Dates
14-16 July 2026
Three full days, 09:00-18:00

Chiang Mai, Thailand · 14-16 July 2026
A three-day hands-on boot camp for students and technical practitioners who want to program high-speed HPC networks, benchmark RDMA performance, and build a mini-project on real systems.
Dates
Three full days, 09:00-18:00
Venue
Andromeda meeting room, Planetarium and Exhibition Building, Mae Rim
Cohort
For CS, computer engineering, AI, data science, HPC, and networking backgrounds
Format
Lectures, real-system labs, open Q&A, instructor research sharing, and student presentations
Meals
Lunch, refreshments, and drinks will be provided at the venue throughout the camp
Why this camp
The proposal frames the camp as a national capacity-building effort: bring RDMA programming out of isolated competitions and into a repeatable learning pathway for Thai students, researchers, and advanced computing teams.
Thailand has growing HPC infrastructure, including LANTA, Chalawan, and ADA, but practical RDMA programming experience on real systems is still scarce.
The HPC-AI Advisory Council has run Asia-Pacific RDMA programming activities for more than nine years. This camp helps make that knowledge transfer continuous inside Thailand.
Participants learn RDMA concepts, verbs, communication patterns, benchmarking, and mini-project design so they can apply high-speed networking in research and industry.
Participant outcomes
The camp combines lectures with guided labs, real-system benchmarking, and mentored mini-projects, so participants leave with practical RDMA experience instead of slide-only exposure.
Understand RDMA and HPC networking principles from hardware through software workflows.
Connect MPI, NCCL, UCX, UCC, InfiniBand, RoCE, and verbs concepts into one working communication stack.
Run collective benchmarks and RDMA demos on available training systems.
Reason about performance tradeoffs in memory registration, OS bypass, GPUDirect RDMA, and GPU collectives.
Join a national network of students, researchers, and HPC practitioners.
Three-day agenda
Sessions run 14-16 July 2026 with lectures, hands-on work, Q&A, and closing presentations.
Day 1 · 14 July 2026 (Tue)
09:00-09:15
Opening + Welcome
On-site · 15 min
09:15-10:05
MPI Foundation + Operations
On-site · 50 min · 10 min break
10:15-11:05
MPI Process Grouping & Communicators
On-site · 50 min · 10 min break
11:15-12:05
Communication Profiling (MPI + NCCL)
Remote · 50 min
12:05-13:30
Lunch
Break · 85 min
13:30-14:20
InfiniBand Fundamentals + Networking Trends + RoCE
On-site · 50 min · 10 min break
14:30-15:20
Verbs Channel Provider Model
On-site · 50 min · 10 min break
15:30-16:20
OS Bypass + User-level Networking
On-site · 50 min · 10 min break
16:30-17:30
Hands-On 1: NCCL + HPC-X collective benchmarks (on ThaiSC LANTA)
On-site · 60 min
17:30-18:00
Day 1 Q&A + Wrap
On-site + remote · 30 min
Day 2 · 15 July 2026 (Wed)
09:00-09:50
UCX Architecture
Remote · 50 min · 10 min break
10:00-10:50
UCC Architecture + APIs
Remote · 50 min · 10 min break
11:00-11:50
UCC TL/CL Implementation + TL Extensibility
Remote · 50 min · 10 min break
12:00-13:30
Lunch
Break · 90 min
13:30-14:20
RDMA Design Patterns + Memory Registration
On-site · 50 min · 10 min break
14:30-15:20
Verbs APIs for High-Performance RDMA Programming
On-site · 50 min · 10 min break
15:30-16:20
Hands-On 2: Verbs P2P + UCX client-server demo
On-site · 50 min · 10 min break
16:30-17:30
Open Q&A + Instructors' Research Sharing
On-site + remote · 60 min
17:30-18:00
Day 2 Wrap
On-site · 30 min
Day 3 · 16 July 2026 (Thu)
09:00-09:50
GPUDirect RDMA + Peer-direct
Remote · 50 min · 10 min break
10:00-10:50
NCCL GPU Collective Communication
Remote · 50 min · 10 min break
11:00-12:00
Course Summary + Take-home Assignment Briefing
On-site + remote · 60 min
12:00-13:30
Lunch
Break · 90 min
13:30-16:00
Student Presentation + Reflection (each team: ~15 min, including Q&A)
On-site + remote · 150 min
16:00-17:00
Closing + Group photo
On-site · 60 min
Instructors and support
The training is led by HPC-AI Advisory Council instructors and supported locally through Thai HPC institutions and the NARIT venue.
HPC-AI Advisory Council
HPC-AI Advisory Council
HPC-AI Advisory Council
Project lead and assistant instructor
Venue
Planetarium and Exhibition Building, National Astronomical Research Institute of Thailand, 260 Moo 4, Don Kaew, Mae Rim, Chiang Mai 50180. Lunch, refreshments, and drinks are provided at the venue throughout the camp.
Open in Google MapsContact and project host
Hosted by the Faculty of Science and Technology, Thammasat University, with support from ThaiSC, NARIT, and the HPC-AI Advisory Council.
ผู้รับผิดชอบโครงการ
Worawan Diaz Carballo, Faculty of Science and Technology, Thammasat University
Contact
Participant support
Lunch, refreshments, drinks, learning documents, and training materials are provided during the camp. Participants are responsible for their own travel, accommodation, and personal expenses.