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

[mhps]

Maybe you can post MD5 of the miner here in the announcement message (and put it in a MD5.txt by the side of the zip file) so that others can setup a mirror that downloaders can trust.

what's wrongwith the SHA256 sum on the download page?

OK. I see it's in the first post.

[fruittool]

Is there any more gain to be made with more warps per block?

The increase from 8 to 16 made a few khash/sec difference but i find with -i0 that my system is still responsive. I would expect it to be really laggy if the card was being pushed to its limits.
I would try it myself but my coding skills are still very limited when it comes to cuda and crypto.

[cbuchner1]

The reason I am asking is that I believe the 780 could do nearly double of what a 660Ti does.

looking at the Google Docs spreadsheet I see this:

GeForce GTX 780   Windows 7 Pro x64   11/2/2013   280+      -H 1 -l K48x6 -C 2   

hmm, if he also specified -i 0, I think more than 300 kHash can be achieved.  Also the device does have Compute 3.5 - maybe the Titan kernel would also be worth a try.

Christian

[hhs99]

Im running the benchmark against a 8600GT and I get this. How do I fix it?

./cudaminer -d 1 -i 0 --benchmark
           *** CudaMiner for nVidia GPUs by Christian Buchner ***
                     This is version 2013-11-01 (alpha)
        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
               Cuda additions Copyright 2013 Christian Buchner
           My donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

[2013-11-05 09:09:52] 1 miner threads started, using 'scrypt' algorithm.
[2013-11-05 09:09:52] Binding thread 0 to cpu 0


[2013-11-05 09:11:33] GPU #1: hAÊ with compute capability -247478710.32664
[2013-11-05 09:11:33] GPU #1: interactive: 0, tex-cache: 0 , single-alloc: 0
Segmentation fault (core dumped)

The other card (internal 9100) runs and gets 1.6khash

[darthreaper4]

Using Gigabyte Windforce OC GTX780 Rev 2.0 getting 330+khash/sec
Options : -i 0 -l K42x6 -C 2

Code:

[2013-11-05 18:53:25] GPU #0: GeForce GTX 780, 1653120 hashes, 330.68 khash/s
[2013-11-05 18:53:30] GPU #0: GeForce GTX 780, 1661184 hashes, 330.52 khash/s
[2013-11-05 18:53:35] GPU #0: GeForce GTX 780, 1653120 hashes, 330.45 khash/s
[2013-11-05 18:53:40] GPU #0: GeForce GTX 780, 1653120 hashes, 330.41 khash/s
[2013-11-05 18:53:45] GPU #0: GeForce GTX 780, 1653120 hashes, 330.44 khash/s
[2013-11-05 18:53:50] GPU #0: GeForce GTX 780, 1653120 hashes, 330.56 khash/s
[2013-11-05 18:53:55] GPU #0: GeForce GTX 780, 1653120 hashes, 330.61 khash/s
[2013-11-05 18:54:00] GPU #0: GeForce GTX 780, 1653120 hashes, 330.38 khash/s
[2013-11-05 18:54:05] GPU #0: GeForce GTX 780, 1653120 hashes, 330.48 khash/s
[2013-11-05 18:54:10] GPU #0: GeForce GTX 780, 1653120 hashes, 330.34 khash/s
[2013-11-05 18:54:15] GPU #0: GeForce GTX 780, 1653120 hashes, 330.56 khash/s
[2013-11-05 18:54:20] GPU #0: GeForce GTX 780, 1653120 hashes, 330.41 khash/s
[2013-11-05 18:54:25] GPU #0: GeForce GTX 780, 1653120 hashes, 330.50 khash/s
[2013-11-05 18:54:30] GPU #0: GeForce GTX 780, 1653120 hashes, 330.47 khash/s
[2013-11-05 18:54:35] GPU #0: GeForce GTX 780, 1653120 hashes, 330.37 khash/s
[2013-11-05 18:54:40] GPU #0: GeForce GTX 780, 1653120 hashes, 330.26 khash/s
[2013-11-05 18:54:45] GPU #0: GeForce GTX 780, 1653120 hashes, 330.39 khash/s
[2013-11-05 18:54:50] GPU #0: GeForce GTX 780, 1653120 hashes, 330.33 khash/s
[2013-11-05 18:54:55] GPU #0: GeForce GTX 780, 1653120 hashes, 330.47 khash/s
[2013-11-05 18:55:00] GPU #0: GeForce GTX 780, 1653120 hashes, 330.43 khash/s
[2013-11-05 18:55:05] GPU #0: GeForce GTX 780, 1653120 hashes, 330.41 khash/s
[2013-11-05 18:55:10] GPU #0: GeForce GTX 780, 1653120 hashes, 330.42 khash/s
[2013-11-05 18:55:15] GPU #0: GeForce GTX 780, 1653120 hashes, 330.47 khash/s
[2013-11-05 18:55:20] GPU #0: GeForce GTX 780, 1653120 hashes, 330.20 khash/s

