We are starting a beta test for a new research project. Here are the basic details:

• 2,000 workunits available initially
  
• The project will use redundant copies so there will be at least 4,000 total results available in this initial phase
  
• Much longer than normal runtime - I would expect the runtime for this initial beta to average 20+ CPU hours
  
• Much larger than usual input and output data - The input data sizes for this initial test are about 29 MB compressed and the output data sizes are about 128 MB compressed.
  

To register for beta testing please log into the website and navigate to My Contribution -> Beta Testing and verify that the "Participate in Beta Testing" checkboxes are checked for the profiles you want participating.

IMPORTANT NOTE - PLEASE READ
I want to set everyone’s expectations appropriately for this beta test. Due to some issues and complexities with the software for this project, our internal development and testing time has been much longer than normal. Due to that, I anticipate a longer beta testing time than usual as well. We will learn a lot from this initial beta test.
Once I have enough data to determine which way it is headed, I will update this thread.

THE FINE PRINT
Since everyone has been looking forward to a new project for awhile, I will include some more details for those that are interested.
Many may guess that this is one of the climate projects. I cannot confirm nor deny that, but I will say the research application being used is the Weather Research & Forecasting (WRF) model.
WRF is a very large, mostly Fortran application, and this has been one of the factors contributing to the longer than usual development time. For this project, the only method available to validate results is to run redundant copies and check for binary equivalence. While the WRF application does include restart capabilities, we have run into bugs that are causing slight variations in output after a restart, which means that the results are not binary equivalent. This is one of the issues that has extended the development time.
Additionally, WRF is typically run on large compute clusters and restarts are not typically used as granular as we need them. Some of the bugs we have found we have been able to track down the cause and fix in the code. However, there is still a bug or two that have proven to be very difficult to squash. We are currently testing a work-around internally that is behaving nicely in our environment. However, the real test will be with our beta environment to prove whether or not the work-around is sufficient.
For those not familiar with the WRF application, it is used to simulate weather conditions over a region over a defined time period. The work for this project will be broken into small geographical regions, and in the end each region will be simulated for one calendar year. Each individual work unit represents 48 hours calendar time for this simulation. Once a result has been validated for the 48 hours, the output will be used to build the input for the next 48 hours of runtime. This is similar to some of our other projects, but the good news here is that the work units for the next piece of the simulation will be generated solely by us. There will be no delay in sending the work units back to the researchers for generation.
For this first beta test, all 2,000 workunits will be only the initial 48 hour simulation period.

As always, thanks so much to all of the beta testers.
armstrdj