CudaMiner for OS X

CudaMiner

Please Note: I am no longer actively updating the CudaMiner and ccMiner utilities for OS X. These pages are being kept for posterity, but will no longer be updated.

The last builds of the utilities were compiled with the CUDA Toolkit 6.0. As newer versions of the CUDA Toolkit are released, you have a couple of options if you wish to keep using these utilities:
 

  1. Keep using the 6.0 version. So far, there really haven’t been any benefits to updating to 6.5.
     
  2. Create a symlink to the new version as follows:

    You can create a symlink so that the program can still find the files it needs. Open Terminal (Applications > Utilities) and do the following:
     

    sudo ln -s /Developer/NVIDIA/CUDA-6.5 /Developer/NVIDIA/CUDA-6.0
     

    You will be prompted for your password.
     

    The above command will create a “fake” CUDA-6.0 directory that is actually a pointer to CUDA-6.5.

CudaMiner is a command-line scrypt coin (i.e. Litecoin) miner for Nvidia GPUs developed by Christian Buchner (cbuchner1).

[ANN] cudaMiner – a litecoin mining application

CudaMiner on GitHub

The developer of CudaMiner provides compiled binaries for Windows.  Mac OS X users have previously had to manually compile the source code in order to use CudaMiner, see Mining Litecoin on Mac OS X with CudaMiner.  To make this simpler for Mac OS X users, I am providing a compiled version of the CudaMiner releases for Mac OS X.

This information has been updated for the new CudaMiner for OS X packages that do not require Xcode and some of the other prerequisites previously required. For the older instructions, as well as binaries for Snow Leopard, see CudaMiner for OS X (Old).

Update (2014-05-28): The 2014-02-28 release of CudaMiner has been been updated to work with both the 5.5 and 6.0 versions of the CUDA toolkit.

Update (2014-03-10): The CudaMiner packages have been updated to include versions for 10.7, 10.8, and 10.9 in a single package. Simply run the the appropriate version for your computer.

Prerequisites

CudaMiner for OS X Releases

CudaMiner is compiled separately for 10.7, 10.8, 10.9, and 10.10. The downloads below include binaries for each.

Mac OS X 10.7, 10.8, 10.9, and 10.10

OSX_10.7_48px OSX_10.8_48px OSX_10.9_48px OSX_10.10_48px

2014-02-28 Release Build (CUDA 5.5 and 6.0) Download
Mac OS X 10.7, 10.8, and 10.9

OSX_10.7_48px OSX_10.8_48px OSX_10.9_48px

2014-02-18 Release Build (CUDA 5.5 only) Download

Using CudaMiner

There are a number of command-line parameters available for CudaMiner outlined in its README file.  CudaMiner, however, will attempt to auto-tune its configuration based on your NVIDIA GPU.  To launch CudaMiner using the auto-tune options, simply specify your pool URL and worker credentials as follows:

CudaMiner

Credits

Christian Buchner (cbuchner1): Developer of CudaMiner –  CudaMiner announcement | CudaMiner on GitHub

96 Comments

  1. Thanks for the new version but I get segmentation fault when running auto tune – otherwise it runs fine.

    2014-02-26 06:30:37] Stratum detected new block

    [2014-02-26 06:30:37] GPU #0: GeForce GTX 660M with compute capability 3.0

    [2014-02-26 06:30:37] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0

    [2014-02-26 06:30:37] GPU #0: 32 hashes / 4.0 MB per warp.

    Segmentation fault: 11

    Reply
      • I’m getting a similar error with the new version, ‘Segmentation Fault: 11’, but I have nothing else running in the background. I updated to 10.8.5 recently and the 2014-02-18 version now says asks: ‘Unable to query CUDA driver version! Is an nVidia driver installed?’

        Any suggestions. Thanks for the great work!

        Reply
          • I’ve tried uninstalling/reinstalling Cuda several different ways, including deleting nvidia’s installer paths. I’m still getting a Segmentation fault: 11 error.

            Any thing else I could try? thx!

          • Try doing the following:

            – Open Terminal
            – In Finder, navigate to the CudaMiner folder > deps.
            – In Terminal, type ¢€œcd¢€ and then a space.
            – Drag the ¢€œlib¢€ folder from Finder to the Terminal. It should add the full path to that folder in the command line (i.e. ¢€œcd /Users/¢€¦¢€)
            – Hit enter to change directory to that folder
            – Run this command to replace the specified dylib file: cp /usr/lib/libSystem.B.dylib ./

            After doing that, try running CudaMiner again.

