ECE IT is launching a group of Linux GPU machines students can use for academic purposes.
Over the past several years, we’ve received quite a few requests from students for GPU-enabled machines. We are initially making six machines available. We will be using this initial launch to gauge demand.
The lab is comprised of the following machines:
- Dell Precision 3630 Tower workstations. These workstations are named sequentially, starting with linux-gpu-001.ece.uw.edu and ending with linux-gpu-006.ece.uw.edu .
- Specifications:
- Dell Precision 3630 Tower running AlmaLinux 9
- NVIDIA GeForce RTX 2060 (1920 CUDA cores)
- Intel Core i7-9700 3.0GHz (Cannon Lake) (8 cores)
- 32GB RAM (and 192GB swap)
- 512GB Samsung SSD
- Access to the GPU Lab machines is currently remote only – use SSH and/or RDP.
- Specifications:
- The CUDA Toolkit is installed, as are the cuDNN, cuBLAS, and TensorRT libraries.
- The lab workstations use the same networked Linux home directories as the ECE Linux Lab (VLSI/IS Lab). This means if you save a file in your home directory while working on one machine (in either lab), you’ll have that same file available to you from all other lab machines.
- These workstations are intended for coursework. Research tasks are not to be run on these computers.
More Info
- Installing GPU-Aware Software <– please read this
- Linux Lab Remote Access Using RDP