MIPP impact statement¶
MIPP has allocated $39,000 for the DAQ upgrade. Of that $39,000, they have allocated $19,200 for the following software development-related tasks from CD:
- porting Linux SLC5 to their new PPC single board computers (about 8 weeks to complete) # consultation regarding the upgrade of their event builder package (small amount anticipated and not specified)
The $19,200 covers about one month of labor costs.
The MIPP plan predicts the porting task to take about eight weeks of effort. As currently planned, the funds allocated only cover half of this task.
The MIPP plan called for an upgrade to the event builder. The existing builder cannot be used to handle the new data format and the 100 times increase in data rate. They envisioned using a new CD-supported, generic event builder that would take minimal effort to integrate instead of the existing MIPP one.
There is no generic event builder to be used from CD to the best of my knowledge. I have informed them of this issue. They are planning to have the collaborate handle the construction of the new event builder and hinted that they would only need minimal input from the people who created the first event builder. I did not get names of the original authors, but I suspect they may not be at the laboratory any longer. It is unclear how much time this consulting effort will be. If it is only consulting work, I would allocate one week for FY11. This does not include participation in the development of this package.
The PPC boards specified in the plan are an older solution. Currently CD has x86 boards deployed and plenty of experience with them. The Linux SLC5 OS is already running on these types of boards.
An alternative plan for the PPC single board computers is to purchase and use x86 boards. This will bring the estimated effort for the porting task down to a couple of weeks and also allow for a more standard installation.
There is about nine weeks of effort needed in the current planning of MIPP. Currently Gennadiy Lukhanin and Ron Rechenmacher would be allocated for these tasks. Both are slotted to work on NOvA, mu2e, and MicroBooNE for FY11. The currently planned MIPP work is only support-related and does not further the software development mission of our department or group, whereas the tasks from the other experiments do.
The alternative plan only involves about three weeks of effort. Such involvement would be minimal. This minimal effort will permit our new group member (Gennadiy) to learn more about operating the x86 single board computer with Fermi Linux, which will have some benefits such as providing a backup for Ron.