          • ah! cp the libSystem.B.dylib file to Cudaminer dir did the trick. I’m too much of noob to understand why, but thanks a lot for your help! it’s up and running.

  2. Any change to add noob instructions here? Downloaded the applicantion, opened cudaminer and allways get an error and don’t have a clue where to add the pool url or worker credintials….. sorry n00b I know :)

    2014-02-28 22:07:35] 1 miner threads started, using ‘scrypt’ algorithm.

    [2014-02-28 22:07:35] HTTP request failed: Failed to connect to 127.0.0.1 port 9332: Connection refused

    [2014-02-28 22:07:35] json_rpc_call failed, retry after 15 seconds

    Reply
    • When you enter the command into the terminal, add your pool and credentials after cudaminer: ./cudaminer -o stratum+tcp://poolurl -O username:password

      Reply
      • How can I stop those 15sec retry’s to have time to type this command into the termianl?

        Reply
        • Ctrl+C will end the current operation.

          Reply
          • And then it is impossibe to write anyting in the terminal window after ctrl C….

            Last login: Fri Feb 28 22:17:04 on ttys000

            kjartans-imac:~ xxxxxxxxxxxxxxxx$ /Users/xxxxxx/Desktop/CudaMinerOSX-10-9-2014-02-181/cudaminer ; exit;

            *** CudaMiner for nVidia GPUs by Christian Buchner ***

            This is version 2014-02-18 (beta)

            based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler

            Cuda additions Copyright 2013,2014 Christian Buchner

            LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

            BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM

            YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4

            [2014-02-28 22:17:43] 1 miner threads started, using ‘scrypt’ algorithm.

            [2014-02-28 22:17:43] HTTP request failed: Failed to connect to 127.0.0.1 port 9332: Connection refused

            [2014-02-28 22:17:43] json_rpc_call failed, retry after 15 seconds

            ^C

            logout

            [Process completed]

          • Open a new Terminal window. Drag the cudaminer file from Finder to the Terminal window. It will enter the full path to the cudaminer file in the Terminal command line. Then you can enter your parameters after that.

  3. Thanks John. Both new to Mac and mining but finally got it to work. Had the lovely Asteroid and was around 65KH/s with my new GTX780M, now it lookd like I am around 250 :).

    Many thanks.

    Looks right?

    [2014-02-28 22:36:40] accepted: 12/12 (100.00%), 236.96 khash/s (yay!!!)

    [2014-02-28 22:36:40] accepted: 13/13 (100.00%), 236.96 khash/s (yay!!!)

    [2014-02-28 22:36:40] accepted: 14/14 (100.00%), 236.96 khash/s (yay!!!)

    [2014-02-28 22:36:59] GPU #0: GeForce GTX 780M, 246.01 khash/s

    [2014-02-28 22:36:59] accepted: 15/15 (100.00%), 246.01 khash/s (yay!!!)

    [2014-02-28 22:37:04] GPU #0: GeForce GTX 780M, 242.70 khash/s

    [2014-02-28 22:37:04] accepted: 16/16 (100.00%), 242.70 khash/s (yay!!!)

    [2014-02-28 22:37:14] GPU #0: GeForce GTX 780M, 245.64 khash/s

    [2014-02-28 22:37:14] accepted: 17/17 (100.00%), 245.64 khash/s (yay!!!)

    Reply
    • I’d say that’s pretty darn good. Little jealous, only get 72 kh/s on my 650M :-)

      Reply
      • :)

        The heat is more with this cuda miner then the Asteroid obviously… can I turn down the intensity of the worker? GPU temp around 95‚°now.

        Reply
        • :-) Yay, that’s the nature of the beast.

          You can add -i 1 -h 2 to the end of the command. That won’t help a huge amount, but it may help some.

          Reply
  4. I seem to be having trouble… I get the error “Unable to query CUDA driver version! Is an nVidia driver installed?”. I have installed the latest drivers AND tried the previous version of drivers to no avail. I have a GeForce 320M, which is apparently not CUDA enabled, but I’ve used CUDA applications in the past on this same machine. Other people have also recorded that CUDA works with my device… Have any ideas?

