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

[ivanlabrie]

is anyone mining max right now? how much more profitable than doge or yac with these prices?

I don't think MAX is profitable for us nVidia miners anymore, YAC is pretty good even though it lowered a bit, Doge IDK, I am mining microCoins: I must say the high numbers are addicting  

and now? haha just doubled again mcxnow.com

Might be good...I get 203mh/s out of my 780, and that would net me 0,013btc/day. Yac is worse than that right now I think, but tdp might go higher.

EDIT: 204w vs 264w...not too shabby.

[1714605653]

1714605653

[1714605653]

1714605653

[1714605653]

1714605653

[AizenSou]

Hey Christian,

Maxcoin's price has increased 3folds. You see, I told you before.  . It's very profitable now.

[Banin]

hi guys again .

I would like to start YACOIN, but i can't made cudaminer work on the latest version, can anyone help me ?

Got a GTX 760

[yupee]

Hello,

It might sound weird, but I have 3 660 ti cards, which I bought for other CUDA related stuff, and I am not sure what to do with them.

660Ti are excellent miners, also for Yacoin. The three cards together do 10kHash or slightly more.
by all means put them to good use. Pick a 3 slot gaming mainboard, e.g. GA-990FXA-UD3
a cheap Sempron CPU and a 750W PSU and mine away.

you could earn about 10 Euros a day with power costs in the order of 2-3 Euros/day,
getting return of investment on the mainboard and additional components in about a month.

Thank for the reply. I made fast calculation and I got 500-600 euros from alternate. I will have to think if it is worth it.

[Morgahl]

Hey Christian,

Maxcoin's price has increased 3folds. You see, I told you before.  . It's very profitable now.

Glad I sold out at 0.0125 right now lol.

That said it look like a few other exchanges opened up for MAX namely Bter causing a spike. Only worth going back to mining it if it stays stable at these new prices. Switching now and watching it drop while still waiting for coins to mature will suck. The ideal Mining coin is a long term stable coin not one valued at a current day snap shot.

If you already have unexchanged MAXCoins you should be rejoicing though

[cbuchner1]

If you already have unexchanged MAXCoins you should be rejoicing though

rejoice.

now I can order 10 more pizzas online without ever using any cash.

I heard some payment services will allow you to use crypto coins for checkout on sites like amazon or ebay.
This could be interesting, too.

EDIT: why did I sell at 0.31 and not at 0.64?  Hit me with a hammer.

[cbuchner1]

Hello,

It might sound weird, but I have 3 660 ti cards, which I bought for other CUDA related stuff, and I am not sure what to do with them.

by all means put them to good use. Pick a 3 slot gaming mainboard, e.g. GA-990FXA-UD3
a cheap Sempron CPU and a 750W PSU and mine away.

Thank for the reply. I made fast calculation and I got 500-600 euros from alternate. I will have to think if it is worth it.

wait I thought you already had three GTX 660Ti's to spare. The mainboard and stuff should be way cheaper. A case is luxury. Rigs love open air

Christian

[Morgahl]

If you already have unexchanged MAXCoins you should be rejoicing though

rejoice.

now I can order 10 more pizzas online without ever using any cash.

I heard some payment services will allow you to use crypto coins for checkout on sites like amazon or ebay.
This could be interesting, too.

I'd get your sell order in quickish though, total volume out for sale has literally halved in about 10 minutes... if it hit a wall price will rocket for a few minutes then slump lol. timing is key.

EDIT: AAAAAND CRASH!!!

[debido666]

https://mcxnow.com/exchange/MAX

Entertainment, in a sad way. Chat Box.

[cbuchner1]

These boxes aren't called troll boxes for no reason!

CUDA 6.0 RC is available. You have to be a registered developer though
https://developer.nvidia.com/user

compile cudaminer and compare results. the T kernel for keccak might be interesting to check out.
In CUDA 5.5 it gives a rather shocking performance.

Christian

[ManIkWeet]

These boxes aren't called troll boxes for no reason!

