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

[peri]

Thanks Ness

Managed to get around 500khs..... could get a little more, but then its not reliable enough to leave for hours unattended.

http://s5.postimg.org/ljtgb5uw3/16_12_2013_14_25_30.jpg

_http://img703.imageshack.us/img703/8964/dhqi.jpg

[cbuchner1]

Code:

*** CudaMiner for nVidia GPUs by Christian Buchner ***
                     This is version 2013-12-01 (beta)
        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
               Cuda additions Copyright 2013 Christian Buchner
           My donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

[2013-12-15 21:17:20] 1 miner threads started, using 'scrypt' algorithm.
[2013-12-15 21:17:20] Starting Stratum on stratum+tcp://stratum.gentoomen.org:3333
[2013-12-15 21:17:35] Stratum detected new block
[2013-12-15 21:17:35] Stratum detected new block
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti with compute capability 3.0
[2013-12-15 21:17:38] GPU #0: interactive: 1, tex-cache: 1D, single-alloc: 1
[2013-12-15 21:17:38] GPU #0: using launch configuration K14x16
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti, 7168 hashes, 3.25 khash/s
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti, 200704 hashes, 833.73 khash/s
[2013-12-15 21:18:01] GPU #0: GeForce GTX 650 Ti result does not validate on CPU!

anyone know how to fix this problem?

cudaminer 2013-12-10 has a fix for validation issues on Kepler devices

Furthermore I suspect K14x16 requires more memory than the WDDM graphics driver model will allow on your system. cudaminer 2013-12-10 will print a warning if this happens.
Try one of these
a) don't use -C 1 (allowing the card to make a bunch of small memory allocations instead (i.e single-alloc: 0)
b) use a smaller launch config than K14x16. e.g. K14x8
c) upgrade your system RAM (e.g. to 4 GB or 6 GB), allowing larger memory allocations
d) Linux and Windows XP don't have the trouble with the WDDM driver model at all. Switch if you can.
e) autotune for the best launch config on your system instead of doing copy&paste from someone else's config

[cbuchner1]

any progress on the optimizations the cloud miner promised to send us?

wanted to ask about the same ...

keep checking his blog for updates.

He's published his secret sauce. So let's race to integrate it - hopefully we can compensate for the last 3 difficulty increases

[Ness]

any progress on the optimizations the cloud miner promised to send us?

wanted to ask about the same ...

keep checking his blog for updates.

He's published his secret sauce. So let's race to integrate it - hopefully we can compensate for the last 3 difficulty increases

Nice! Hope to see a new build soon

[Laska_Forum]

Anybody had problem with Notebook and Cuda? I have reinstaled all drivers. After that installed new one, I have also installed cuda etc. and i can't mine. Always problem with CUDA. I have GT540M 

[leshow]

Code:

*** CudaMiner for nVidia GPUs by Christian Buchner ***
                     This is version 2013-12-01 (beta)
        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
               Cuda additions Copyright 2013 Christian Buchner
           My donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

[2013-12-15 21:17:20] 1 miner threads started, using 'scrypt' algorithm.
[2013-12-15 21:17:20] Starting Stratum on stratum+tcp://stratum.gentoomen.org:3333
[2013-12-15 21:17:35] Stratum detected new block
[2013-12-15 21:17:35] Stratum detected new block
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti with compute capability 3.0
[2013-12-15 21:17:38] GPU #0: interactive: 1, tex-cache: 1D, single-alloc: 1
[2013-12-15 21:17:38] GPU #0: using launch configuration K14x16
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti, 7168 hashes, 3.25 khash/s
[2013-12-15 21:17:38] GPU #0: GeForce GTX 650 Ti, 200704 hashes, 833.73 khash/s
[2013-12-15 21:18:01] GPU #0: GeForce GTX 650 Ti result does not validate on CPU!

anyone know how to fix this problem?