    Reply
  5. Installed the cuda toolkit and the driver, when I run I get:

    dyld: Library not loaded: /usr/lib/system/libsystem_asl.dylib

    Any ideas?

    Thanks

    Reply
    • Ooops. I was running 10.9 version on my 10.8.4 osx. Anyway, I got the correct version, and now I’m getting;

      dyld: Library not loaded: /Developer/NVIDIA/CUDA-5.5/lib/libcudart.5.5.dylib

      Referenced from: /Users/abaybas/Downloads/CudaMinerOSX-10-7-2014-02-28/./cudaminer

      Reason: Incompatible library version: cudaminer requires version 5.5.0 or later, but libcudart.5.5.dylib provides version 0.0.0

      Trace/BPT trap: 5

      Thanks

      Reply
      • Make sure you have the 10.8 version of CUDA.

        Reply
        • Excellent! Thanks. I’ll leave this overnight, see how it does. Thanks!

          Reply
  6. Thanks a lot.

    Works perfectly, but i had to disable samples installation in cuda packages in order to install it correctly. (OS X 10.9 + CUDA 5.5.28)

    Go from 80kH/s with asteroid to 250-300kH/s with your command line compiled cudaminer. (GTX 660 TwinFrozr)

    Did you think we can use the inboard graphic cards simultaneously ? (It’s a gigabyte z87 + corei5 Hackintosh)

    Regards

    Reply
    • You could use SGMiner for the non-CUDA GPU. You’d run that first and disable the CUDA GPU before starting CudaMiner.

      Reply
  7. Hi, have had a cuda driver update, thought I’d have another crack at this, got you latest for 10.9 above but got this error:

    [2014-03-11 19:06:58] 1 miner threads started, using ‘scrypt’ algorithm.

    [2014-03-11 19:06:58] HTTP request failed: Failed to connect to 127.0.0.1 port 9332: Connection refused

    [2014-03-11 19:06:58] json_rpc_call failed, retry after 15 seconds

    I did change the details in the “start.sh” file, do I need to make changes elsewhere?

    thanks in advance (once again ;-)

    regards,

    VO

    Reply
    • Whoops, that file wasn’t supposed to be included in the packages (I have updated them). Just run CudaMiner normally: ./cudaminer.10.9 -o stratum+tcp://yada -O username:password

      Reply
      • Hi John,

        thanks for that, ok, that worked, my readout is as follows:

        This is version 2014-02-28 (beta)

        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler

        Cuda additions Copyright 2013,2014 Christian Buchner

        LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

        BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM

        YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4

        [2014-03-11 19:56:40] 1 miner threads started, using ‘scrypt’ algorithm.

        [2014-03-11 19:56:40] Starting Stratum on stratum+tcp://pool1.cryptopoolmining.com:4511

        [2014-03-11 19:56:41] Stratum detected new block

        [2014-03-11 19:56:41] Stratum detected new block

        [2014-03-11 19:58:57] GPU #0: GeForce 9400 with compute capability 1.1

        [2014-03-11 19:58:57] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0

        [2014-03-11 19:58:57] GPU #0: 32 hashes / 4.0 MB per warp.

        [2014-03-11 19:59:09] GPU #0: Performing auto-tuning (Patience…)

        [2014-03-11 19:59:09] GPU #0: maximum total warps (BxW): 32

        [2014-03-11 19:59:12] GPU #0: 3893.27 hash/s with configuration F4x4

        [2014-03-11 19:59:12] GPU #0: using launch configuration F4x4

        [2014-03-11 19:59:13] GPU #0: GeForce 9400, 3.56 khash/s

        [2014-03-11 19:59:21] Stratum detected new block

        [2014-03-11 19:59:21] GPU #0: GeForce 9400, 3.66 khash/s

        [2014-03-11 19:59:41] Stratum detected new block

        [2014-03-11 19:59:41] GPU #0: GeForce 9400, 3.68 khash/s

        I do not think i can squeeze out more (yep it’s an older mac mini (with smc fan control) with only a 256 MB card) i was only getting 2.5-2.7 with asteroid i’ll use this with mac miner (cpu) & & will probably get 20-24 k/hashes / sec not too shabby for my macmini for some reason it is showing as mining at 20-36khs on my pool dashboard, probably will get close to 40-50khs when macminer (cpu) is activated. that’s a big woo-hoo for me! ;-) do you take solarcoin? i could send you some of them…. ;-)

        regards,

        VO

        ps: thanks for you hard work!