[darthreaper4]

Anyone know how to fix result does not validate on CPU! ?

To be more specific  
DEBUG: got new work in 0 ms
DEBUG: hash <= target
Hash:   0000000aedf02a04de2512c94f2e5ffe2e4bc5732b7eaba033a1fc9520578464
Target: 00000010b07d0000000000000000000000000000000000000000000000000000

cbuchner1 if there is a way to fix this, it might be a breakthrough (Or a bug ). I'm getting the following on my GTX 780.

Code:

[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 56448 hashes, 166.85 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 846720 hashes, 1665 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:52] GPU #0: GeForce GTX 780, 8354304 hashes, 2049 khash/s
[2013-11-05 22:16:52] DEBUG: got new work in 0 ms
[2013-11-05 22:16:57] GPU #0: GeForce GTX 780, 10273536 hashes, 2023 khash/s
[2013-11-05 22:16:57] DEBUG: got new work in 0 ms
[2013-11-05 22:17:02] GPU #0: GeForce GTX 780, 10160640 hashes, 2047 khash/s
[2013-11-05 22:17:02] DEBUG: got new work in 1 ms
[2013-11-05 22:17:07] GPU #0: GeForce GTX 780, 10273536 hashes, 2040 khash/s
[2013-11-05 22:17:07] DEBUG: got new work in 0 ms

[RotoX]

I am getting this error using cudaMiner:
http://shrani.si/f/1J/zA/18PFvNsm/11.jpg
http://shrani.si/f/1N/jk/3m4BYXU4/2.jpg

I tried different pools, different instructions, but it seems it just won't connect to the server. And I am sure it's not an issue with my net/port-forwarding/firewall since all my other miners work without a problem...
Anyone has any idea what might be causing this?

[didjaydisteele]

Take a look at the images below:







Are those appalling to you?

I have a whole collection of them!

[didjaydisteele]

Anyone know how to fix result does not validate on CPU! ?

To be more specific  
DEBUG: got new work in 0 ms
DEBUG: hash <= target
Hash:   0000000aedf02a04de2512c94f2e5ffe2e4bc5732b7eaba033a1fc9520578464
Target: 00000010b07d0000000000000000000000000000000000000000000000000000

cbuchner1 if there is a way to fix this, it might be a breakthrough (Or a bug ). I'm getting the following on my GTX 780.

Code:

[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 56448 hashes, 166.85 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 846720 hashes, 1665 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:52] GPU #0: GeForce GTX 780, 8354304 hashes, 2049 khash/s
[2013-11-05 22:16:52] DEBUG: got new work in 0 ms
[2013-11-05 22:16:57] GPU #0: GeForce GTX 780, 10273536 hashes, 2023 khash/s
[2013-11-05 22:16:57] DEBUG: got new work in 0 ms
[2013-11-05 22:17:02] GPU #0: GeForce GTX 780, 10160640 hashes, 2047 khash/s
[2013-11-05 22:17:02] DEBUG: got new work in 1 ms
[2013-11-05 22:17:07] GPU #0: GeForce GTX 780, 10273536 hashes, 2040 khash/s
[2013-11-05 22:17:07] DEBUG: got new work in 0 ms

Yes there is. Try validating the event.

