CudaMiner for OS X (Old)



CudaMiner for OS X has been updated to no longer require Xcode and some other prerequisites listed below. For the updated version, see CudaMiner for OS X.


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 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.

CudaMiner for OS X Releases

CudaMiner is compiled separately for 10.6, 10.7/10.8, and 10.9. Download the appropriate version for your system below.

Mac OS X 10.6

Snow Leopard



Mac OS X 10.7 & 10.8

Lion & Mountain Lion



Mac OS X 10.9




Release Builds
2014-02-18 Release Build Download Download Download
2014-02-09 Release Build Download Download Download
2014-02-07 Release Build Download Download Download
2014-02-06 Release Build

Adds MaxCoin Support

Download Download Download
2014-02-04 Release Build Download Download Download
2014-02-02 Release Build Download Download Download
2013-12-18 Release Build Download Download Download
2013-12-01 Release Build Download Download Invalid download ID.
  • The 5.0 version of the CUDA driver and toolkit for OS X 10.6 only supports 32-bit.


  • Xcode Command Line Tools (download Xcode from the Mac App Store or the Apple Developer website, ensure the Command Line Tools are installed in the Xcode preferences, and then run xcode-select --install  in the Terminal.  Lastly run sudo xcodebuild -license  to agree to the Xcode license)
  • CUDA Toolkit 5.5 for 10.7, 10.8, and 10.9 (install AFTER Xcode, reboot after installing)
  • CUDA Toolkit 5.0 for 10.6 (install AFTER Xcode, reboot after installing)
  • MacPorts

How to Setup CudaMiner

  • Open
  • Install the gcc46 and curl packages using MacPorts:

    For 10.7, 10.8, and 10.9:

    For 10.6:

  • Open your bash shell profile in a text editor, such as vi from the Terminal as follows:

  • Hit the o key to start editing the file.
  • Ensure the following lines are added to the file (for 10.6, replace CUDA-5.5 with CUDA-5.0):


  •  Hit the Escape key to stop editing the file.  Save the file and close vi by entering :wq and hitting the Enter key.


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:



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


  1. Does going through any of these procedures install kexts named foo.tun and foo.tap?

    Because if it does it¢€™s screwing with my VPN application. I need both my VPN and cudaminer but if they can¢€™t co-exist then I¢€™m SOL.

    Another question: Does cudaminer kill launchpad via the trackpad?

    After following the installation instructions above I can no longer use trackpad gestures. I can press F4 on the keyboard but I¢€™m quite attached to my trackpad.

    I¢€™m using a Hackintosh with 32gb ram, 4.2ghz Core i7-3770K, OSX Mavericks 10.9.1, Nvidia GTX-660Ti, etc.. I have plenty of horsepower and have never had any issues, this machine has always run flawlessly.


  2. Thank you for your help. The world needs people like you to become truly free.

    Sorry, I translated with Google, my English is very bad.

    • I am happy you found this useful. Thanks.

  3. I am getting the following error:

    onyx:mine dave$ ./cudaminer

    dyld: Library not loaded: @rpath/libcudart.5.5.dylib

    Referenced from: /Users/dave/mine/./cudaminer

    Reason: image not found

    Trace/BPT trap: 5

    onyx:mine dave$

    • Few questions: what OS are you on? Which version of the CUDA toolkit do you have installed? And did you add the lines to your bash profile and restart

    • David, as discovered in the discussion with Anthony, try re-installing the CUDA toolkit/drivers. Apparently CUDA needs to be installed AFTER Xcode.

  4. Having exactly the same problem. I’m running OSX mavericks and CUDA toolkit 5.5. I’ve added the lines to the bash profile and restarted the Terminal but same error…..

    • Ok, setting up a clean environment to see if I can replicate. Just to verify, you also have Xcode, MacPorts, etc., installed?

    • Oh, and which build are you trying?