        Reply
        • Hmm, do you know if there is an OS X wallet for SolarCoin? Don’t see one on their website.

          Reply
  8. whoops my bad, I forgot all about that, it’s not made yet… say you’re a pretty decent code monkey, you up for a challenge? ;-)

    Reply
    • Never played around with Qt before, installing the tools now to see if I can get it to build. In the codebase, it looks like there were some recent OSX-specific checkins. Looks like a Mac wallet might be close. Will see if there is anything I can do to help :-)

      Reply
  9. Hi , i have this error when i try to run the cudaminer 10.9 ” dyld: Library not loaded: /Developer/NVIDIA/CUDA-5.5/lib/libcudart.5.5.dylib ” im in mavericks 10.9.2 whit cuda driver 5.5.47 , can you halp me ? Thanks !

    Reply
    • I would try reinstalling the CUDA toolkit and then rebooting.

      Reply
      • Thanks.. i download and install the cuda-mac-5.5.28_10.9_64 and now work… but when i try this command : ./cudaminer.10.9 -d i got only this : *** CudaMiner for nVidia GPUs by Christian Buchner ***

        This is version 2014-02-28 (beta)

        based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler

        Cuda additions Copyright 2013,2014 Christian Buchner

        LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm

        BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM

        YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4

        cudaminer.10.9: option requires an argument — d

        Try `cudaminer –help’ for more information.

        Nothing about my GPU..

        but in cuda-z in got all the info of my card ( gtx 780 ) Driver version 8.24.9 310.40.25f01

        Do i need something more ?

        Reply
        • The -d parameter is only needed to specify which GPUs to use. You can leave that out and just do: ./cudaminer.10.9 -o stratum+tcp://URL -O username:password

          Reply
  10. Hi John,

    what a star you are! i recon that solarcoin.org would gift you some coins too! ;-) they have been after someone of your caliber to tackle the wallet for osx issue, good luck!

    Reply
    • HA! We’ll see about that. Did find a few easy thing to fix in building for OSX, however, found a few things I am clueless on. Hopefully StackOverflow will have the answers ;-)

      Reply
    • How does this look? ;-)

      SolarCoin

      I should have a build for you to test here soon, getting some kinks worked out.

      Reply
      • looking good john! ;-) I have notified solarcoin.org of your good work & asked them to consider gifting you some solar coin (or lite/bit coin) when you get a wallet, i’ll send you 10 solarcoin for helping me out! ;-) thanks in advance.

        Reply
    • Alright, I have a build ready for you to try out: http://www.johnchapman.net/?ddownload=3598

      For me, it works on 10.7 and 10.9 (those are the two environments I have to test with). My SLR address is 8PvmL9gNtHUiFKJc6B7sDnfuLTqgX6FrSS. :-)

      Hopefully the Solarcoin folks will let me push the necessary code changes for OSX to the main GitHub branch.

      Reply
  11. Hi John,

    I bought a second card, MSI gtx 760. Everything is running well except it’s running very hot, 79F before i shut it down. Is there anyway in the commands to lower the temperature? Is there software to lower the clock, msi has afterburner but windows only.

    Thanks!

    Reply
    • You could try adding -H 1 to the command. This will offload some of the processing to your CPU. It may not help a lot, but could help some.

      Reply
  12. Hi John, I sent this to the solarcoin.org people:

    I was trying out cudaminer for osx & got it to work via the good work of John Chapman (see here: bottom of page: http://www.johnchapman.net/cudaminer/ ) anyhoo… that aside, I said I’d send him some solarcoins, but of course no osx wallet, so he said, he’d take a look & see if he could get it working… anyways long story short, seeing as he’s working on the wallet for osx, any chance that the community (or solarcoin.org) could gift / send him some solarcoins or litecoins / bitcoins

    they replied back:

    I am CC:ing justin the SolarCoin tech lead on this. There is a 10K bounty for a delivered and community approved solarCoin MacOS client in the forums. I think the thing to do would be to post the issue for discussion there. The 10k is delivered to whomever delivers the client, but a lot of good work has gone into the project as a team effort. In the spirit of a transparent and open community a discussion about sharing the bounty in a public forum might be the best thing to do.

