudaminer -o http://94.23.199.14:9327/ -O USERNAME.WORKERNAME:PASSWORD -i 0 -l 30x8

Just that, nothing else. Should be getting upwards of 200 khash/s

[FaT]

Nice work.
Just one feature that would be useful - some kind of intensity equivalent. I don't want to run my card full tilt because of the fan noise.

[BrewCrewFan]

Nice work.
Just one feature that would be useful - some kind of intensity equivalent. I don't want to run my card full tilt because of the fan noise.

You can add the -i 1 flag then.  so after that "cudaminerxxxxxxx" stuff just add the -i 1 and that should help.

[BrewCrewFan]

I'll start with the CPU miner.
I'm getting the following error following his instructions, with my launch.bat saved as follows:

minerd -o http://coinotron.com:3334 -O myuser.worker1:workerpassword

"4 miner threads started, using 'scrypt' algorithm.
HTTP requrest failed: coudln't connect to host
json_rpc_call failed, retry afer 30 seconds."


Appreciate the help. This looks fun.

One point:

• Coinotron LTC is Stratum only. You will need to run a stratum proxy program and point your minerd.exe (and your cudaminer.exe when you get to that) to http://170.0.0.1:8332 or http://localhost:8332 (same user/pass info).

This would be my suggestion for the .bat file for your setup on coinotron:
start mining_proxy.exe -pa scrypt -o coinotron.com -p 3334
start cudaminer.exe -i 0 -C 2 -l 30x8 -o http://localhost:8332 -O worker1username:worker1password
start minerd.exe --threads 4 --url http://localhost:8332 --userpass worker2username:worker2password

You will see json_rpc_call failed errors until the mining proxy program is up and running. I suggest running minerd with --threads 3 if you're seeing significantly lower hashrate for your GPU when also running minerd.

Thanks, but I have no idea what Stratum is or want to get into that.

I have been able to get launched on WinCoin for CPU mining. here is my command line (I guessed on # of threads...)

minerd.exe --algo scrypt --s 6 --threads 8 --url http://94.23.199.14:9327/ --

I'm getting like 50 khash/s on a Intel core i-5 2320 @ 3.00 GHZ

What do I put on CudaMiner launch.bat? I was able to launch cgminer like this (didn't change a single setting, don't know what to put.):
cgminer --scrypt -o http://94.23.199.14:9327/ --intensity 10 --worksize 128 -g 1 but only got like 60 Khash a second and it sounded like my video card was going to explode.

I have a Geforce GTX 570. I removed the user/pass parts.

.01 BTC donation to whoever gets me over 150 gpu hash with this pool.
I've found some kind of gdoc but don't know what I'm looking for.

Thanks guys.

https://docs.google.com/spreadsheet/ccc?key=0AjMqJzI7_dCvdG9fZFN1Vjd0WkFOZmtlejltd0JXbmc#gid=2

Your threads (for CPU miner) should be 3 when mining with cudaminer on at the same time.

For Cudaminer, you should put this:
cudaminer -o http://94.23.199.14:9327/ -O USERNAME.WORKERNAME:PASSWORD -i 0 -l 30x8

Just that, nothing else. Should be getting upwards of 200 khash/s

No disagree. He should Autotune first to see what it says. I know none of my cards got anything with the postings listed.

[Bakemono]

Nice work.
Just one feature that would be useful - some kind of intensity equivalent. I don't want to run my card full tilt because of the fan noise.

+1

The -i flag does little to no difference to me.

Actually it reduces my hashing performance by about 2% and probably reduces desktop lag by the same percent ... 

[BrewCrewFan]

Nice work.
Just one feature that would be useful - some kind of intensity equivalent. I don't want to run my card full tilt because of the fan noise.

+1

The -i flag does little to no difference to me.

Actually it reduces my hashing performance by about 2% and probably reduces desktop lag by the same percent ... 

It does make me wonder if you use a smaller block size if that would help. Its not something I have played too much with yet because I just got this working a few days ago myself... To be honest it takes a lot of playing around to find the right settings for your card itself.

[FaT]

I tried using a block size that reduced my hashrate by 4 x, but the card still ran very hot.

[Schleicher]

I dont see the "No Thanks" option anywhere. Im either blind or retarded.

Screenshot:
https://dl.dropboxusercontent.com/u/34410241/mega-co-nz-screenshot.jpg

[statdude]

Your threads (for CPU miner) should be 3 when mining with cudaminer on at the same time.

For Cudaminer, you should put this:
cudaminer -o http://94.23.199.14:9327/ -O USERNAME.WORKERNAME:PASSWORD -i 0 -l 30x8

Just that, nothing else. Should be getting upwards of 200 khash/s

No disagree. He should Autotune first to see what it says. I know none of my cards got anything with the postings listed.

Thanks.

GPU seemed to be working. Was giving me high hash rates and my computer was running slow like it was working, with a mixture of this error intermittently:

GPU #0: Geforce GTX 570 result does not validate on CPU!