cudaminer 2013-12-10 has a fix for validation issues on Kepler devices

Furthermore I suspect K14x16 requires more memory than the WDDM graphics driver model will allow on your system. cudaminer 2013-12-10 will print a warning if this happens.
Try one of these
a) don't use -C 1 (allowing the card to make a bunch of small memory allocations instead (i.e single-alloc: 0)
b) use a smaller launch config than K14x16. e.g. K14x8
c) upgrade your system RAM (e.g. to 4 GB or 6 GB), allowing larger memory allocations
d) Linux and Windows XP don't have the trouble with the WDDM driver model at all. Switch if you can.
e) autotune for the best launch config on your system instead of doing copy&paste from someone else's config

im on archlinux actually, and i have 8gb of ram... so that shouldn't be an issue. ill give the new settings a shot

[Notanon]

any progress on the optimizations the cloud miner promised to send us?

wanted to ask about the same ...

keep checking his blog for updates.

He's published his secret sauce. So let's race to integrate it - hopefully we can compensate for the last 3 difficulty increases

Nice. I can then finish mining on Coinotron and send you a 1LTC donation finally.

[cbuchner1]

He's published his secret sauce. So let's race to integrate it - hopefully we can compensate for the last 3 difficulty increases

Nice. I can then finish mining on Coinotron and send you a 1LTC donation finally.

github has the new code in test_kernel.cu currently (kernel prefix is X). It requires the -m 1 option.
still experimenting with it... There's still some work to be done before it can be part of a cudaminer release (i.e. support Compute 3.0 and also the -C option)

What seems apparent already is that there is no 20-30% speed gain. Weird.

On a non overclocked 780Ti I was going from ~440 kHash to 487 kHash  (a 10% improvement)
On a GT 750M I was going from 55 kHash/s to 59 kHash/s (no texture read caching implemented so far)

BTW: 1 LTC is too much of a donation to ask for with today's exchange rates.I need to update the readme file regarding this.

[SavellM]

What seems apparent already is that there is no 20-30% speed gain. Weird.

On a non overclocked 780Ti I was going from ~440 kHash to 487 kHash  (a 10% improvement)
On a GT 750M I was going from 55 kHash/s to 59 kHash/s (no texture read caching implemented so far)

Well a 10% increase is better than a kick in the teeth

Just a question, would it not be worth trying to contact the author and ask him if he can take a quick look and see if there are any other improvements he could make for us?

Lastly, could you release your latest update as like a Beta or something, I would like to test it but cant compile from Github.

And keep it up

[GoldBit89]

according to the spreadsheet this is what i should be getting: i have the 1 GB version

9600 GT 1GB   Win 7 x64   "01.12.2013
"   30-300   99      Non work         1175   1550   MSI                        
9600GT 512mb   Win 7 x64   29/11/2013   290+/-20          

what im getting is 12-13 khash average and 15 khash top using auto tune and using computer at the same time.   
driver 331.65 cuda 5.5/5.0 versions
cudaminer 11/14/13 11/20/13 12/1/13 versions
this config:
\x64\cudaminer.exe -o stratum+tcp://stratum01.hashco.ws:8888 -u user.1 -p forumpost

autotunes to max warps :239 L16x3
[2013-12-17 07:53:10] Stratum detected new block
[2013-12-17 07:53:11] GPU #0: GeForce 9600 GT with compute capability 1.1
[2013-12-17 07:53:11] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2013-12-17 07:53:11] Stratum detected new block
[2013-12-17 07:53:12] GPU #0: Performing auto-tuning (Patience...)
[2013-12-17 07:53:12] GPU #0: maximum warps: 239
[2013-12-17 07:53:17] GPU #0:   13.24 khash/s with configuration L16x3
[2013-12-17 07:53:17] GPU #0: using launch configuration L16x3


where is this extra 200 + khash?                                    

[blackraven1425]

On a 9600GT, I would doubt you can get 200+ khash. Some of the newer, high end cards don't even break that, and yours is a very old, mid range card.

