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

[rudekid]

Hello all. I've been lurking for a couple of months and now it's time to join in a bit.

Firstly, a thousand thanks Christian for this great piece of software. I'm doing this for a bit of fun with 2 * GTX 780 cards, not a pro miner.

I'm about to get a 1x -> 16x riser to move a card outside of my case (after market, not reference cooler. Doh! Why do they even put SLI connectors on these things when they dump so much heat back into the case that I needed to underclock card #1 to stop the temp running away.)

First tip ... PCI-e bus usage:

EVGA Precision Hardware Monitor has an item on its monitoring graphs called "BUS usage, %". I'm getting 10% bus use at the moment mining Mintcoin in a 16x slot, using "-H 2" to not throw calcs back across the bus to the CPU.

So when I get the riser I'll look again and report back, but I'd assume performance would be squeezed down somewhat. If that's a linear relationship between bus and hash rate, then by my calcs that could drop performance to 63% of what I'm currently getting.

Second tip ... a great web page on GPU power use:

http://www.enthusiastpc.net/articles/00002/1.aspx

Very informative, it explains what power the PCI-e bus can provide, the reasons for 6 pin and 8 pin GPU power connectors and much more.

[phillipsaur]

So I'm having problems running two 750 Ti on Ubuntu Mini at the moment. It spits out "Unable to query CUDA driver version! Is an nVidia driver installed?". I had them working together fine 2 days ago before the update but I decided to reinstall Linux for funnies and now it doesnt run with more than 1 card

[bigjme]

Are you sure it runs with 1 card now? You may not have set cud a back up again properly

[ManIkWeet]

I am seeing no keccak hashrate improvement with the 28-2-14 build, on my GTX 780

[ManiacMiner]

Here's my tentative roadmap for continued development. Note the lack of a time table

1) fix the issue with PCIe bandwidth, so people with 1x risers are more happy. CUDA can do all the
    hash evaluation so we save a LOT of bus bandwidth.

2) Add blake256 for blakecoin and related coins.

3) look at implementing some failover options

4) see if I can get a monitoring API included

Hi Christian, can add to your pretty miner momentum protocol as this miner https://github.com/reorder/pts-1gh from reorder? On my 750Ti i have 860 cpm, other miners eg. dga issue +/- 630 cpm.

[djm34]

Here's my tentative roadmap for continued development. Note the lack of a time table

1) fix the issue with PCIe bandwidth, so people with 1x risers are more happy. CUDA can do all the
    hash evaluation so we save a LOT of bus bandwidth.

2) Add blake256 for blakecoin and related coins.

3) look at implementing some failover options

4) see if I can get a monitoring API included

Hi Christian, can add to your pretty miner momentum protocol as this miner https://github.com/reorder/pts-1gh from reorder? On my 750Ti i have 860 cpm, other miners eg. dga issue +/- 630 cpm.

Don't understand what you're saying... (If I take individually each words I understand, but the full sentence absolutely no clue except that you are speaking about pts...)

[cbuchner1]

I guess he's asking about adding the Momentum specific extensions to the stratum protol, plus a CUDA implementation of a Momentum miner. What I did not quite understand is where from I am supposed to take the momentum implementation from. DGA's cuda code? or port the OpenCL miner from the given link?

Anyway, that's too much work  NEXT!

I do take code submissions for new features. So if anyone really needs this, submit the required code please

Christian

[cbuchner1]

I am seeing no keccak hashrate improvement with the 28-2-14 build, on my GTX 780

switch from K to T kernel and lower from x32 to x24

[ManIkWeet]

I am seeing no keccak hashrate improvement with the 28-2-14 build, on my GTX 780

switch from K to T kernel.

Got ya, does autotune work properly yet?
Because it looks like it wants to do 4000 warps...

[djm34]

I am seeing no keccak hashrate improvement with the 28-2-14 build, on my GTX 780

switch from K to T kernel.

Got ya, does autotune work properly yet?
Because it looks like it wants to do 4000 warps...

You need to use K#cudacore x24

[ManIkWeet]

I am seeing no keccak hashrate improvement with the 28-2-14 build, on my GTX 780

switch from K to T kernel.

Got ya, does autotune work properly yet?
Because it looks like it wants to do 4000 warps...

You need to use K#cudacore x24

I use T2048x24, seems to get me 205mhash average

[ManiacMiner]

I guess he's asking about adding the Momentum specific extensions to the stratum protol, plus a CUDA implementation of a Momentum miner. What I did not quite understand is where from I am supposed to take the momentum implementation from. DGA's cuda code? or port the OpenCL miner from the given link?

Anyway, that's too much work  NEXT!

I do take code submissions for new features. So if anyone really needs this, submit the required code please

Christian

I wanted to say that code from Reorder miner much better optimized and work, then code from DGA`s and others miners for CUDA.

[guzzzi]

Im very new to this Mining thing. At the moment i got 2x Gainward gtx780ti SLI Cards.
My Problem is dont know how many kh they really give. At the moment with my Settings i just get 610kh/s with both together, not each.

Possible that anyone give me some Hints to improve this, if its possible.

If i use settings like T2048x24 my cudaminer 32bit version breaks totaly. Or i get a lot of cudaminer error 6.

At the moment i run T14x18.

Many thanks.

[ManIkWeet]

Im very new to this Mining thing. At the moment i got 2x Gainward gtx780ti SLI Cards.
My Problem is dont know how many kh they really give. At the moment with my Settings i just get 610kh/s with both together, not each.

Possible that anyone give me some Hints to improve this, if its possible.

If i use settings like T2048x24 my cudaminer 32bit version breaks totaly. Or i get a lot of cudaminer error 6.

At the moment i run T14x18.

Many thanks.

Pretty sure you need something like T15x24 for optimal scrypt on that card

[djm34]

Im very new to this Mining thing. At the moment i got 2x Gainward gtx780ti SLI Cards.
My Problem is dont know how many kh they really give. At the moment with my Settings i just get 610kh/s with both together, not each.

Possible that anyone give me some Hints to improve this, if its possible.

If i use settings like T2048x24 my cudaminer 32bit version breaks totaly. Or i get a lot of cudaminer error 6.

At the moment i run T14x18.

Many thanks.

Pretty sure you need something like T15x24 for optimal scrypt on that card

2048 is for keccak... it won't work with other algo.

[Severaz]

Price just went up $10 today on newegg for all the ti's once there in your cart for the ones that are left..

ANYONE HAVE HASH RATES FOR THE NON TI VERSION OF THE 750??

[djm34]

Price just went up $10 today on newegg for all the ti's once there in your cart for the ones that are left..

we should start mining gtx750ti, do we have an algo ?

[Severaz]

Price just went up $10 today on newegg for all the ti's once there in your cart for the ones that are left..

we should start mining gtx750ti, do we have an algo ?

or just buy stock

[debido666]

Checking Out the PCI Express x1-x16 USB 3.0 Extenders
http://cryptomining-blog.com/1153-checking-out-the-pci-express-x1-x16-usb-3-0-extenders/

[cbuchner1]

Price just went up $10 today on newegg for all the ti's once there in your cart for the ones that are left..

ANYONE HAVE HASH RATES FOR THE NON TI VERSION OF THE 750??

I have two Palit 750 StormX OC devices at home, but it will be a while before I can connect them.

Christian