add cuDNN v5 support, drop cuDNN v3 support

[ningzhou]

The change was purely based on BVLC/caffe#4159

[Cysu]

Could you please consider switch back to the original lib64 path? As other linux distributions (ubuntu 14.04, arch linux) do not have lib/x86_64-linux-gpu under cuda.

[Cysu]

@ningzhou I have merged the PR. Thank you very much!

[ningzhou]

You are very welcome. In fact, thank you very much for open sourcing the
excellent work.

Ning

On Mon, Oct 24, 2016 at 11:19 PM, Tong Xiao [email protected]
wrote:

@ningzhou https://github.com/ningzhou I have merged the PR. Thank you
very much!

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#8 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABNQXuPRseEJkAppTdHJBNimWOv1qsbwks5q3Z9ogaJpZM4KfjGE
.

[ningzhou]

sure. I will create a new pull request.

On Mon, Oct 24, 2016 at 10:05 PM, Tong Xiao [email protected]
wrote:

@Cysu requested changes on this pull request.

Could you please consider switch back to the original lib64 path? As other
linux distributions (ubuntu 14.04, arch linux) do not have
lib/x86_64-linux-gpu under cuda.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#8 (review),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABNQXjTnD9DwRtBbcHLKxuF3JWGK1L34ks5q3Y4CgaJpZM4KfjGE
.