Unfortunately, no Hash is being credited on WinPool. What gives?

Also, what is Autotune?

[BrewCrewFan]

Your threads (for CPU miner) should be 3 when mining with cudaminer on at the same time.

For Cudaminer, you should put this:
cudaminer -o http://94.23.199.14:9327/ -O USERNAME.WORKERNAME:PASSWORD -i 0 -l 30x8

Just that, nothing else. Should be getting upwards of 200 khash/s

No disagree. He should Autotune first to see what it says. I know none of my cards got anything with the postings listed.

Thanks.

GPU seemed to be working. Was giving me high hash rates and my computer was running slow like it was working, with a mixture of this error intermittently:

GPU #0: Geforce GTX 570 result does not validate on CPU!

Unfortunately, no Hash is being credited on WinPool. What gives?

Also, what is Autotune?

You need to drop the -l 30x8 and delete that line. Then autotune should so right off when you start cuda. That error means you got the *x* wrong and is more than it can handle.

[Signus]

How are you guys getting 120-130 khash/s out of GTX 460? Overclocking? Because mine is only doing around 100 khash/s.

It mostly depends on the specs of your card. Each manufacturer has different stock clock settings that can affect your output. Which GTX 460 is it?

[statdude]

Sweet! brew, the reward is yours.

So, I'm hashing ~150 + ~45, how often am i paid?

I wonder why I have to be paid at separate addresses with this pool.

Also, tips to optimize from here?

[BrewCrewFan]

Sweet! brew, the reward is yours.

So, I'm hashing ~150 + ~45, how often am i paid?

I wonder why I have to be paid at separate addresses with this pool.

Also, tips to optimize from here?

Keep your hard earned coins

Well you can use any number of things. the "-i" flag is for intesity... so like you keep your computer on a lot but not using it, you can remove the -i flag, and then when you do need to use it and the lag is tooo bad for you... just open the .bat file and add the -i 1 flag back in and save.
There is also the -C flag which uses the 1D and 2D memory.... so if you want to try the 1D you put in the -C 1 and if the 2D it would be -C 2. I found that it helped on my 660 and not so much on the 450... but the 460 is better than the 450.

once you start seeing a consistent number kernel launch number ( say the 48x2 or whatever ) you can start playing around with combos with the -l flag....you will know when your out of spec because then you start getting those "can not write to cpu" errors .

Then if you really want to have fun and push that extra 15k out... you can then overclock.

As what was said before, just because one person got it on a certain setting for your modle of card by no means your going to get the same results with yours because there are too many factors into it... just like overclocking.

As a side note, if your going to overclock make darn sure your power supply can handle the extra load and that you can keep the card cool... you start hitting 80 c with fan at 100% i would back off some.

[statdude]

Sweet - Thanks again.. I'm sure I'll be back tweaking and setting up more rigs if it's profitable!

I am showing as
146.743 KH/s 0.36249 LTC/d on the Wincoin page.

But, I've not received any transactions. How often should I expect to receive LTC?

[BrewCrewFan]

Sweet - Thanks again.. I'm sure I'll be back tweaking and setting up more rigs if it's profitable!

I am showing as
146.743 KH/s 0.36249 LTC/d on the Wincoin page.

But, I've not received any transactions. How often should I expect to receive LTC?

Sweet - Thanks again.. I'm sure I'll be back tweaking and setting up more rigs if it's profitable!

I am showing as
146.743 KH/s 0.36249 LTC/d on the Wincoin page.

But, I've not received any transactions. How often should I expect to receive LTC?

That depends on the pool. Also, what limits you have set on your account at whatever pool you joined. I think every pool I have joined for any coin has always been set at no auto withdraw ... and that is usually 0.

[Redjuice]

Hi guys,

Was wondering if anyone can direct me to a readme as to how to set up cudaminer with multiple pools for failover.

I've read the thread and the readme, but not sure if this feature has been added yet.

Tyvm!

[beppe]

CUDA programs cannot be started on Remote Desktop. Need to use VNC instead.

Only exception is Tesla on TCC drivers.

damn, i lost hours before discovering this!

[Lacan82]

CUDA programs cannot be started on Remote Desktop. Need to use VNC instead.

Only exception is Tesla on TCC drivers.

damn, i lost hours before discovering this!

can it start if you use the /console switch?

[cbuchner1]

CUDA programs cannot be started on Remote Desktop. Need to use VNC instead.

Only exception is Tesla on TCC drivers.

damn, i lost hours before discovering this!

can it start if you use the /console switch?

Open a DOS prompt with administrator privileges

and run a script like this

Code:

tscon 0 /dest:console
cd "cudaminer directory"
cudaminer.exe arguments blah blah

Works for me, but it temporarily logs you out of your Remote Desktop session
This also gets OpenGL programs to run (these have the same issue with Remote
Desktop. Thank you Microsoft for obstructing competing technologies).