Also, someone completely wrecked the spreadsheet. I restored it to an old version (only a couple hours ago). Can you modify permissions so people can only add and update things they have added themselves?

[GoldBit89]

On a 9600GT, I would doubt you can get 200+ khash. Some of the newer, high end cards don't even break that, and yours is a very old, mid range card.

Also, someone completely wrecked the spreadsheet. I restored it to an old version (only a couple hours ago). Can you modify permissions so people can only add and update things they have added themselves?

thats what i thought too--has always been in the 15 khash range tops

just wanted to make sure i wasnt missing something updated.


Thanks.

[waterbit]

cbuchner1 you were running a gtx 260 before you replaced it. Did you ever find an ideal config for it? Your autotune runs it at 27x3 or 25x3 but I see you had it at 54x3 for some time. Just trying to dial this thing in for the time being before I get a new GPU.

[Notanon]

He's published his secret sauce. So let's race to integrate it - hopefully we can compensate for the last 3 difficulty increases

BTW: 1 LTC is too much of a donation to ask for with today's exchange rates.I need to update the readme file regarding this.

Fair enough.

[dga]

What seems apparent already is that there is no 20-30% speed gain. Weird.

On a non overclocked 780Ti I was going from ~440 kHash to 487 kHash  (a 10% improvement)
On a GT 750M I was going from 55 kHash/s to 59 kHash/s (no texture read caching implemented so far)

Well a 10% increase is better than a kick in the teeth

Just a question, would it not be worth trying to contact the author and ask him if he can take a quick look and see if there are any other improvements he could make for us?

We're chatting.  I suspect that what it's going to come down to is:
  - OK speedups on some cards (5-10%) using Cuda 5.0
  - Decent speedups on some other cards (20%) using Cuda 5.5
  - Really good speedups (80%) on some low-cuda-core-count mobile platforms using Cuda 5.5.

It's a little tricky, though - there's something in my code that really likes Cuda 5.5, and the previous code targets Cuda 5.0 in order to be more widely usable.  Likely nothing that can't be solved, but it may take some work.

[69charger]

What seems apparent already is that there is no 20-30% speed gain. Weird.

On a non overclocked 780Ti I was going from ~440 kHash to 487 kHash  (a 10% improvement)
On a GT 750M I was going from 55 kHash/s to 59 kHash/s (no texture read caching implemented so far)

Well a 10% increase is better than a kick in the teeth

Just a question, would it not be worth trying to contact the author and ask him if he can take a quick look and see if there are any other improvements he could make for us?

We're chatting.  I suspect that what it's going to come down to is:
  - OK speedups on some cards (5-10%) using Cuda 5.0
  - Decent speedups on some other cards (20%) using Cuda 5.5
  - Really good speedups (80%) on some low-cuda-core-count mobile platforms using Cuda 5.5.

It's a little tricky, though - there's something in my code that really likes Cuda 5.5, and the previous code targets Cuda 5.0 in order to be more widely usable.  Likely nothing that can't be solved, but it may take some work.

Thanks to both Christian and Dave for all the hard work!

Anxiously awaiting your masterpiece on two GTX 660's 

[juggs]

I know ~nothing~ about coding this stuff, so forgive me if this is a dumb question.

How hard would it be to have cudaminer display the currently mined scrypt's difficulty, the GPU utilisation, temp and fan speed %ge?

I only ask as I used to get to get temp and fan speed info through nVidia's X Server Settings application (on linux). Now I've installed the latest nVidia drivers and CUDA from source, moved my desktop display to the motherboard onboard graphics and run cudaminer with the non-interactive flag (works fine) that route no longer works to query the information. I've looked around but not been able to find any way to query the activity of the card.

I probably missed something obvious - so all suggestions welcome

[juggs]

And having looked for hours, having posted the above I find a way that should query those things.

