PP: Can we use the OpenHPI (or a custom) daemon, one per CPU?
FDSA: We should give it a try. Prepare the software and we will evaluate it.
PP: Is there an infrastructure allocated to run the IPMI IOCs?
FDSA: The MicroTCA machines will run disk-less. The image servers could be used for crate monitoring.
PP: What distribution of GNU/Linux is used on those machines? We currently use CentOS for development.
FDSA: The MicroTCA CPUs will run Yocto. For the server side development continue using CentOS.
PP: Who to contact about details on the GNU/Linux distributions used in ESS?
FDSA: Olivier or Han
IDK: However, currently they are on leave/vacation/something
New piezo driver RTM (HPD200):
- 1 circuit less (Programmable Gain Amplifier - PGA) - better results (gain was not the same between the boards)
- channel switching and protection: UART to be replaced by 12 control lines (Diagnostic MCU <=> FPGA) + IPMI
BSP:
- MicroBlaze is provided by NCBJ but not used.
- DDR3 controler provided by NCBJ is for different memory than soldered (NCBJ promised the new BSP till the end of June), at the moment BRAM is used.
Anders Johanson is main contact regarding Piezo Driver. TUL is a part of ICS now (in that topic).
WJ Talk:
e3-icblmapp was created and is too big due to tosca as submodule.
It was requested for Han a merge request.
Asyn instead NDS3.
A lot of threads runnings in parallel.
Standard way to write data on PVs.
512 samples per time.
It's needed a final algorithm for icBLM.
How to use the calibration file? Read, apply and Publish as PV's?
How the control panels for icBLM should look like?
ICBLM Specification can be found at https://jira.esss.lu.se/browse/PBIICBLM-15
Irena presentation of how the IOC should look like:
Table 6 - ID's were discussed and how it should appear on IOC. Historic acquisition should be shown.
Raw data and Calibrated data. It should be selected.
Range selected on real time?
Same decimation on everything.Multi axes or check box for previous data.