I am interested in speeding up the ‘number crunching’ ability of my nearly 10 year old laptop.
The language I am learning is Python, and wish to use the GPU version of Tensorflow.
My laptop has a CPU 2630QM and an NVIDIA GeForce GT 540M.
My questions are similar to the suggested questions, but NOT for my NVIDIA GeForce GT 540M.
There are two SDKs that I need to install. My questions are:
- Is there a CUDA sdk specific/appropriate for my GT 540M
- Which is the appropriate cuDNN that I need for a GT540M. There are so many cuDNN versions I don’t know what is the appropriate version to install.
- Will any version of the cuDNN especially for older machines work with the CUDA sdk?
- What are the environmental paths needed to set up the CUDA sdk and the cuDNN.
Anthony of Sydney