Profile : Research Scholar with no background of GP
System Info :Win 8, nVidia GeForce 635 GT635M(2 GB)
Intentions: Develop the Ac-BM(Aho-Corasick Boyer Moore) algorithm in for the GPU preferably in openCL
Help : What toolkit is needed to develop/compile the openCL applications?
Suggestions: Is it better to use CUDA( as I didn’t find much of a mention of opencl on NVidia’s site?
toolkit: I am not sure if the nvidia CUDA developers toolkit supports openCL, but I know intel and AMD both have an openCL SDK for their CPU, so if you don’t find it here, maybe you can try those?
How I did it: When I started programming openCL, I created the openCL architecture, developed it in C99, backported it to a dll, and did the debugging from there using an 3d-iterator :p. Afterwards, in your program, you’ll have a call clBuildProgram(). If this call fails, you cal retrieve the error log with clGetProgramBuildInfo()
My suggestion: use CUDA. I don’t think you want to use your CPU aside from your GPU, and CUDA will provide you more flexibility when you want to develop nvidia-GPU-only applications