CUDA 11.2 (CentOS8)Launch at CHI@TACC Launch at CHI@UC
The CentOS 8 CUDA 11.2 appliance is built from the CentOS 8 appliance and customized with CUDA 11.2. CUDA is a parallel computing platform and programming model developed by NVIDIA. It enables to use "CUDA-enabled" GPUs for general purpose processing (GPGPU).
This appliance is based on the default CentOS 8 which has been customized to provide the NVIDIA drivers and the CUDA toolkit. CUDA version 8+ is required for the NVIDIA P100 GPUs.
CUDA samples are installed in: /usr/local/cuda/samples
- Image name: CC-CentOS8-CUDA11
- Default user account: cc
- Remote access: Key-Based SSH
- Root access: passwordless sudo from the cc account
- Chameleon admin access: enabled on the ccadmin account
- Cloud-init enabled on boot: yes
- Repositories (Yum): EPEL, RDO (OpenStack Liberty)
- Installed packages:
- CUDA 11.2.460
- Standard development tools such as make, gcc, gfortran, etc.
- Config management tools: Ansible
- OpenStack command-line clients
- OpenStack Heat configuration agents
- Chameleon packages: rubygem-g5k-checks (providing cc-checks), cc-snapshot, cc-cloudfuse
KeywordsCentOS cuda gpu
|Created By:||zzhen on Jan. 26, 2021, 9:02 a.m.|
|Updated By:||zzhen on April 2, 2021, 8:03 a.m.|