CUDA 8.0 - driver version is insufficient for CUDA runtime version

Hi All,

We currently have the driver version 384.90. However, CUDA 8.0 reports “driver version is insufficient for CUDA runtime version”. We have tried both CUDA-7.5 and CUDA-9.0, and both works.

Can anyone shine some lights on what the issue could be?

===================================
[kaisong@n0273 deviceQuery]$ ./deviceQuery
./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 35
-> CUDA driver version is insufficient for CUDA runtime version
Result = FAIL
[kaisong@n0273 deviceQuery]$ nvidia-smi
Tue Dec 5 14:38:50 2017
±----------------------------------------------------------------------------+
| NVIDIA-SMI 384.90 Driver Version: 384.90 |
|-------------------------------±---------------------±---------------------+
| 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 C2050 Off | 00000000:02:00.0 Off | 0 |
| 30% 49C P0 N/A / N/A | 0MiB / 2622MiB | 0% Default |
±------------------------------±---------------------±---------------------+
| 1 Tesla C2050 Off | 00000000:85:00.0 Off | 0 |
| 30% 43C P0 N/A / N/A | 0MiB / 2622MiB | 0% Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
±----------------------------------------------------------------------------+
[kaisong@n0273 deviceQuery]$ uname -a
Linux n0273.mako0 3.10.0-327.36.3.el7.x86_64 #1 SMP Mon Oct 24 09:16:18 CDT 2016 x86_64 x86_64 x86_64 GNU/Linux

===================================

Thanks in advance!

Kai