NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Software > SddMetrics > MetricsC32008

Changes | Index | Contents | Search | Statistics | Go

Software Development Metrics & Information Summary for C3 2008

31 March 2008 through 16 May 2008

The Plan of Record used for this cycle can be found at PlanOfRecordC32008.

• Overview
  • Project Allocation
  • Cycle Commitments
• Operational and User Support
• Commitment Highlights by Project
  • Dynamic Scheduling
  • PTCS
    • Holography
    • Active Surface
    • Servo
  • Mustang
  • GUPPI
  • KFPA

Overview

11 SDD members worked ~3360 hours in C3 2008 over 35 days (4% division administration and planning, 77% cycle commitments, 3% GBT operations and user support, 7% overhead, and 10% excused absences).

Project Allocation

The cycle commitments were allocated over 6 major projects: Dynamic Scheduling, PTCS, Mustang, Cicada/GUPPI, K-band Focal Plane Array, and Software Continuing Maintenance and Enhancements. We also worked on 1 other project this cycle to which resources were not allocated at the beginning of the cycle - RRI. In the chart below, "other" is used as a catch-all for those cycle commitments or activities which are not associated with a major project.

The projects are further broken down into the individual commitments in the next section.

Cycle Commitments

The cycle commitments for C3 consisted of 31 commitments. A summary of each major commitment is discussed later in its project section below. Of the 31 commitments, 31 were met satisfactorily (of these 17 were part of larger efforts which will continue into future cycles). A graph showing actual vs. expected person-hours of effort for each project's commitments is discussed below in the commitments section below.

Operational and User Support

The items listed below pertain to time spent supporting GBT operation. These tasks usually involve helping users, fixing/troubleshooting system problems, and fixing/troubleshooting lab test environment problems. Note: Effort estimates reflect total time elapsed, i.e. how long did it take to fix the problem start to finish - not the sum total of each person's time.

Items:

• date - who - what - how long? (low, medium, high)
• 4 Apr 2008 - MarkClark - Reminded observer that the Spectrometer can select 3 vs 9 level only with slow samplers - low
• 7 Apr 2008 - AmyShelton - Minor BOS support. Added wording to remind visitors to submit and finalize reservations. - low
• 18 Apr 2008 - JoeBrandt, MarkClark, RamonCreager - Astrid hanging because of corrupted ports on earth. - low
• 18 Apr 2008 - PaulMarganian - BOS Support for Journal Entry bug. - low
• 18 Apr 2008 - PaulMarganian - GBTIDL user support for issues related to large file sizes. - low
• 29 Apr 2008 - PaulMarganian - SDFITS online filler not keeping up w/ data rate. - low
• 29 Apr 2008 - RamonCreager, MarkClark - AntennaCharacterization manager messages persisting even after normal operation resumes. low
• 28 Apr 2008 - JoeBrandt Az oscillation/servo shutdown issue medium
• 28 Apr 2008 - PaulMarganian - GBTIDL support - liscence issue - low
• 29,30 Apr 2008 - JoeBrandt X1/X2 actuator problem debugging medium
• 30 Apr 2008 - JoeBrandt, PaulMarganian - SDFITS/GBTIDL support: could not connect to online file due to testing - low
• 1 May 2008 - PaulMarganian - getting servo monitor samplers running again. low
• 11 May 2008 - JoeBrandt - Power failure recovery low
• 12 May 2008 - MarkClark - Spectrometer failure, referred to electronics, hardware failure. low

Key:

• low - Things that take less than two hours.
• medium - Things that take more than two hours, but less than eight hours.
• high - Things that take more than eight hours.

Commitment Highlights by Project

Dynamic Scheduling

Information on the Dynamic Scheduling project can be found at the Dynamic Scheduling Project Page and Dynamic Scheduling Web Home. The effect of weather conditions, both wind and opacity, on high frequency astronomy is well known. Currently, the GBT is using a fairly basic dynamic scheduling system which neither allow for rapid changes in the weather nor for long spells of good or bad weather. To maximize our use of good weather, we need to implement a true dynamic scheduling system on the telescope. The design and implementation of such a system is the goal of this project. This cycle, the Dynamic Scheduling Project team spent most of their time readying the system for testing during the 08B trimester. This cycle we have worked toward adding more information fields to the database, creating statistical analysis reports for scheduling simulations, creating a scheduling overview calendar covering months at a time, integrating observer information and accounts into the DSS via the BOS, and scheduling telescope periods on 15-minute intervals. We have been working closely with the NRAO e2e group, who are contributing 0.25 FTE software development effort.