  5. Yes i have both Xcode and MacPorts installed. i’ve been trying the latest build (2014-01-22 GitHub Build)

    • Can you try the latest “release” build? I wonder if it’s something with one of the new features the developer is working on (but isn’t yet finished).

      • Ok i’ve just tried the 2013 release build and still the same error:

        dyld: Library not loaded: @rpath/libcudart.5.5.dylib

        Referenced from: /Users/Anthony/Documents/cudaminer-osx-2013-12-18/./cudaminer

        Reason: image not found

        Trace/BPT trap: 5

        • Hmm. I did have an issue in one of my build environments where installing Xcode after installing CUDA caused issues. Try running the CUDA installer again, I’m still working on a clean Mavericks environment, shouldn’t be too much longer.

          • Yes! thank you very much for all your help!

          • So, installing CUDA after Xcode worked?

          • yes it now runs, although it has now returned another error:

            [2014-01-24 17:27:57] JSON decode failed(1): ‘[‘ or ‘{‘ expected near ‘HTTP’

            [2014-01-24 17:27:57] JSON decode failed(1): ‘[‘ or ‘{‘ expected near ‘Date’

            [2014-01-24 17:27:57] …retry after 15 seconds

          • What’s the full command you are running (just mask your username/pwd)? And do you get this with both the GitHub and Release builds?

          • dyn1137-175:cudaminer-osx-2013-12-18 Anthony$ ./cudaminer -o stratum+tcp:// -O username:password

          • Hmm. I’m running the latest Release built at the moment and I am not seeing that error. What happens if you point to a different pool?

          • ok i tried the pool as per your example and it seems to have worked!? i don’t have an account with them, how does this work with the wallet etc?

          • You need to create an account with them and setup your worker accounts. I just tried with port 3334 and it seems to work…

          • Ok everyhing is working perfectly. Thanks a lot for all your help!

  6. Hi John,

    I’m getting this error and I have not seeming a solution around the web:

    Unable to query CUDA driver version! Is an nVidia driver installed?

    CUDA is installed and I reinstalled twice. CUDA shows in the “System Preferences” and it update ones to 5.5.28 and the GPU Driver Version is

    I had tried the 2014-01-24 and the 2013-12-18 and I get the same error. The 2014-01-20 did not work at all (Illegal instruction:4).

    I also try to test the sample from NVIDIA, but I get the following error:

    g++: error: i686: No such file or directory

    g++: error: unrecognized option ‘-arch’

    make[1]: *** [asyncAPI.o] Error 1

    make: *** [0_Simple/asyncAPI/Makefile.ph_build] Error 2

    However, I do see the missing file (it is an image), and I have not clue about the “-arch”

    I have a MacPro1,1 with two NVIDIA 8800 GT and a 7300 GT

    Thanks in advance for your help.

    • Which version of Mac OS X are you running?

      According to the Nvidia web page, I don’t think your 7300 GT will work at all for CUDA-related tasks:

      The 8800 GT cards (which are supported at the Compute 1.X level) won’t work with the latest version of the CUDA drivers (5.5). The latest drivers only work with Compute 2.x and 3.x devices. You could download the 5.0 version of the drivers: The Snow Leopard binaries are compiled with the 5.0 toolkit, so you might give those ones a try once you install these drivers (I’m not sure if they will work or not on newer versions, but worth a shot). You will need to replace the 5.5 with 5.0 in your bash profile (as shown in the above instructions) and restart

      If running the Snow Leopard binaries doesn’t work, you can compile CudaMiner with the 5.0 CUDA drivers for your system. See

  7. Hello, I just tried to use the newest version (I am on Mavericks), but it gives me the following error:

    dyld: Library not loaded: /opt/local/lib/libcurl.4.dylib

    Referenced from: /Users/arnold/cudaminer/./cudaminer

    Reason: Incompatible library version: cudaminer requires version 8.0.0 or later, but libcurl.4.dylib provides version 7.0.0

    I think this is quite weird, says the latest version of libcurl is 7.35 which I have installed via homebrew¢€¦ any idea?