[Lacan82]

Anyone know how to fix result does not validate on CPU! ?

To be more specific  
DEBUG: got new work in 0 ms
DEBUG: hash <= target
Hash:   0000000aedf02a04de2512c94f2e5ffe2e4bc5732b7eaba033a1fc9520578464
Target: 00000010b07d0000000000000000000000000000000000000000000000000000

cbuchner1 if there is a way to fix this, it might be a breakthrough (Or a bug ). I'm getting the following on my GTX 780.

Code:

[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 56448 hashes, 166.85 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:48] GPU #0: GeForce GTX 780, 846720 hashes, 1665 khash/s
[2013-11-05 22:16:48] DEBUG: got new work in 0 ms
[2013-11-05 22:16:52] GPU #0: GeForce GTX 780, 8354304 hashes, 2049 khash/s
[2013-11-05 22:16:52] DEBUG: got new work in 0 ms
[2013-11-05 22:16:57] GPU #0: GeForce GTX 780, 10273536 hashes, 2023 khash/s
[2013-11-05 22:16:57] DEBUG: got new work in 0 ms
[2013-11-05 22:17:02] GPU #0: GeForce GTX 780, 10160640 hashes, 2047 khash/s
[2013-11-05 22:17:02] DEBUG: got new work in 1 ms
[2013-11-05 22:17:07] GPU #0: GeForce GTX 780, 10273536 hashes, 2040 khash/s
[2013-11-05 22:17:07] DEBUG: got new work in 0 ms

From what I am seeing. it means you are using an incorrect configuration based on the inflated hash rates. CPU validation is to verify the GPU is actually doing work. If you are randomly trying bigger configs when you used -D then yes it will not work. Autotune picks the best valid one.

[Lacan82]

For the life of me don't know why you'd block out your address since people share them to allow donating.   you have :password at the end of that?

also

try using -u address -p password 

[RotoX]

@didjaydisteele
What exactly is your point? It is a litecoin mining pool...

@Lacan82

you have :password at the end of that?

It doesn't matter what you have for pass since that pool doesn't use passwords...

try using -u address -p password

Doesn't work...

[Lacan82]

@didjaydisteele
What exactly is your point? It is a litecoin mining pool...

@Lacan82

you have :password at the end of that?

It doesn't matter what you have for pass since that pool doesn't use passwords...

try using -u address -p password

Doesn't work...

Password doesn't matter, but having :anyrandomthing does

Makes the difference between me connecting and not connecting. I just tested

[RotoX]

I have 123 if your so curious... Doesn't change a thing...

[Schleicher]

I have 123 if your so curious... Doesn't change a thing...

Doesn't work here either.
I think the server is down.
Does it work with http://rav3n.dtdns.net:9327/ ?

[Lacan82]

I have 123 if your so curious... Doesn't change a thing...

Doesn't work here either.
I think the server is down.
Does it work with http://rav3n.dtdns.net:9327/ ?

Code:

C:\Users\Jerec\Downloads\cudaminer-2013-11-01\cudaminer-2013-11-01>cudaminer -o
http://p2pool.org:9327 -i 0 -O LaQ5B3K79BxAttK2wjrmW3PPhmzv2pCkcK:123 -m 1
           *** CudaMiner for nVidia GPUs by Christian Buchner ***
                     This is version 2013-11-01 (alpha)
        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
               Cuda additions Copyright 2013 Christian Buchner
           My donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

[2013-11-05 18:54:33] 1 miner threads started, using 'scrypt' algorithm.
[2013-11-05 18:54:33] Binding thread 0 to cpu 0
[2013-11-05 18:54:34] Starting Stratum on stratum+tcp://p2pool.org:9327
[2013-11-05 18:54:34] Stratum detected new block
[2013-11-05 18:54:34] Stratum detected new block
Ctrl-C
[2013-11-05 18:54:47] workio thread dead, waiting for workers...