The project was allocated ~752 person-hours of effort, and expended ~894. We spent more time on Dynamic Scheduling than anticipated due to ongoing issues with scheduling simulations as well as preparations for the upcoming test trimester.

PTCS

The project was allocated ~643 person-hours of effort, and expended ~704. This project received extra effort because other items to which software team members were assigned did not require the anticipated level of effort, e.g. less Mustang support and lower overhead than expected. A summary of the commitments for PTCS is shown below:

Holography

The traditional holography backend was originally deployed years ago on a VxWorks system with an MCB interface. The system was limited to 200 ms dumps due to the combination of the VxWorks OS and the MCB interface. We need the backend to dump much faster in order to properly map the GBT's panels. A new host platform and operating system is required, and a new interface to the data stream and the command stream will be created. The external interface to the M&C system will remain unchanged, except for the volume of data recorded, and the legal command values for integration time. This cycle responsibility for the development of the traditional Holography Manager was transferred to a member of PTCS from the Electronics Division due to a resignation within the Software Development Division. Work on this is expected to wrap-up shortly.

Additionally, some work was done on the Out-Of-Focus Holography Software Revival. This effort will continue into next cycle.

Active Surface

The purpose of ongoing Active Surface rework is to reduce Active Surface RMS deviation error to meet goals specified in "Refined Scientific Requirements for the PTCS, section 4.1" (Condon, 2003). And additionally, improve reliability of Active Surface control computers. This is an ongoing activity expected to span many cycles. Work on this item this cycle moved from existing code inspections and analyzing surface error data from current system in order to determine what improvements will need to be made to meet PTCS specification to incremental improvements to the software system. We allocated ~67 person-hours for this effort and expended ~67 person-hours this cycle. This effort will continue into next cycle.

The Active Surface manager currently has the ability to implement a set of Zernike coefficients. These coefficients yield surface corrections that are a function of the observed elevation, and which do a good job in ameliorating the large-scale gravitational distortions of the dish. In order to attempt to counteract thermal deformations of the surface (e.g. during daytime observations), we need a way to conveniently implement another set of Zernike coefficients in addition to the gravitational terms. Work was done this cycle to add this capability to the Active Surface manager. We allocated ~26 person-hours for this effort and expended ~34 person-hours this cycle. This effort will continue into next cycle.

Servo

We supported ongoing servo PLC upgrades to the CCU and SCU for the GBT servo system. The CCU and SCU software communicates with antiquated hardware wire-wrapped Control, Status and Rate-Loop boards. All control commands and status data are communicated via hardware analog and digital I/O lines between these boards and the SCU/CCU. The Servo PLC upgrade requires that all digital status data be communicated to the SCU and CCU over a Modbus TCP Ethernet link. To simplify the design and to poise us for future design improvements we would like the digital SCU and CCU commands, as well as all digital status data, to be also transmitted to the PLC via the Modbus interface. The analog interface between the SCU/CCU and the Interface Cabinet remains totally unchanged. We allocated ~80 person-hours for this effort and expended ~88 person-hours this cycle. This effort will continue into next cycle.

The next generation servo development work pertains to the replacement of the servo control systems on the Green Bank Telescope (GBT). This effort is being undertaken as part of a larger set of enhancements to the GBT under the PTCS program. The goal of the servo replacement project is to give the GBT a level of performance markedly better than current performance. This will allow both an increase in the environmental operational limits of the telescope yielding increased observation time and an improvement and extension of the telescope’s high frequency capabilities. This is an ongoing activity expected to span many cycles. Work this cycle continued to focus on creating a ServoReplacementSpecification, existing code inspections, and preliminary software analysis design activities. We allocated ~126 person-hours for this effort and expended ~139 person-hours this cycle. This effort will continue into many future cycles.