    Thanks in advance!

    • I’m not entirely sure why this happens with the version included with OS X. The only way I was able to get this to work (or even compile) was by installing CURL with MacPorts. I had tried some of the other home-brews to see if I could make the prerequisites simpler, but ultimately CURL from MacPorts was the only one I got to work.

  8. I compiled the latest on a late 2012 iMac (gtx 675gx) and it works fabulously. For a bit. Every so often it randomly wants to freeze up the machine. I’ve tried a few command line options, but here is my current cudaminer -H 0 -i 1 -l K7x16 -C 1 -m 1 -o stratum+tcp:// -O user:pw.

    I’ve also just done cudaminer -l auto -o stratum+tcp:// -O user:pw and it tends to do the same thing, basically the system gradually slows down until it either freezes completely or dumps and reboots. Anyone else run into anything similar?

    • I haven’t experienced anything like this. However, you might do -H 2 to not put any burden on the CPU. The combination of load on the CPU and GPU could have an impact. Worth a shot.

      You might also ask your question on the main thread for CudaMiner:

  9. So, This is really very helpful and I’m at the last Stage. When I run the command to start cudaminer with the pool url and username/password I get this message:

    dyld: Library not loaded: @rpath/libcudart.5.5.dylib

    I’m so close :) Any idea about why this is coming up? I’m running Maverick, by the way


    • Also states this…

      Reason: image not found

      Trace/BPT trap: 5

    • Thomas, check to see if you setup you bash profile correctly and then restart the terminal app (so it loads the profile). If that doesn’t help, try reinstalling the CUDA toolkit.

      • I am very new to this so i would never know how to tell if a bash profile were correctly set up or not. Speak to me as if i were a chld, please :)

        • Walk through the instructions on this page again. Review the part where you add the three lines in the Terminal and make sure there are no typos. Lastly, completely close the terminal before launching it again to run the miner. If that doesn’t work, install the CUDA toolkit again and reboot.

          • I am having the exact same problem (and am in the same childlike state). I am running this on a 10.6.8 machine, with Xcode 3.2.6. One thing I find odd is that I’m not sure if Command Line Tools are installed — I did a full install of XCode (with Unix tools, which I believer contain the Command Line Tools). But when I run the ./cudaminer -o …url instruction in the Terminal I get the “./cudaminer: No such file or directory” message.

          • Hi Tony,

            In Terminal, you need to change directory to the directory containing CudaMiner. For instance, if CudaMiner is in a folder named “CudaMiner” in your Downloads folder, enter this:

            cd ~/Downloads/CudaMiner

            The ~ represents your home directory.

          • John, thanks for the help — you are correct, I just needed to navigate to the folder (cudaminer-osx-10-6-2014-02-09). I have tried opening the app in terminal, but get the message: “Unable to query CUDA driver version! Is an nVidia driver installed?”

            Per the forums I went to the Nvidia site to update the latest driver, still no luck. I confirmed that both of the video cards on my system (a G5, 10.6.8) are compatible with Cudaminer. Not sure what I should do next.

          • I would re-install the CUDA toolkit. This sometimes resolves the inability to find it. And then reboot.

          • Thanks again. I tried re-installing the tookit but no luck. I have changed the CUDA driver versions from 5.0.61, to 5.0.45, to 4.2.10, but none of those have any luck either. I’ll keep on plugging away and let you know if anything works.

          • I’m not sure which CUDA version supports PPC (your G5). I would download 4.2 (the driver and toolkit): And then follow the instructions here to build it manually: (you’d replace 5.5 with 4.2 I think, you can CD to the /Developer/NVIDIA directory to see what the directory should be). The unfortunate part is that I can only find CUDA documentation for the latest version, not the archive.

  10. Hey John,

    I’m running into this error:

    “Unable to query CUDA driver version! Is an nVidia driver installed?”

    I followed your instructions up to

    “./cudaminer -o stratum+tcp://yourpoolurl -O username:password” but when I ran it, I got the error mentioned above.

