Skip to content

Commit

Permalink
# Version 1.2
Browse files Browse the repository at this point in the history
Code & binaries updated with libsvm version 3.17 & CUDA SDK version 5.5

Details:
Exchanged the header file “cublas.h” for “cublas_v2.h”.

Exchanged the type cublasStatus for cublasStatus_t.

Exchanged the function cublasAlloc() and cublasFree() for cudaMalloc() and cudaFree(), respectively.

Declared the cublasHandle_t CUBLAS library handle.

Initialized the handle using cublasCreate(). Also, released the handle once finished using cublasDestroy().

Added the handle as the first parameter to all the CUBLAS library function calls.

Changed the scalar parameters to be passed by reference, instead of by value (usually simply adding “&” symbol in C/C++ is enough, because the parameters are passed by reference on the host by default).

Changed the parameter characters 'N' or 'n' (non-transpose operation), to CUBLAS_OP_N.

Checked the function prototypes in the header files “cublas.h” and “cublas_v2.h” for code correctness.
  • Loading branch information
tpapaz committed Oct 4, 2013
1 parent ceb9ae3 commit 091f78a
Show file tree
Hide file tree
Showing 101 changed files with 14,806 additions and 4,960 deletions.
14 changes: 11 additions & 3 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,14 @@ GPU-accelerated LIBSVM is a modification of the [original LIBSVM](http://www.csi

Watch a [short video](http://www.youtube.com/watch?v=Fl99tQQd55U) on the capabilities of the GPU-accelerated LIBSVM package

###CHANGELOG

V1.2

Updated to LIBSVM version 3.17
Updated to CUDA SDK v5.5
Using CUBLAS_V2 which is compatible with the CUDA SDK v4.0 and up.

###FEATURES

Mode Supported
Expand All @@ -21,7 +29,7 @@ Functionality / User interface
LIBSVM prerequisites
NVIDIA Graphics card with CUDA support
Latest NVIDIA drivers for GPU

###PERFORMANCE COMPARISON

To showcase the performance gain using the GPU-accelerated LIBSVM we present an example run.
Expand Down Expand Up @@ -71,7 +79,7 @@ We are interested in porting our imlementation in Matlab but due to our workload

* Visual Studio will not let me build the provided project.

The project has been built in both 32/64 bit mode. If you are working on 32 bits, you might need the x64 compiler for Visual Studio 2008 to build the project.
The project has been built in both 32/64 bit mode. If you are working on 32 bits, you might need the x64 compiler for Visual Studio 2010 to build the project.

* Building the project, I get linker error messages.

Expand All @@ -80,7 +88,7 @@ Please go to the project properties and check the library settings. CUDA librari
* I have built the project but the executables will not run (The application has failed to start because its side-by-side configuration is in
correct.)

Please update the VS2008 redistributables to the PC you are running your executable and install all the latest patches for visual studio.
Please update the VS2010 redistributables to the PC you are running your executable and install all the latest patches for visual studio.

* My GPU-accelerated LIBSVM is running smoothly but i do not see any speed-up.

Expand Down
Loading

0 comments on commit 091f78a

Please sign in to comment.