CUDA 6.0 RC is available. You have to be a registered developer though
https://developer.nvidia.com/user

compile cudaminer and compare results. the T kernel for keccak might be interesting to check out.
In CUDA 5.5 it gives a rather shocking performance.

Christian

Meh too much questions, can't really say I am THAT much of a developer...

Edit:

Unified Memory -- Simplifies programming by enabling applications to access CPU and GPU memory without the need to manually copy data from one to the other, and makes it easier to add support for GPU acceleration in a wide range of programming languages.

That, scrypt-jane anyone?

[cbuchner1]

Unified memory isn't really helping in terms of speed, only in terms of code simplicity.
I'd rather be in full control, having to write a bit more code though.

I am more interested in possible compiler / code generation improvements...

[gg234]

I am running cudaminer 2014-02-09 (beta) and i am seeing very slow hash rate

card model GeForce GTS 240

Current settings

cudaminer -l K600x16 -a keccak -H 1 -i 1 -C 1 -o stratum+tcp://xxx -u x -p x

[Morgahl]

Unified memory would cause a huge burden on the PCI-x lane. You'd need powered risers and would get severely impacted by boards that split the lanes to different cards when multiple cards are in use

[ManIkWeet]

Unified memory isn't really helping in terms of speed, only in terms of code simplicity.
I'd rather be in full control, having to write a bit more code though.

Hmm yeah I read a bit more, sounds like it "hides" the code to copy from CPU to GPU, rather than extending CPU to GPU.

[cbuchner1]

that K kernel really likes to have an x32 at the end...

but your problem is a different one: The GTS 240 is a "legacy" compute 1.x device, asking for the F kernel (which also supports Compute 1.x)

so -l F600x16 it is..

Christian

[Morgahl]

I am running cudaminer 2014-02-09 (beta) and i am seeing very slow hash rate

card model GeForce GTS 240

Current settings

cudaminer -l K600x16 -a keccak -H 1 -i 1 -C 1 -o stratum+tcp://xxx -u x -p x

Code:

-i 0 -H 1 -l F<...>x16 -o stratum+tcp://xxx -u x -p x

Please replace the <...> in the above with about the highest value you can place without getting an out of memory error when launching CUDAminer.

That is a fairly old card I'm not sure you will see much more then 20MH/s and that's generous.

[Morgahl]

Unified memory isn't really helping in terms of speed, only in terms of code simplicity.
I'd rather be in full control, having to write a bit more code though.

Hmm yeah I read a bit more, sounds like it "hides" the code to copy from CPU to GPU, rather than extending CPU to GPU.

I'm sure control of this is retained but it just adds functionality to easy assigning this for extra compatibility

[chapmanjw]

These boxes aren't called troll boxes for no reason!

CUDA 6.0 RC is available. You have to be a registered developer though
https://developer.nvidia.com/user

compile cudaminer and compare results. the T kernel for keccak might be interesting to check out.
In CUDA 5.5 it gives a rather shocking performance.

Christian

Using CUDA 6.0 with my GeForce GT 650M (1024MB) on OS X (MacBook) seems to have reduced my performance.  With 5.5 I was seeing 70 - 75 Kh/s.  With 6.0 I am seeing 45 - 50 Kh/s. 

With both versions I have been using K4x32 as my configuration.  Any suggestions on other configuration options that might work better with 6.0?

Thanks!

[Morgahl]

These boxes aren't called troll boxes for no reason!

CUDA 6.0 RC is available. You have to be a registered developer though
https://developer.nvidia.com/user

compile cudaminer and compare results. the T kernel for keccak might be interesting to check out.
In CUDA 5.5 it gives a rather shocking performance.

Christian

Using CUDA 6.0 with my GeForce GT 650M (1024MB) on OS X (MacBook) seems to have reduced my performance.  With 5.5 I was seeing 70 - 75 Kh/s.  With 6.0 I am seeing 45 - 50 Kh/s. 

With both versions I have been using K4x32 as my configuration.  Any suggestions on other configuration options that might work better with 6.0?

Thanks!

CUDA 6.0 might need adjusting or tweaking and it's not an official release yet just an RC for people to look at and devs to develop on, there may be issue or adjustments to software needed