Recurring Meeting of Cleveland Clinic - Stanford University teams
Date: October 7th, 2013
Time: 5.00 PM EST
Means: Conference Call
Attendees:
Cleveland Clinic
- Ahmet Erdemir
- Snehal Chokhandre
- Craig Bennetts
Stanford University
- Joy Ku
- Henry Kwong
Agenda
- Project description
- Expectations from Stanford University team
- Scheduling of monthly conference calls
- Action items for following month
- Other
Notes
- Project description
- Ahmet gave a general overview of the grant.
- Stanford team will be involved in achieving the specific aim 1 which is 'To provide an open, freely available, and collaborative development, testing, simulation and dissemination platform for in silico exploration of the biomechanics of healthy and diseased knees.'
2. Expectations from Stanford University team
- In the existing Simtk infrastructure have a link to provide a gateway to high performance computing platforms.
- High performance computation allocation from XSEDE will be obtained.
- Users can submit jobs through the web based interface to the high performance computing systems (Similar to Robetta).
- Integration between such a gateway and Simtk will be accomplished in this grant.
- Users can click on the 'Compute' link which will open a gateway, users can specify settings/ or upload a workflow file and run the job on the cloud.
- Authentication and keeping track of users will be required.
- Gateway will be inherent to Simtk infrastructure/ Simtk user access and authentication will be used.
- Potential resources : google compute engine/ server at Stanford site/ XSEDE (allocation is easy to gain and gateway development is supported)
- For commercial purposes a different high performance computing gateway can be made available.
- Tools needed to run : FEBio
- Results will be kept in an abridged version on the cloud.
- All the results will be kept public.
- Computing power requirements will be based on how much the gateway is used.
- Gateway can be initially prototyped on a simple machine and then deployed to high performance facility.
- Initial goal is to prototype the framework.
- Prototype can be developed with XSEDE.
- Open knee generation 1 model can be used in prototyping.
- The goal is to be able to pull models from the source code repository or the release version and use the same authetication as Simtk and run models.
- The user should be able to hit compute- open a server page - choose which model to use- pick workflow ( loading,boundary conditions can be changed )-submit job to computing server. When done email is send to user.
- All the scripts will be in python.
- FEBio will be on the compute server.
3. Scheduling of monthly conference calls
- The Cleveland Clinic and Stanford University teams will meet on every third Wednesday of the month.
- Next meeting: November 20th, 2013.
4. Action items for following month
- Ahmet and Craig will write specification for exact intial workflow.
- Stanford team in the meantime can get familiar with XSEDE framework.