C:\Users\Jerec\Downloads\cudaminer-2013-11-01\cudaminer-2013-11-01>cudaminer -o
http://p2pool.org:9327 -i 0 -O LaQ5B3K79BxAttK2wjrmW3PPhmzv2pCkcK -m 1
           *** CudaMiner for nVidia GPUs by Christian Buchner ***
                     This is version 2013-11-01 (alpha)
        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
               Cuda additions Copyright 2013 Christian Buchner
           My donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

Try `cudaminer --help' for more information.

Firewall maybe?


Schleicher's even works.

[Schleicher]

Firewall maybe?

There's two different URLs:
server2.p2pool.org and p2pool.org
The first doesn't work, and the seconds works.

[darthreaper4]

Oh well, those numbers seemed too good to be true anyways.  Well on my GTX 780, Autotune gives T576x1 which gives a pathetic hash rate of 260kh/s.

So I use K42x6 for my Gold Coin mining which gives a hash rate of 330kh/s. TDP is at 75% @ 67C

With the below settings I already mined 3 GLD

Code:

[2013-11-06 08:27:22] 1 miner threads started, using 'scrypt' algorithm.
[2013-11-06 08:27:22] Binding thread 0 to cpu 0
[2013-11-06 08:27:22] GPU #0: GeForce GTX 780 with compute capability 3.5
[2013-11-06 08:27:22] GPU #0: interactive: 0, tex-cache: 2D, single-alloc: 1
[2013-11-06 08:27:22] GPU #0: using launch configuration K42x6
[2013-11-06 08:27:22] GPU #0: GeForce GTX 780, 8064 hashes, 42.44 khash/s
[2013-11-06 08:27:23] GPU #0: GeForce GTX 780, 217728 hashes, 320.17 khash/s
[2013-11-06 08:27:28] GPU #0: GeForce GTX 780, 1604736 hashes, 330.90 khash/s
[2013-11-06 08:27:33] GPU #0: GeForce GTX 780, 1661184 hashes, 329.58 khash/s
[2013-11-06 08:27:38] GPU #0: GeForce GTX 780, 1653120 hashes, 330.64 khash/s
[2013-11-06 08:27:43] GPU #0: GeForce GTX 780, 1661184 hashes, 330.89 khash/s
[2013-11-06 08:27:48] GPU #0: GeForce GTX 780, 1661184 hashes, 330.19 khash/s

Code:

Confirmed Date Type Label Address Amount
TRUE 2013-11-06T04:07:57 Mined 45
TRUE 2013-11-05T18:37:09 Mined 45
TRUE 2013-11-05T17:36:29 Mined 45

[cbuchner1]

hash rate of 330kh/s. TDP is at 75%

The power utilization of 75% is surprising. Are you running with -i 0 or -i 1 flag?

my 660ti has 7 SMX (multiprocessors) and runs best on K14x16

Your GTX 780 has 12 SMX and should run nicely on K24x16 - maybe try that and compare.

The GTX 780 Ti which was announced recently has 13 working SMX.

I need to do something about the lousy Titan kernel

Christian

[darthreaper4]

Well with K24x16 (-i 0 -l K24x16 -C 2) I get 271kh/s @ 63TDP.

I usually run -i 0 -l K42x6 -C 2 which gives me a steady 330kh/s @ 75% TDP. So far mined 6 blocks of GLD with this.

Auto tune keep picking up Titan kernal and gave me T575x1 and T576x1 so far. Both which only give 260kh/s.

When I was getting those ridulously high numbers I was using -i 0 -l K6x42 -C 2 (for 543kh/s @ 25% TDP) and -i 0 -l K6x42 -C 2 (for 2010kh/s @ 25% TDP). Wierd part is K42x42 also gave the same results. Christian, is it maybe we are not utilizing these cards properly? Like if we changed the way the code was calculating the hash, the performance can be more? When its running those crazy numbers, my PC actually slows down (it does not when running K42x6) and I cannot even use the browser.