/usr/bin/nvidia-smi -a

Sadly it is mostly N/A with my card (GTS 250)

Code:

==============NVSMI LOG==============

Timestamp                           : Wed Dec 18 02:40:11 2013
Driver Version                      : 319.37

Attached GPUs                       : 1
GPU 0000:02:00.0
    Product Name                    : GeForce GTS 250
    Display Mode                    : N/A
    Display Active                  : N/A
    Persistence Mode                : Disabled
    Accounting Mode                 : N/A
    Accounting Mode Buffer Size     : N/A
    Driver Model
        Current                     : N/A
        Pending                     : N/A
    Serial Number                   : N/A
    GPU UUID                        : GPU-83625be0-08b8-2a8c-44c8-82f560e2d9b7
    VBIOS Version                   : 62.92.7E.00.00
    Inforom Version
        Image Version               : N/A
        OEM Object                  : N/A
        ECC Object                  : N/A
        Power Management Object     : N/A
    GPU Operation Mode
        Current                     : N/A
        Pending                     : N/A
    PCI
        Bus                         : 0x02
        Device                      : 0x00
        Domain                      : 0x0000
        Device Id                   : 0x061510DE
        Bus Id                      : 0000:02:00.0
        Sub System Id               : 0x110319DA
        GPU Link Info
            PCIe Generation
                Max                 : N/A
                Current             : N/A
            Link Width
                Max                 : N/A
                Current             : N/A
    Fan Speed                       : 43 %
    Performance State               : N/A
    Clocks Throttle Reasons         : N/A
    Memory Usage
        Total                       : 511 MB
        Used                        : 235 MB
        Free                        : 276 MB
    Compute Mode                    : Default
    Utilization
        Gpu                         : N/A
        Memory                      : N/A
    Ecc Mode
        Current                     : N/A
        Pending                     : N/A
    ECC Errors
        Volatile
            Single Bit           
                Device Memory       : N/A
                Register File       : N/A
                L1 Cache            : N/A
                L2 Cache            : N/A
                Texture Memory      : N/A
                Total               : N/A
            Double Bit           
                Device Memory       : N/A
                Register File       : N/A
                L1 Cache            : N/A
                L2 Cache            : N/A
                Texture Memory      : N/A
                Total               : N/A
        Aggregate
            Single Bit           
                Device Memory       : N/A
                Register File       : N/A
                L1 Cache            : N/A
                L2 Cache            : N/A
                Texture Memory      : N/A
                Total               : N/A
            Double Bit           
                Device Memory       : N/A
                Register File       : N/A
                L1 Cache            : N/A
                L2 Cache            : N/A
                Texture Memory      : N/A
                Total               : N/A
    Retired Pages
        Single Bit ECC              : N/A
        Double Bit ECC              : N/A
        Pending                     : N/A
    Temperature
        Gpu                         : 64 C
    Power Readings
        Power Management            : N/A
        Power Draw                  : N/A
        Power Limit                 : N/A
        Default Power Limit         : N/A
        Enforced Power Limit        : N/A
        Min Power Limit             : N/A
        Max Power Limit             : N/A
    Clocks
        Graphics                    : N/A
        SM                          : N/A
        Memory                      : N/A
    Applications Clocks
        Graphics                    : N/A
        Memory                      : N/A
    Default Applications Clocks
        Graphics                    : N/A
        Memory                      : N/A
    Max Clocks
        Graphics                    : N/A
        SM                          : N/A
        Memory                      : N/A
    Compute Processes               : N/A

That's with cudaminer running - all I have to go on is the temp and fan speed. Doesn't seem to be working that hard really - it used to get hotter and higher fans than that gaming on it.

Any pointer welcome

[69charger]

Does the massive update on GitHub mean a release is close ? 

[cbuchner1]

Does the massive update on GitHub mean a release is close ? 

I present the 2013-12-18 release with all new Kepler, Titan kernels. Huge thanks to David Andersen who came up with a more efficient way to do scrypt mining on the Kepler architecture.

We now use CUDA 5.5, which has increased driver requirements. Be sure to use a recent nVidia driver.

When mining on Kepler devices, please autotune again and report your findings.

Christian