View source
spinor
  • Contributors
    1. Loading...
Author Commit Message Date Builds
250+ commits behind default.
Shaun Johnstone
Quick fix for phasematrix buffered mode, forcing output to be on every experiment Should be reverted once labscript implementation of StaticDigitalOutput for phasematrix is fixed (needs pseudoclock parent faked)
Russell Anderson
Made h5_lock debug messages of bad3fb4093e7 more consistent and exhaustive.
Russell Anderson
Start branch spinor for 7a898f1ffae3. This version has been running on bec-spinor-control for several weeks.
Chris Billington
Undid unnecessary changes and misleading comments
Chris Billington
Merge
Chris Billington
Gave BLACS processes meaningful names in zlock
Russell Anderson
Merge
Russell Anderson
Now require and use lyse port from LabConfig.
Chris Billington
Merge
Chris Billington
Added skeleton for sr400 photon counter tab. (added hgignore with globs for common untracked file extensions)
Chris Billington
ADWin now gets firmware from filepaths, so they are user-configurable. ADWin restores previous values when restarted. IPC for static update requests now uses a unique ID for each request so that tabs don't get confused by old requests when one tab restarts or whatnot.
Chris Billington
adwin can now do buffered runs
Philip Starkey
* Added some debug messages to track h5file locking issues. Leaving them in because they are very useful! * Reordered some code in the queue manager to only open the h5file once when checking which devices are in use in the current experiment, instead of opening it once for each device tab in BLACS.
Philip Starkey
merge
Philip Starkey
Fixed printing of programming time in the BLACS status bar Fixed a bug in compile_and_restart to print error messages to the output box correctly. That said, it still sometimes doesn't print the non-error messages properly (see comments in the code on line 87 and 89)
Chris Billington
Merge
Chris Billington
Modified a win7 specific line so that it doesn't crash in WinXP when there is no such win32 API call.
Chris Billington
Merge
Chris Billington
Added ADWin files
Chris Billington
Made device tabs with a fatal error not wait until timeout before closing when BLACS is shutting down. They were never going to execute their shutdown anyway, due to their event loop not running. BLACS closes faster.
Philip Starkey
Fixed a bug with the indentation of some code. The NI_pcie_6363 was trying to save wait information even when there were no waits. Indenting the block of code placed it within an already existing conditional statement that checked if we had waits or not! The question I have is that given the WaitMonitorWorker does not complete transition_to_buffered when there are no waits, why does transition to static get called? Guess it doesn't really matter but it seems bizarre!
Philip Starkey
Applied a bugfix that was uncomitted in the SVN local copy on krb-control
Chris Billington
Renamed a variable for consistency
Chris Billington
Acquire kill_lock during the critical section of wait monitoring, so that the process is not killed by lack of heartbeats when it needs to be timing out waits. This ensures a correctly wired experiment will still complete even if the rest of BLACS crashes.
Chris Billington
Everythong about the pineblaster works now, it can be a master clock, or not. It can transition to buffered, static, and all that jazz. Seems to be working.
Chris Billington
Pineblaster debugging
Chris Billington
Removed a bug left in 2 commits ago
Chris Billington
Waits have names now
Chris Billington
Data chunking of waits still not implemented, but wait durations are computed and saved to the file.
Chris Billington
Much work on waits, they work now, but I haven't got data chunking done yet
Chris Billington
Working on waits in BLACS
Chris Billington
Half-implemented wait-monitor process. Committing to move to control computer for testing.
beclab@bec-knarbli-control.physics.monash.edu.au
Fixed another bug from the change. An experiment runs now!
beclab@bec-knarbli-control.physics.monash.edu.au
Fixed a few remaining bugs that stopped BLACS from starting after recent changes
Chris Billington
Fixed bugs in move to zmq IPC. Had to move the logging setup to a separate file where it can be imported for use by all processes (previously they just ran the main file!). This should however be removed in time, as zmq itself should be used for slinging log messages around to a central server. Also fixed path issues that prevented modules from being imported in the child process, and thus prevented classes from being able to be unpickled.
Chris Billington
Fixed bug in pulseblaster, waits_in_use attribute not initialised
Chris Billington
Ported to use zmq for interprocess communication, via the subproc_utils module. This was fairly simple as the subproc_utils.Process class is very similar to the multiprocess.Process class.
Chris Billington
Modified the pulseblaster so that it checks that all waits have completed before it announces that the experiment is over. It does this by polling a subproc_utils.Event, which the wait monitoring device tab will have to send. However if there are no waits in use the pulseblaster tab does not wait for this event -- so current experiments still run, event thogh I haven't written the wait monitor code yet
Chris Billington
BLACS now triggers the master pseudoclock, rather than the hard-coded pulseblaster.
Chris Billington
Made BLACS shut the hell up Its logging is now error or higher in the terminal by default, so you can actually see prinlines if you're debugging.
Chris Billington
Added support for the new event distribution functionality from subproc_utils. BLACS tabs can now instantiate subproc_utils.Event objects and post and wait on events that all subprocesses can see.
Chris Billington
Added pineblaster functions