    Once the MacOS client is done please forward it to Justin CC:d above. Thanks for your involvement and volunteer efforts for the project. The SolarCoin community is all the stronger because of great people like yourself.

    regards,

    Nick Gogerty

    Portfolio manager, analyst & author

    +1.858.366.8674

    linkedin:Nick Gogerty

    skype:nickgogerty7166

    Greenwich, Connecticut

    you should get in touch directly, so they can pay you the bounty! ;-)

    Reply
  13. Cannot get it to launch at all.

    iMac:lib bfindlay$ cp /usr/lib/libSystem.B.dylib ./

    iMac:lib bfindlay$ cudaminer

    -bash: cudaminer: command not found

    iMac:lib bfindlay$ /Applications/CudaMinerOSX-2014-02-28/cudaminer.10.9

    dyld: Library not loaded: /Developer/NVIDIA/CUDA-5.5/lib/libcudart.5.5.dylib

    Referenced from: /Applications/CudaMinerOSX-2014-02-28/cudaminer.10.9

    Reason: image not found

    Reply
    • Looks like it cannot find the CUDA toolkit. Try reinstalling it and rebooting.

      Reply
      • Hmmm. Dont really know what that means. I downloaded the osx 10 .9 innstaller. The instructions say to ‘simply run the appropriate version’. Never geard of a Cuda installer before 30 seconds ago. Is it part of the package in the cudaminer disk image?

        Reply
        • In the prerequisites listed on this page, you’ll see the CUDA toolkit listed with a link to get it. Download the 5.5 version for your OS.

          Reply
          • John – thanks for your patience with this. I can now get it to run no problem, but it does nothing but return ‘cudaminer –help’ for more information.

            Is the password:username supposed to be for my ‘worker’ or for logging into the pool website? Also, is -o and -O small ‘ohhh’ and capital ‘ohhhh’ or is the second one a zero?

          • Update – never mind, I think I have it running now! Thanks again…

  14. I got this purring like a kitten with my Geforce GTX 780M @ ~250 kH/s. But is there a way to run this (cudaminer-terminal/GPU) simultaneously as Asteroid (asteroid/Antminer U1)? When I try to do this Asteroid mines but it terminates the cudaminer… [Terminated: 15]

    Reply
    • Launch Asteroid and then in its preferences disable GPU mining. Then close and reopen Asteroid. It should no longer be interacting with the GPU and you can start CudaMiner.

      Reply
      • Tried disabling GPU and reopening and still terminating and/or preventing CudaMiner from starting.

        Not sure if this is contributing but I do get a conflicting USB driver warning message from Asteroid on boot when the Antminers are plugged in. Doesn’t seem to matter if i choose to disable the drivers or not.

        Also seems to only conflicting if Asteroid is actually mining. If it’s just open but not mining it lets CudaMiner run. BUT, if CudaMiner is running and I open (or quit, actually) Asteroid it will terminate CudaMiner mining.

        Befuddled!

        Reply
        • Asteroid ultimately uses CGMiner as its backend. You might try downloading the latest CGMiner (which has no GPU support) and running it from the command line.

          Reply
  15. Hi, FIrst thanks a lot for the port. :)

    I have observed that I get twice the hashing power using the original cudaminer under windows (Bootcamp) than using the port under iOSX. it is a big difference indeed! so I was wondering if there’s anything I can do to rise up the iosx performance.

    Of course I run both with the same set of parameters, but I get twice the hashing with the GTX 680MX.

    Any ideas?

    Cheers,

    Reply
  16. hi,

    First of all thanks a lot for the port and all the work on it :).

    I have noteice that cudaminer on windows (bootcamp) can get up to twice the hashing power out of my GPU than this port on OSX.

    Any idea why this may be happening?

    Thanks!!

    Reply
    • Ultimately the only differences at this point would be OSX vs. Windows and the NVIDIA graphics driver for each. I don’t think there is much that could be done differently from the CudaMiner perspective. Maybe just see if the same launch configuration is being used. If you are just letting it auto tune, on Windows, grab the configuration (something like K32x4 or T16x2, etc). Then add that to the OSX command to see if you get better performance.

      Reply
      • You were right… for some reason the auto-config was not optimizing the settings well for OSX. I can get the same performance in iOS by taking the configuration from Windows…

        That’s interesting… Thanks!!