    My GPU is a nVidia GeForce 9400m, I’m on a late 2009 Macbook Pro that started with Snow Leopard and now I’m on Mavericks. I installed CUDA 5.5 before installing XCODE.

    I’m trying to use this to mine Maxcoin with my GPU. Anyhelp would be greatly appreciated. Also, if you know of any other way to GPU mine Maxcoin on a mac that would be useful too. Thanks!

    • Justin, according to, you have a Compute 1.x device. The 5.5 CUDA drivers do not support those on the Mac side. You could uninstall it, install the CUDA 5.0 drivers and follow the Snow Leopard instructions (should work on 10.9, it’s the driver verso that makes the difference).

      • Hi John,

        I installed CUDA 5.0 and downloaded the Snow Leopard version with Maxcoin support, but when I run it I get this error:

        /Users/justinwu/Downloads/cudaminer-osx-10-6-2014-02-06/cudaminer ; exit;

        justin-wus-mbp:~ justinwu$ /Users/justinwu/Downloads/cudaminer-osx-10-6-2014-02-06/cudaminer ; exit;

        dyld: Library not loaded: /opt/local/lib/libcurl.4.dylib

        Referenced from: /Users/justinwu/Downloads/cudaminer-osx-10-6-2014-02-06/cudaminer

        Reason: Incompatible library version: cudaminer requires version 8.0.0 or later, but libcurl.4.dylib provides version 7.0.0

        Trace/BPT trap: 5



        Also, in terminal, after entering this code:

        sudo port install gcc46 +universal

        It runs for a really long time and it get stuck at:

        Attempting to fetch llvm-3.3.src.tar.gz from

        —> Verifying checksums for llvm-3.3

        —> Extracting llvm-3.3

        —> Applying patches to llvm-3.3

        —> Configuring llvm-3.3

        —> Building llvm-3.3

        and it has been here for over an hour.

        So I basically tried two things, compiling it in Terminal via your instructions and getting stuck at Building llvm-3.3, and then I tried just opening the pre-compiled file in terminal and getting the Incompatible library version: cudaminer requires version 8.0.0 or later, but libcurl.4.dylib provides version 7.0.0 error.



  11. Hi John,

    Great program, and thanks!

    I am running cudaminer 2014-02-09 with two GTX580 cards in my hackintosh running 10.8.5. I am getting great khash rates, but my cards are running very hot and the fans are spinning on full tilt. Is there a way to limit the temperature of both cards to 80c like I can in cgminer? I don’t want to burn these cards out prematurely if I can avoid it.

    Here are my settings:

    ./cudaminer -o stratum+tcp:// -O user.worker:password123


  12. Hi, but where to I add the ./cudaminer -o stratum+tcp://yourpoolurl -O username:password part?

    I followed all your steps but when I open cudaminer it just saying

    HTTP request failed: Failed to connect to port 9332: Connection refused

    json_rpc_call failed, retry after 15 seconds

    i’ve tried inputting the line in to it, and it just keeps repeating them.

    I feel as if you’ve skipped a step for a novice such as myself.

    • In the Terminal, you use that to start CudaMiner.

  13. I get the following error right off the bat when I ran:

    sudo port install gcc46 curl

    I get:

    Warning: The Command Line Tools for Xcode don’t appear to be installed; most ports will likely fail to build.

    Warning: See for more information.

    —> Computing dependencies for gcc46Error: Unable to execute port: can’t read “build.cmd”: Failed to locate ‘make’ in path: ‘/opt/local/bin:/opt/local/sbin:/bin:/sbin:/usr/bin:/usr/sbin’ or at its MacPorts configuration time location, did you move it?

    I followed all the prior steps, except that I installed Xcode from the app store and it just auto-installs it. When I tried to run this:

    xcode-select –install

    I got this error:

    xcode-select: Report or change the path to the active

    Xcode installation for this machine.