Also, this cycle we made changes to the Servo Monitor, which is a data acquisition system which samples and logs GBT Servo Room data. Currently, the sampling rate is 50Hz for all I/O points which is, in some cases below Nyquist rate. The changes involved increasing the sampling bandwidth to 200 Hz to avoid aliasing effects. We allocated ~84 person-hours for this effort and expended ~84 person-hours this cycle. This item is complete.

Along with the changes to the Servo Monitor, we modified the sampler2log program. The sampler2log program is a general use logging program which attaches to specific data streams and produces a FITS formatted log file. Some data streams may be sampled at rates significantly higher than the desired logging rate. For example, many data streams are sampled at 50Hz rates, where only a 10Hz or 1Hz logging rate is desired. Currently there is an option to 'pick' (i.e. just log one of every Nth sample), but it does this without any signal processing to band-limit the resulting data. The work this cycle on sampler2log added the correct signal processing to implement decimation by a integral number. We allocated ~66 person-hours for this effort and expended ~70 person-hours this cycle. This item is complete.

Mustang

The Ygor Manager for Mustang is now in routine use for Mustang observations. Work continued this cycle and will continue into the coming cycles to put the finishing touches on the software as well as extend its capabilities.

There were at least 15 separate activities associated with the Mustang effort. These include the Mustang Meta-MR, Cryogenics Control, Cryogenics Monitoring, Cryogenics FITS, Crate Control, Tower Control, FITS Writer, High Level Parameters, Manager Startup, Spider Tests, Data Monitoring, Calibration Diode Control, Collect Detector Bias Curves, Messaging Modifications, and also general support.

Of these 15 activites, 7 were completed this cycle. These were: Mustang Meta-MR, Spider Tests, Mustang Crate Control, Mustang Tower Control, Mustang FITS Writer, High Level Parameters, and Messaging Modifications. The rest will continue into next cycle.

We allocated ~471 person-hours of effort this cycle and expended ~337. The bulk of the discrepancy was due to unexpected excused absence.

GUPPI

The Green Bank Ultimate Pulsar Processing Instrument (GUPPI) is the next-generation pulsar processor for the GBT and is built on the CASPER platform. GUPPI development aims for first-light in the early months of 2008 and requires at least an expert-user interface to provide simple controls and access to its parameters. The short-term goal is to provide a lightweight interface to be used during the engineering and commissioning phase, with support for later integration with the Configuration Tool and Astrid. This lightweight interface is to be used to prototype what the user would see in the Astrid backend tab for GUPPI. The Very-Expert-User Interface and Control Framework for GUPPI (i.e., GUPPI Controller) aims to support GUPPI's first light requirements and provide a simplified user interface with full-access to all GUPPI parameters in a framework which is portable -- for potential use at other institutions -- and is modular -- to support various client-server communication models upstream. This development is being done in coordination with the NRAO e2e group, who are graciously lending 0.25 FTE.

There were 5 separate activities associated with the GUPPI effort: Conceptual Design and Very-Expert-User Interface, IBOB and BEE2 Interfaces, Guppi M&C Manager, GPU R&D, and general support.

We allocated ~253 person-hours of effort for this task and expended ~235.

KFPA

The goal of the generic GBT data analysis planning activity is to evaluate the current effectiveness of the GBT data reduction package, GBTIDL, identify its current deficiencies, and to plan for the future of GBT data reduction. We expect this effort to continue into the coming cycles and evolve into an effort to implement the recommendations which sprout from the analysis. However for the next fiscal year, our efforts will focus on the data reduction pipeline software for the K-band Focal Plane Array specifically. We plan to use this effort as a case in point for future pipeline and data reduction planning because much of the KFPA pipeline is generalizable to other GBT receivers and backends as well. This cycle we have continued working to create a GBT SDFITS to ACSIS (JCMT format) conversion in order to leverage third party data visualization software in the Starlink library (especially GAIA). We are also investigating the reuse of our sdfits code for the pipeline. We allocated ~237 person-hours of effort for this task and expended ~107. Of the 130-hour discrepancy, 42 hours arose from support of Dynamic Scheduling and 60 hours from preventive maintenance items associated with GBT data reduction.

We also started development of low-level M&C functionality for the single-pixel test system. We allocated ~66 person-hours of effort for this task and expended ~57. Work will continue on this item for the coming cycles.