HPC Phase IV Platform Training | vLLM-Ascend Inference Deployment and Tuning in Practice

To better support the research and teaching of our university's faculty and students, the Information Technology and Data Intelligence Department will hold "HPC Phase IV Platform Training | vLLM-Ascend Inference Deployment and Tuning in Practice" on Tuesday, November 10, 2026, from 9:30 to 11:30.

 

This training focuses on vLLM-Ascend inference deployment and tuning, and will carry out AI development practice on the Phase IV domestic cluster of the HPC platform. It is aimed at beginners and entry-level users to help everyone quickly become familiar with the platform's core functions. The training features a dedicated hands-on session, and on-site participants will be provided with dedicated accounts, computing resources, and accompanying training manuals. Participants are requested to bring their own laptops for on-site operation.

 

The specific arrangements are as follows:

Theme: vLLM-Ascend Inference Deployment and Tuning in Practice

Time: Tuesday, November 10, 2026, 9:30 - 11:30

Venue: E1, 1st Floor, 134

Online participation: Tencent Meeting: 838-465-263

Participants: All faculty and students of the university

Language: Mandarin

 

Training content outline:

1. Introduction to Ascend Super Node and Architecture (10 min)

2. vLLM-Ascend inference deployment process + hands-on practice (45 min)

3. Inference accuracy analysis methods (10 min)

4. Inference performance analysis methods (15 min)

5. Inference performance tuning methods + hands-on practice (30 min)

 

Faculty and students are warmly welcome to participate! Training registration link: https://qnr.hkust-gz.edu.cn/vm/rXLgqEj.aspx

 

If you have any questions, please contact hpc@hkust-gz.edu.cn.

 

Information Technology and Data Intelligence Department(ITDID)

November 6, 2026