    Usage: xcode-select –print-path

    Prints the path of the active Xcode folder

    or: xcode-select –switch

    Sets the path for the active Xcode folder

    or: xcode-select –version

    Prints the version of xcode-select

    Any thoughts?

    • What version of OS X are you running? You might launch Xcode and in the preferences, you may have the option to download the Command Line Tools.

      • Thanks – I’m running 10.8.x.

        I checked the preferences of Xcode and under preferences, Command Line Tools is there and it appears to tell me that it is working (although I have no option to alter it).

        • There should be a download button for the Command Line Tools. Select it or right click it, there should be a way to force it to download.

          • Got it! Mining away – thank you!

  14. OK,

    did all of the required steps, then went to run the cudaminer and got this error:

    ***obscured for security***/Downloads/cudaminer-osx-10-9-2014-02-18/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

    [2014-02-20 18:53:42] Unable to query CUDA driver version! Is an nVidia driver installed?


    [Process completed]

    here is the detail ,of my machine:

    osx mac mini 2009 running osx mavericks

    Processor 2 GHz Intel Core 2 Duo

    Memory 5 GB 1067 MHz DDR3

    Graphics NVIDIA GeForce 9400 256 MB

    any help would be great ;-) thanks in advance!

    regards VO

    • Your 9400 is a Compute 1.x device. The current CUDA 5.5 drivers do not work with Compute 1.x devices. You can try installing CUDA 5.0 and using the Snow Leopard binaries (and follow the instructions on the page that are specific to Snow Leopard). If this doesn’t work, you could either try compiling the miner yourself against the 5.0 driver (see or dual boot with Snow Leopard (or even Windows).

      • Hi, thanks for the advice, I was able to change to 5.5 to 5.0 but on compiling i get these errors and cudaminer will not run Please see:

