Is NVIDIA K40C compatible with cuDNN v5.1 ?

Hi all,

I am using cuDNN 4007 at the moment with 2 NVIDIA K40C.

I was wondering if they are fully compatible with the brand new cuDNN 5.1?

For info, here’s my nvidia-smi output.

Thanks in advance!

Best,
Philippe

±-----------------------------------------------------+
| NVIDIA-SMI 352.63 Driver Version: 352.63 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla K40c Off | 0000:6E:00.0 Off | 0 |
| 27% 55C P0 64W / 235W | 5963MiB / 11519MiB | 0% Default |
±------------------------------±---------------------±---------------------+
| 1 Tesla K40c Off | 0000:6F:00.0 Off | 0 |
| 24% 45C P8 20W / 235W | 23MiB / 11519MiB | 0% Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1599 G /usr/bin/X 15MiB |
| 0 2638 C /usr/bin/python 73MiB |
| 0 3085 C python 5847MiB |
±----------------------------------------------------------------------------+

Yes, any version of cuDNN should be compatible with any cc3.0 or higher (i.e. “Kepler”) GPU. Your K40c is cc3.5, a member of the Kepler family.

https://developer.nvidia.com/cudnn

“cuDNN is supported on Windows, Linux and MacOS systems with Pascal, Kepler, Maxwell, Tegra K1 or Tegra X1 GPUs.”