CODEWITHSUNDEEP
javawindows-764-bitnvidiajcuda
duffymo
duffymo

Reputation: 308763

Can't Run JCUDA 0.4.1 Over CUDA 4.0 On Windows 7 64-bit, Sun JDK 7, and IntelliJ 11.1

The question lays out all the pertinent versions.

I'm trying to run one of the sample files at http://www.jcuda.org: JCublasMatrixInvert.java.

I've got the NVIDIA CUDA drivers installed; I've downloaded the JCUDA 0.4.1 bindings and added their /lib directory to my java.library.path. Here are the JVM parameters I set when I run:

-ea -XX:+UseParNewGC -XX:+UseNUMA -XX:PermSize=128m -XX:MaxPermSize=256m -Xmx1024m -Djava.library.path=F:\Projects\Java\learning\out\production\learning

Yet when I try to run the class I get the following stack trace:

Error while loading native library "JCublas-windows-x86_64" with base name "JCublas"
Operating system name: Windows 7
Architecture         : amd64
Architecture bit size: 64
Stack trace from the attempt to load the library as a resource:
java.lang.NullPointerException: No resource found with name '/lib/JCublas-windows-x86_64.dll'
    at jcuda.LibUtils.loadLibraryResource(LibUtils.java:151)
    at jcuda.LibUtils.loadLibrary(LibUtils.java:83)
    at jcuda.jcublas.JCublas.initialize(JCublas.java:82)
    at jcuda.jcublas.JCublas.<clinit>(JCublas.java:70)
    at matrix.jcuda.JCublasMatrixInvert.main(JCublasMatrixInvert.java:34)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:601)
    at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120)
Stack trace from the attempt to load the library as a file:
java.lang.UnsatisfiedLinkError: F:\Projects\Java\learning\out\production\learning\JCublas-windows-x86_64.dll: Can't find dependent libraries
    at java.lang.ClassLoader$NativeLibrary.load(Native Method)
    at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1928)
    at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1854)
    at java.lang.Runtime.loadLibrary0(Runtime.java:845)
    at java.lang.System.loadLibrary(System.java:1084)
    at jcuda.LibUtils.loadLibrary(LibUtils.java:94)
    at jcuda.jcublas.JCublas.initialize(JCublas.java:82)
    at jcuda.jcublas.JCublas.<clinit>(JCublas.java:70)
    at matrix.jcuda.JCublasMatrixInvert.main(JCublasMatrixInvert.java:34)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:601)
    at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120)

Exception in thread "main" java.lang.UnsatisfiedLinkError: Could not load the native library
    at jcuda.LibUtils.loadLibrary(LibUtils.java:129)
    at jcuda.jcublas.JCublas.initialize(JCublas.java:82)
    at jcuda.jcublas.JCublas.<clinit>(JCublas.java:70)
    at matrix.jcuda.JCublasMatrixInvert.main(JCublasMatrixInvert.java:34)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:601)
    at com.intellij.rt.execution.application.AppMain.main(AppMain.java:120)

Process finished with exit code 1

I can see the allegedly missing dll at the path 

F:\Projects\Java\learning\out\production\learning\JCublas-windows-x86_64.dll

What have I missed? What am I doing wrong? Can anyone who has had success with running this example advise me? Thanks.

Upvotes: 0

Views: 1880

Answers (2)

Mysterion
Mysterion

Reputation: 9320

If standard way is too hard - I recommend you to use mavenized version of JCuda. You could clone standard project here - https://github.com/MysterionRise/mavenized-jcuda it's working automatically for Windows/Linux 32/64 bit

All you need to do is to run:

  • You need to install Cuda 5.5 for your platform here - https://developer.nvidia.com/cuda-downloads
  • Call mvn clean package to build project
  • Call mvn exec:exec to run main class with "Hello, JCuda" sample :)
  • After that you could replace main class with your JCuda code

Upvotes: 1

djmj
djmj

Reputation: 5544

Download that file from the jCuda download section and put it into your library path.

Upvotes: 0

Related Questions