        my-mac-mini-4:CudaMiner-master my$ ./

        checking build system type… i386-apple-darwin13.0.0

        checking host system type… i386-apple-darwin13.0.0

        checking target system type… i386-apple-darwin13.0.0

        checking for a BSD-compatible install… /usr/local/bin/ginstall -c

        checking whether build environment is sane… yes

        /Users/my/Downloads/CudaMiner-master/missing: Unknown --is-lightweight' option

        Try /Users/my/Downloads/CudaMiner-master/missing –help’ for more information

        configure: WARNING: ‘missing’ script is too old or missing

        checking for a thread-safe mkdir -p… /usr/local/bin/gmkdir -p

        checking for gawk… no

        checking for mawk… no

        checking for nawk… no

        checking for awk… awk

        checking whether make sets $(MAKE)… yes

        checking whether make supports nested variables… yes

        checking whether to enable maintainer-specific portions of Makefiles… no

        checking for style of include used by make… GNU

        checking for gcc… gcc

        checking whether the C compiler works… yes

        checking for C compiler default output file name… a.out

        checking for suffix of executables…

        checking whether we are cross compiling… no

        checking for suffix of object files… o

        checking whether we are using the GNU C compiler… yes

        checking whether gcc accepts -g… yes

        checking for gcc option to accept ISO C89… none needed

        checking whether gcc understands -c and -o together… yes

        checking dependency style of gcc… gcc3

        checking for gcc option to accept ISO C99… -std=gnu99

        checking how to run the C preprocessor… gcc -std=gnu99 -E

        checking for grep that handles long lines and -e… /usr/bin/grep

        checking for egrep… /usr/bin/grep -E

        checking whether gcc -std=gnu99 needs -traditional… no

        checking dependency style of gcc -std=gnu99… gcc3

        checking for ranlib… ranlib

        checking for g++… g++

        checking whether we are using the GNU C++ compiler… yes

        checking whether g++ accepts -g… yes

        checking dependency style of g++… gcc3

        checking for gcc -std=gnu99 option to support OpenMP… -fopenmp

        checking for ANSI C header files… yes

        checking for sys/types.h… yes

        checking for sys/stat.h… yes

        checking for stdlib.h… yes

        checking for string.h… yes

        checking for memory.h… yes

        checking for strings.h… yes

        checking for inttypes.h… yes

        checking for stdint.h… yes

        checking for unistd.h… yes

        checking sys/endian.h usability… no

        checking sys/endian.h presence… no

        checking for sys/endian.h… no

        checking sys/param.h usability… yes

        checking sys/param.h presence… yes

        checking for sys/param.h… yes

        checking syslog.h usability… yes

        checking syslog.h presence… yes

        checking for syslog.h… yes

        checking for sys/sysctl.h… yes

        checking whether be32dec is declared… no

        checking whether le32dec is declared… no

        checking whether be32enc is declared… no

        checking whether le32enc is declared… no

        checking for size_t… yes

        checking for working alloca.h… yes

        checking for alloca… yes

        checking for getopt_long… yes

        checking whether we can compile AVX code… no

        configure: WARNING: The assembler does not support the AVX instruction set.

        checking for json_loads in -ljansson… yes

        checking for pthread_create in -lpthread… yes

        checking for SSL_library_init in -lssl… yes

        checking for EVP_DigestFinal_ex in -lcrypto… yes

        checking for gawk… (cached) awk

        checking for curl-config… /opt/local/bin/curl-config

        checking for the version of libcurl… 7.35.0

        checking for libcurl >= version 7.15.2… yes

        checking whether libcurl is usable… yes

        checking for curl_free… yes

        checking that generated files are newer than configure… done

        configure: creating ./config.status

        config.status: creating Makefile

        config.status: creating compat/Makefile

        config.status: creating compat/jansson/Makefile

        config.status: creating cpuminer-config.h

        config.status: executing depfiles commands

        ed-blundells-mac-mini-4:CudaMiner-master my$ make

        /Applications/ all-recursive

        Making all in compat

        make[3]: Nothing to be done for all-am'.

        gcc -std=gnu99 -DHAVE_CONFIG_H -I. -msse2 -I/opt/local/include -fopenmp -pthread -fno-strict-aliasing -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME -O3 -MT cudaminer-cpu-miner.o -MD -MP -MF .deps/cudaminer-cpu-miner.Tpo -c -o cudaminer-cpu-miner.o test -f ‘cpu-miner.c’ || echo ‘./’`cpu-miner.c

        cpu-miner.c:36:21: fatal error: jansson.h: No such file or directory

        compilation terminated.

        make[2]: *** [cudaminer-cpu-miner.o] Error 1

        make[1]: *** [all-recursive] Error 1

        make: *** [all] Error 2

        my-mac-mini-4:CudaMiner-master my$

        many thanks in advance ;-) your help is greatly appreciated !

        regards VO

          • HI John,

            thanks for that dis that, got it installed & also installed gawk and endian (as endian.h was failing) but I still get this error:

            gcc -std=gnu99 -DHAVE_CONFIG_H -I. -msse2 -I/opt/local/include -fopenmp -pthread -fno-strict-aliasing -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME -O3 -MT cudaminer-cpu-miner.o -MD -MP -MF .deps/cudaminer-cpu-miner.Tpo -c -o cudaminer-cpu-miner.o test -f 'cpu-miner.c' || echo './'cpu-miner.c

            cpu-miner.c:36:21: fatal error: jansson.h: No such file or directory

            compilation terminated.

            make[2]: *** [cudaminer-cpu-miner.o] Error 1

            make[1]: *** [all-recursive] Error 1

            make: *** [all] Error 2

            I have stopped & re-started the mac as well, just to make sure all changes are bedded in.

            sorry to be pain, any further ideas? thanks in advance ;-)



          • Hmm, I wonder if this is due to the 5.0 toolkit being 32-bit (and you are building on a 64-bit machine). On the post about compiling CudaMiner, be sure to follow the steps that had you install the +universal version of the MacPorts stuff for Snow Leopard. Then, before you build, run this:

            perl -pi -w -e 's{"CFLAGS=-O3" "CXXFLAGS=-O3"}{"CFLAGS=-m32" "CXXFLAGS=-m32" "LDFLAGS=-m32"}g' ./

            That should force it to build 32-bit. I have this as part of my Snow Leopard build script.

