jcuda.org

Downloads



JCuda is published under the terms of the MIT/X11 License



Downloads


The following packages contain the binaries, documentation and source code of all libraries available on jcuda.org


The binary distribution archives contain the JCuda JAR files and the respective native libraries. The JAR files have to be present in the CLASSPATH, and the native library files must be located in a path that is visible either as a java.library.path, or as a default project path, or via an environment variable like the PATH environment variable on Windows.


Version 0.3.1. These libraries have been compiled for CUDA 3.1 (build 9)

Binaries for Windows 32bit The JAR files and the DLLs of all libraries for 32 bit Windows.
Binaries for Windows 64bit The JAR files and the DLLs of all libraries for 64 bit Windows.
Binaries for Linux 32bit The JAR files and the SOs of all libraries for 32 bit Linux.
Binaries for Linux 64bit The JAR files and the SOs of all libraries for 64 bit Linux.
These libraries have been provided by Kashif Rasul.
Binaries for MacOS 64bit The JAR files and the JNILIBs of all libraries for 64 bit MacOS.
These libraries have been provided by Christoph von Tycowicz and Ihor Smal.
Other binaries Binaries for other combinations of Windows, Linux, MacOS or Solaris on x86 32/64bit, PowerPC, PowerPC 64, SPARC, ARM, RISC or MIPS processors.
API documentations of all libraries This archive contains the Java API documentations of all libraries. Most of the documentation is directly taken from the documenations of the respective native libraries.
Source code of all libraries This archive contains the Java- and C- source code for all libraries, including the project- and makefiles for Windows, Linux and MacOS.

Note that in any case, in order to compile the native libraries, you will have to download the CUDA Toolkit and SDK from the NVIDIA CUDA developer web site.

Additionally, for compiling the native part of JCudpp, you will have to download and compile the corresponding CUDPP libraries. The source code may be obtained from the CUDPP home page.









Previous versions:

Version 0.3.0a. These libraries have been compiled for CUDA 3.0 (build 14)

Binaries for Windows 32bit The JAR files and the DLLs of all libraries for 32 bit Windows.
Binaries for Windows 64bit The JAR files and the DLLs of all libraries for 64 bit Windows.
Binaries for Linux 64bit The JAR files and the SOs of all libraries for 64 bit Linux.

These libraries have been provided by Adrian Challinor. Adrian also has created a bundle including the source code, documentation, binaries, precompiled libraries for CUDPP, and a utility class for handling CUBIN files: Developer bundle for 64 bit Linux.
Binaries for MacOS 32bit The JAR files and the JNILIBs of all libraries (except JCudpp) for 32 bit MacOS.

These libraries have been provided by Teruhisa Haruguchi from Cornell University
Other binaries Binaries for other combinations of Windows, Linux, MacOS or Solaris on x86 32/64bit, PowerPC, PowerPC 64, SPARC, ARM, RISC or MIPS processors.
API documentations of all libraries This archive contains the Java API documentations of all libraries. Most of the documentation is directly taken from the documenations of the respective native libraries.
Source code of all libraries This archive contains the Java- and C- source code for all libraries, including the project- and makefiles for Windows, Linux and MacOS.

Note that in any case, in order to compile the native libraries, you will have to download the CUDA Toolkit and SDK from the NVIDIA CUDA developer web site.

Additionally, for compiling the native part of JCudpp, you will have to download and compile the corresponding CUDPP libraries. The source code may be obtained from the CUDPP home page.


Version 0.3.0. These libraries have been compiled for CUDA 3.0 (build 14)
Binaries for Windows 32bit The JAR files and the DLLs of all libraries for 32 bit Windows.
Binaries for Windows 64bit The JAR files and the DLLs of all libraries for 64 bit Windows.
Binaries for Linux 32bit The JAR files and the SOs of all libraries for 32 bit Linux.
Other binaries Binaries for other combinations of Windows, Linux, MacOS or Solaris on x86 32/64bit, PowerPC, PowerPC 64, SPARC, ARM, RISC or MIPS processors.
API documentations of all libraries This archive contains the Java API documentations of all libraries. Most of the documentation is directly taken from the documenations of the respective native libraries.
Source code of all libraries This archive contains the Java- and C- source code for all libraries, including the project- and makefiles for Windows, Linux and MacOS.

Note that in any case, in order to compile the native libraries, you will have to download the CUDA Toolkit and SDK from the NVIDIA CUDA developer web site.

Additionally, for compiling the native part of JCudpp, you will have to download and compile the corresponding CUDPP libraries. The source code may be obtained from the CUDPP home page.


Version 0.2.3. These libraries have been compiled for CUDA 2.3


Binaries for Windows 32bit The archive contains the JAR files and the DLLs of all libraries for 32 bit Windows. The JAR files have to be present in your CLASSPATH, and the DLLs must be located in a path that is visible either as a java.library.path or from the path specified in the Windows PATH property.
Binaries for Windows 64bit The archive contains the JAR files and the DLLs of all libraries for 64 bit Windows. The JAR files have to be present in your CLASSPATH, and the DLLs must be located in a path that is visible either as a java.library.path or from the path specified in the Windows PATH property.
Binaries for Linux 32bit The archive contains the JAR files and the SOs of all libraries for 32 bit Linux. The JAR files have to be present in your CLASSPATH, and the SOs must be located in a path that is visible either as a java.library.path or as one of these Linux library paths...
Binaries for Linux 64bit The archive contains the JAR files and the SOs of all libraries (except JCudpp) for 64 bit Linux. The JAR files have to be present in your CLASSPATH, and the SOs must be located in a path that is visible either as a java.library.path or as one of these Linux library paths...

These libraries have been provided by Bibrak Qamar from NUST Pakistan.
Other binaries Binaries for other combinations of Windows, Linux, MacOS or Solaris on x86 32/64bit, PowerPC, PowerPC 64, SPARC, ARM, RISC or MIPS processors.
API documentations of all libraries This archive contains the Java API documentations of all libraries. Most of the documentation is directly taken from the documenations of the respective native libraries.
Source code of all libraries This archive contains the Java- and C- source code for all libraries, including the project- and makefiles for Windows, Linux and MacOS.

Note that in any case, in order to compile the native libraries, you will have to download the CUDA Toolkit and SDK from the NVIDIA CUDA developer web site.

Additionally, for compiling the native part of JCudpp, you will have to download and compile the corresponding CUDPP libraries. The source code may be obtained from the CUDPP home page.