        Reply
  17. Worked perfectly on my Macbook Pro. No issues whatsoever. Running around 62 khash/s on my Geforce GT 650M. I don’t know how you got it up to 72 :D

    Reply
  18. Any chance of a simple step by step guide please? I downloaded it but when I double click it the OS just tells me there is no application assigned. I know i’m supposed to use terminal, but not sure what! Thank you.

    Reply
    • Hello Hugh,

      The easiest thing to do here is to:

      1. Open Terminal (Applications > Utilities)
      2. Drag the cudaminer executable (cudaminer10.9 for ex) to the Terminal window (this will paste the full path to the executable in the Terminal window)
      3. Hit the space bar and then enter the parameters for cudaminer (like what you see on this page under “Using CudaMiner”)
      4. Hit the enter key to execute the command

      Thanks!

      Reply
      • Thanks John, although I haven’t got it working. It keeps saying “unsupported non-option argument” for various things i’ve typed in. What exactly do I type in after I’ve dragged in the executable and pressed space?

        My pool url is stratum+tcp://usa.tompool.org:3328

        My username is 12MuTr6CnNQfiQz72mJuoVJ94KztPU5qfd

        My password is x

        Thanks for this!

        Reply
        • Hello Hugh,

          You would enter -o stratum+tcp://usa.tompool.org:3328 -O 12MuTr6CnNQfiQz72mJuoVJ94KztPU5qfd:x

          (Those are both “o”‘s, one is lowercase, one is upper).

          So it would look like:

          /path/cudaminer.10.9 -o stratum+tcp://usa.tompool.org:3328 -O 12MuTr6CnNQfiQz72mJuoVJ94KztPU5qfd:x

          Reply
          • Ahh Brilliant! I feel silly now, i could have sworn I tried that! All working now and hashing about 215kh/s with my GTX 660 in my hackintosh. I’ll send you some coins as soon as I have mined some! Many thanks.

  19. Getting an error when trying to open the most recent zip of the pre compiled software. any idea why?

    Reply
  20. Hi John!

    Im getting this error, same as an earlier poster.

    Last login: Tue Apr 22 12:16:31 on ttys000

    user1s-Mac-Pro-3:~ user1$ /Users/user1/Downloads/CudaMinerOSX-2014-02-28/cudaminer.10.8 ; exit;

    dyld: Library not loaded: /Developer/NVIDIA/CUDA-5.5/lib/libcudart.5.5.dylib

    Referenced from: /Users/user1/Downloads/CudaMinerOSX-2014-02-28/cudaminer.10.8

    Reason: image not found

    Trace/BPT trap: 5

    logout

    [Process completed]

    But I have the toolkit installed-however its the new nvidia cuda toolkit 6 and not 5.5, could this be a problem? I cant seem to find 5.5 for osx anywhere to download.

    Reply
  21. Hi John I’m having issues installing cudaminer with the Nvidia CUDA Driver Version: 6.0.37. Is there an updated guide/installer that supports the latest CUDA?

    Thanks

    Reply
  22. BTW this is the error im getting

    Library not loaded: /Developer/NVIDIA/CUDA-5.5/lib/libcudart.5.5.dylib

    Referenced from: /Users/Ed/Desktop/CudaMinerOSX-2014-02-28/cudaminer.10.9

    Reason: image not found

    Trace/BPT trap: 5

    Reply
  23. Hi John, Any word on cudamining with X11 ? And thank you so much for all you do.

    Reply
  24. I get the message “Abort trap: 6” right after stratum authentication every time. Any idea what I’m doing wrong? I’m on OSX 10.9.2 with CUDA 6 and your software both downloaded this morning.

    Reply
    • What is the full command you are entering (mask your username and password)?

      Reply
      • cudaminer -o stratum+tcp://ypool.net:9090 -u username.user1 -p notmyrealpassword

        Reply
      • if it helps, here are the last few lines of output:

        [2014-05-12 13:51:21] 1 miner threads started, using ‘scrypt’ algorithm.

        [2014-05-12 13:51:21] Starting Stratum on stratum+tcp://ypool.net:9090

        [2014-05-12 13:51:21] Stratum session id: 22222222222222222222222222222222

        [2014-05-12 13:51:21] Stratum difficulty set to 512

        Abort trap: 6

        Reply
        • I’m not sure what would cause that. You might post this over on the bitcointalk thread for CudaMiner (see the top of this page for the link).