          • Hello John, all was going well until i typed the wrong command into the Terminal. I have 10.9 and I typed the line: “sudo port install gcc46 +universal” which was meant for 10.6. Is there a way of ‘undoing’ the last command? i should have typed “sudo port install gcc46 curl” which is the correct line for my OSX version. Please help

          • I think if you just run the correct command, the necessary files will get installed. If not, you can always sudo port uninstall gcc46.

  15. Hi John, I had to ‘sudo port clean gcc46 ‘ and do the same for each of the four listed dependencies before it would let me run the ‘sudo port install gcc46 curl’ command. I got it working but now it appears Nvidea do not do a 10.9 driver for my GTX 285!! as i am now getting ‘Unable to query CUDA driver version! Is an nVidia driver installed?’ Its back to trying to get a cpuminer running. bummer! thanks for your help anyway. Good informative site & the easiest for a noob like me to follow.

  16. I followed you intrustions explicitly and encountered the “dyld: Library not loaded: @rpath/libcudart.5.5.dylib” issue as well. The solution for me was to either close the current terminal instance, or run source ~/.bash_profile

    Thanks John!

  17. Hi John,

    I’ve gone through the installation several times, removed/reinstalled cuda +macports with no luck. But I think my probable is basic and related to command line tools. After installing command tools within xcode and then again with amanual d/l, i’m still get an odd error when using –install :

    xcode-select: Error: unknown command option ‘–install’.

    any suggestions??


    • The -install is only for Mavericks. If you have an older version of OS X you’ll need to install the Command Line Tools from the Preferences window in Xcode.

  18. Thanks for the instructions. I managed to set-it up on my iMac!

    But on my MBP it doesn’t seem to work even though I did check and the card GeForce GT 330M is supported. And I did everything the same way as on the iMac.

    When running the command: ./cudaminer -o stratum+tcp:// -O username:password

    I get this: Unable to query CUDA driver version! Is an nVidia driver installed?

    I did reinstall CUDA and I did double check every step. Any suggestions of what else can I try? Any ideas would be greatly appreciated.

    • I realized the 1.x cards are not supported currently on CUDA 5.5. I re-installed it all for 5.0 and changed the bash profile to this:

      export PATH=/Developer/NVIDIA/CUDA-5.0/bin:$PATH


      export PATH=/opt/local/bin:/opt/local/sbin:$PATH

      Now I get this error:

      dyld: Library not loaded: @rpath/libcudart.5.5.dylib

      Referenced from: /Users/iraszl/./cudaminer

      Reason: image not found

      Trace/BPT trap: 5

      I don’t understand why is it trying to load @rpath/libcudart.5.5.dylib, should it not try to load @rpath/libcudart.5.0.dylib?

      How can I fix this please?

        • The MBP’s card is 1.x so it will only run‚ CUDA 5.0. But CUDA 5.0 will only run on Snow Leo which the MBP can’t run because it’s too new. Catch 22? :)

          I guess all I can do is to wait and hope CUDA 5.5 will support 1.x cards, right?

          • I suppose so. You could dual boot Windows with Bootcamp. CudaMiner in Windows should run fine with your Compute 1.x card.

          • Thanks! Which version of Windows do you recommend?

          • Bootcamp on Mavericks supports Windows 7 and 8. CudaMiner will run on both. Which ever one you get your hands on should be fine. I usually opt for the latest and greatest myself (I run Windows 8 in Parallels). I’m probably not the best resource for running CudaMiner in Windows, I would check out the main threat for CudaMiner on Bitcointalk:

          • Thanks again!

Leave a Reply

Pin It on Pinterest

Share This
%d bloggers like this: