A hands-on week to learn DynEarthSol workflows, discuss geodynamic modeling challenges, and build new tools with the GeoFLAC/DynEarthSol community. Three days of workshops and a two-day hackathon in Daejeon — beginners welcome.
Whether you maintain the code, run simulations, or are simply curious about geodynamic modeling, this is your week. Learn the internals, sharpen your skills, and team up to build something new.
Guided sessions on GeoFLAC/DynEarthSol — from first build to advanced modeling techniques, led by core developers.
Two developer-centred days on the 2026 roadmap — features, fixes, tooling, docs. Bring a modeling question and we will pair you with a developer.
Meet the people behind the code. Connect with developers and researchers working on the same problems you are.
GeoFLAC/DynEarthSol developers and anyone interested are welcome — beginners included. No prior contribution required.
DynEarthSol is the open-source code we use throughout the tutorials — a finite element solver for the long-term deformation of Earth's lithosphere in 2D and 3D. Instead of assembling and storing the enormous matrices that traditional implicit methods require, DynEarthSol uses an explicit, matrix-free formulation based on the FLAC algorithm, so it sidesteps the "memory wall" and scales efficiently across CPUs and GPUs.
Solves the momentum equations explicitly with dynamic relaxation — no global matrix assembly, no memory wall as models grow.
A deterministic two-stage summation gives bitwise-reproducible results regardless of execution order, enabling robust GPU acceleration (OpenACC).
Couples tectonics, surface processes (via goSPL), and earthquake-cycle physics (rate-and-state friction) in one 2D/3D framework.
Results, simulations, and visualizations made with DynEarthSol. To contribute, email an image, video, or public link, or open a GitHub issue. Include a short caption, credit or affiliation, and confirm that you have permission to publish it. The organizers will review and add selected submissions.
Days 1–3 are hands-on workshops pairing tutorials with talks. Days 4–5 are a developer-centred hackathon and developer meeting — open collaborative work and roundtable discussion, with no presentation sessions. Sessions run 09:30–17:00 daily, with a one-hour lunch break at 12:00. The program below is tentative — times, titles and topics may still change.
Welcome and introductions, GeoFLAC and DynEarthSol overviews — including GPU acceleration in DynEarthSol v2.0 — then talks and a hands-on block on earthquake-cycle modeling.
Talks and tutorials on melt migration and two-phase flow. Workshop banquet in the evening.
Coupling DynEarthSol to surface-process modeling with goSPL, related talks, and hackathon preparation.
Code roadmap and pull-request guidelines, topic and issue selection, then open collaborative work. Workshop participants with something they want to try in the model are paired with a developer.
Continued work and PR preparation, collaborative review, and a developer discussion of future development and the next workshop.
Click a day to expand. Tutorial topics are still being decided — sessions marked topic TBA will be announced before the event.
| 09:30–09:40 | Welcome and orientationSungho Lee |
| 09:40–09:50 | Meet the GeoFLAC/DynEarthSol teamDevelopment team |
| 09:50–10:05 | Participant introductionsEveryone |
| 10:05–10:45 | Overview · GeoFLACLuc L. Lavier |
| 10:45–11:00 | Coffee break |
| 11:00–11:40 | Overview · DynEarthSol v2.0, incl. GPU accelerationEh Tan, Chase J. Shyu & Sungho Lee |
| 11:40–12:00 | Questions and discussionEveryone |
| 12:00–13:00 | Lunch — on-site cafeteria |
| 13:00–13:40 | Talk · Earthquake-cycle models: history and current stateSo Ozawa |
| 13:40–15:00 | Talk · Fault-zone thickness and the earthquake cycleSungho Lee |
| 15:00–15:15 | Coffee break |
| 15:15–16:40 | Tutorial / hands-on sessiontopic TBAEunseo Choi & DynEarthSol developers |
| 16:40–17:00 | Day 1 wrap-upEveryone |
| 09:30–10:45 | Tutorial 1topic TBAEunseo Choi & Chase J. Shyu |
| 10:45–11:00 | Coffee break |
| 11:00–11:40 | Talk · From Magma to Ice — a two-phase flow solver for Europa's chaos terrainChase J. Shyu |
| 11:40–12:00 | Questions and discussionEveryone |
| 12:00–13:00 | Lunch — on-site cafeteria |
| 13:00–13:40 | Talk · Continental rifting and lithospheric extensionLuc L. Lavier |
| 13:40–14:55 | Tutorial 2topic TBAEunseo Choi & Chase J. Shyu |
| 14:55–15:10 | Coffee break |
| 15:10–16:40 | Tutorial / hands-on sessiontopic TBAEunseo Choi |
| 16:40–17:00 | Day 2 wrap-upEveryone |
| 18:30–20:30 | Workshop banquetvenue to be confirmedEveryone |
| 09:30–10:40 | Tutorial 1topic TBAEunseo Choi |
| 10:40–10:55 | Coffee break |
| 10:55–11:35 | Talk · Coupling ASPECT and FastScapeJaeyoon Keum |
| 11:35–12:00 | Questions and discussionEveryone |
| 12:00–13:00 | Lunch — on-site cafeteria |
| 13:00–13:40 | Talk · Thermochronology and thrust systems — the Taiwan caseEh Tan |
| 13:40–14:20 | Tutorial 2topic TBAEunseo Choi |
| 14:20–15:20 | Tutorial 3topic TBAEunseo Choi |
| 15:20–15:35 | Coffee break |
| 15:35–16:40 | Tutorial / hands-on sessiontopic TBAEunseo Choi |
| 16:40–17:00 | Workshop wrap-up and hackathon preparationEveryone |
| 09:30–10:00 | Hackathon orientationSungho Lee |
| 10:00–11:15 | Code roadmap and pull-request guidelinesRoundtable — everyone |
| 11:15–11:30 | Coffee break |
| 11:30–12:00 | Team topics and issue selectionEveryone |
| 12:00–13:00 | Lunch — on-site cafeteria |
| 13:00–14:30 | Team formation and collaborative workTeams |
| 14:30–14:45 | Coffee break |
| 14:45–16:40 | Open coding, testing and documentationTeams |
| 16:40–17:00 | Team check-in and next stepsEveryone |
| 09:30–10:00 | Morning sync and team plansEveryone |
| 10:00–11:15 | Open collaborative workTeams |
| 11:15–11:30 | Coffee break |
| 11:30–12:00 | Open collaborative workTeams |
| 12:00–13:00 | Lunch — on-site cafeteria |
| 13:00–15:00 | Open work and PR preparationTeams |
| 15:00–15:15 | Coffee break |
| 15:15–16:00 | Collaborative PR review and documentation checkTeams |
| 16:00–16:40 | Future code development and workshop planningRoundtable — everyone |
| 16:40–17:00 | ClosingSungho Lee |
* Schedule subject to change. Detailed times will be shared with registered participants. Please bring cash for lunch.
The tutorials are run with DynEarthSol. We'll do a quick environment check on Day 1, but we will not walk through the full build process during the workshop — so please build it on your own laptop beforehand and confirm it runs. This way we can dive straight into modeling.
Clone the DynEarthSol repository (with submodules) and follow the build instructions in its README.
Make sure these are installed before building:
Come with DynEarthSol already built and running on your laptop. Run one of the examples to confirm it works.
⚠️ There is no public Wi-Fi on-site, so download any data, dependencies, and docs you'll need beforehand.
Stuck on the build? Email us before the event so we can help in advance: slee91@kigam.re.kr
DES user manual ↗ — setup, model configuration, and example tutorials.
Coupling DynEarthSol with GoSPL ↗ — preparation and a worked example for coupling tectonic deformation with surface processes.
Days 4–5 are developer-centred. Instead of a call for team proposals, we run an open working sprint around the project's 2026 development roadmap, together with a developer meeting on the contribution workflow and future direction. Workshop participants are very welcome to stay: if there is something you would like to try with the model, come along and we will pair you with the developer who works on that part of the code. No proposal, no team and no demo required.
Orientation, then a community discussion of the code roadmap and the GitHub contribution workflow (pull requests, review, push policy). Topics and issues are picked after lunch, and open collaborative work begins.
Continued work and PR preparation with developers on hand, a collaborative PR and documentation review, then a roundtable on future development and planning for the next workshop.
The sprint is run by and for the developers this year, so there is nothing to prepare and nothing to present. If the workshop sparked an idea — a setup you want to try, a feature you wish existed, a result you cannot reproduce — bring it along on Day 4 and we will connect you with the developer who knows that part of the code.
Presentations confirmed so far. Titles and abstracts are provisional unless the full abstract is shown, and slides are linked once available.
Thank you to everyone who submitted! Authors will be notified of acceptance by Oct 14, 2026, and accepted talks appear in the Talks section above.
Speakers: need to send or update your slides? Email them to us with your talk title — large files such as videos too.
Email your slides ↗Questions? slee91@kigam.re.kr
This is an in-person event in Daejeon. There is no online/remote option — join us on-site.
124 Gwahak-ro, Yuseong-gu, Daejeon 34132, Republic of Korea
Google Maps is limited in Korea — for directions, Naver Map or KakaoMap work best:
For Korean apps, copy & paste this address:
The venue is at 124 Gwahak-ro, Yuseong-gu, Daejeon 34132. Daejeon sits at the center of Korea and is easy to reach by train or bus. Here are the most common routes.
Airport limousine bus — direct to Daejeon (Government Complex / Bus Terminal). ~2.5–3 hrs, ₩26,000–33,000. Tickets at the airport bus counters.
Or by KTX — AREX train to Seoul Station, transfer to KTX to Daejeon Station (~50–65 min).
The closest airport (~1 hr away).
Direct bus to Daejeon Complex Terminal (~1 hr, limited departures).
Or via Osong Station — shuttle to Osong (~20 min), then one KTX stop to Daejeon (~15 min). Taxi ~50 min.
KTX / SRT from Seoul Station or Suseo to Daejeon Station — ~50–60 min. Fastest and most frequent.
Express bus from Seoul to Daejeon Complex Terminal — ~2 hrs.
From Daejeon Station or the bus terminal, the venue is ~20–30 min by taxi; from the Yuseong / Government Complex area it's ~10–20 min.
A few tips for international visitors — the right apps and cards make travel much easier.
Google Maps gives limited directions in Korea. Install Naver Map or KakaoMap for reliable walking, transit, and driving routes.
Use the Kakao T app to hail taxis — far easier than flagging one down, and you can pay in the app.
A T-money card works on buses and subways. Buy and top it up with cash at any convenience store (GS25, CU, 7-Eleven). iPhone users can add T-money to Apple Wallet and tap to pay.
A few hotels participants often use. These are suggestions only — please book early, as availability and prices vary.
Business hotel in the Expo/Yuseong area, one of the most convenient options for the venue.
Popular full-service hotel in the Yuseong area, near the Yuseong hot-spring district.
Affordable, reliable chain hotel near Government Complex Daejeon subway station.
Travel times and fares are approximate — please check official timetables before departure. For Korean map apps, search the address: 대전 유성구 과학로 124.
Last updated: June 2026.
Thank you to everyone who signed up!
Already registered?
Registration does not guarantee a seat. We will send confirmation emails by mid-October 2026. Spots are allocated on a first-come, first-served basis by registration time; if the event is full you will be placed on a waiting list. To keep participation balanced, we may limit the number of confirmed participants per institution.
Participation options (as selected in your registration):
The Developer meeting is a community session on the project roadmap and contribution workflow (push policy, future direction), held alongside the hackathon.
Questions? Email slee91@kigam.re.kr
We use your registration information only to manage participation and event communication.
GeoFLAC/DynEarthSol developers and anyone interested in geodynamic modeling. Beginners are welcome — you don't need prior experience with the codebase.
Nothing. The event is completely free to attend. (Travel and accommodation are not covered.)
No. This is an in-person event held on-site in Daejeon only.
No. Days 4–5 are developer-centred: an open sprint on the 2026 roadmap plus a developer meeting, with no proposals, teams or demos required. Workshop participants are very welcome to join — if there is something you want to try with the model, we will pair you with the developer who works on that part of the code.
No — registration is now closed. If you already registered, we will email you by mid-October 2026 to confirm your participation. Places are limited by venue capacity: seats are allocated first-come, first-served and as evenly as possible across institutions.
Yes. The tutorials use DynEarthSol, and the build process is not covered during the workshop. Please build it on your laptop in advance — see Before you come. If you run into build problems, email us before the event and we'll help.
A laptop with DynEarthSol already built and running, and a power adapter. Also bring cash for lunch (on-site cafeteria, ~₩7,000).