          Reply
          • I posted as you recommended. FWIW, ccMiner works fine on my Mac.

          • Additional information that I discovered today: Dogecoin mining with cudaminer works fine on my machine on another pool. It’s just ypool.net that seems to throw the abort trap.

  25. I’ve got a 650M, too. Can’t find settings that push it above 60 kh/s without it the miner spamming back “result does not validate on CPU”.

    With CUDA 5.x I ran ” -lK8x32 -m1 -C2 “.

    With CUDA 6.0 it seems to prefer ” -lK4x32 -C2 “.

    If I change it to -C3 or -C4, I get a doubling of hash rate, and –benchmark doesn’t report any errors. But once I try those when connected to a server, I get the “result does not validate” errors.

    Any insights?

    Reply
  26. Hey John,

    Thanks for this article. I was having trouble setting up on OS X with a CUDA based card. I have a Quadro FX 4800 which is technically a compute 1.X card but it still works and I’m getting 51.53khash/s. Must be an update?

    Reply
  27. Thanks!!

    [2014-05-28 01:43:19] accepted: 0/68 (0.00%), 644.83 khash/s (booooo)

    [2014-05-28 01:43:19] accepted: 0/69 (0.00%), 644.83 khash/s (booooo)

    [2014-05-28 01:43:19] accepted: 0/70 (0.00%), 644.83 khash/s (booooo)

    [2014-05-28 01:43:19] GPU #0: GeForce GTX 780, 622.32 khash/s

    [2014-05-28 01:43:20] GPU #0: GeForce GTX 780, 639.99 khash/s

    [2014-05-28 01:43:21] accepted: 0/71 (0.00%), 639.99 khash/s (booooo)

    Reply
  28. Any chance you know about getting this going to 10.10 Yosemite. I understand it’s a beta, but have my machines upgraded already for work, and would like the added efficiency??

    Reply
    • I have updated the 2/28 release to include a binary for 10.10 Yosemite.

      Reply
  29. I keep getting a stratum_recv_line failed and stratum connection interrupted any help?

    Reply
    • I’m not sure what would be going on there except maybe for a malformed URL or something might be up with your pool. You might also ask over on the CudaMiner’s forum: http://www.cudaminers.net/forum/

      Reply
  30. Hi John, When i hash with scrypt jane N13 i would get 7mh/s on my vista and my windows 8.1 Now with N 14 i get 3.5 mh/s on my vista and 2.3 mh/s on windows. Memory on my windows is more. I tried every combo that i could think of in my bat file especially the t 5×24 ect command. Any thoughts why i can’t get 3.5 mh/s on my windows? thank you

    Reply
  31. I was actually able to get this working on OS X 10.10 with the Cuda 6.5 drivers (iMac Geforce GT 755m) which wasn’t even on the list of supported devices but it does have Cuda support. Initially I got a few errors but if you notice & follow the directory paths to the files that it’s missing, the miner is looking for an older version than you might have. Mine was looking for CudaDriver6.0 or something and I had CudaDriver6.5. I simply renamed it to 6.0 and it worked just fine with the latest drivers as long as I renamed the directory for the older code to find the files.

    Reply
  32. can you build it for cuda-6.5 libs

    Reply
    • You can create a symlink so that the program can still find the files it needs. Open Terminal (Applications > Utilities) and do the following:

      sudo ln -s /Developer/NVIDIA/CUDA-6.5 /Developer/NVIDIA/CUDA-6.0

      You will be prompted for your password.

      The above command will create a “fake” CUDA-6.0 directory that is actually a pointer to CUDA-6.5. 6.5 doesn’t offer any benefit for CudaMiner at this point and for many people results in worse performance.

      Reply
  33. Finally I can make it work with CUDA-7.5 under Mac OSX 10.11:

    $ sudo ln -s /Developer/NVIDIA/CUDA-7.5 /Developer/NVIDIA/CUDA-6.0
    $ sudo ln -s /Developer/NVIDIA/CUDA-6.0/lib/libcudart.7.5.dylib /Developer/NVIDIA/CUDA-6.0/lib/libcudart.6.0.dylib

    Reply

Leave a Reply

Pin It on Pinterest

Share This
%d bloggers like this: