FLAG issueshttps://gitlab.ras.byu.edu/ras-devel/flag/-/issues2019-05-31T14:36:41Zhttps://gitlab.ras.byu.edu/ras-devel/flag/-/issues/8Kernel reports "page allocation errors" of several processes when processing2019-05-31T14:36:41ZMitch BurnettKernel reports "page allocation errors" of several processes when processinghttps://gitlab.ras.byu.edu/ras-devel/flag/-/issues/7bfFitsWriter process reports a segmentation fault when being forcefully closed2019-05-31T14:36:17ZMitch BurnettbfFitsWriter process reports a segmentation fault when being forcefully closedhttps://gitlab.ras.byu.edu/ras-devel/flag/-/issues/6Hashpipe process reports a segmentation fault in libbeamformer.so2019-05-31T14:35:44ZMitch BurnettHashpipe process reports a segmentation fault in libbeamformer.sohttps://gitlab.ras.byu.edu/ras-devel/flag/-/issues/5Clean up install script/process2019-05-28T19:53:17ZMitch BurnettClean up install script/processRight now the default way to install FLAG is to source the dibas.bash script in /home/groups/flag/dibas to setup the environment and then run the FLAG `install` script in the root level repo directory. However, if you wanted to install F...Right now the default way to install FLAG is to source the dibas.bash script in /home/groups/flag/dibas to setup the environment and then run the FLAG `install` script in the root level repo directory. However, if you wanted to install FLAG in a location other than `/home/groups/flag/hash/ you would have to update your path for that session. And so this is kind of a chicken and the egg situation.
It seems the best thing to do may be to have the dibas.conf script generated by the `install` script that is then sourced after building each time using flag.https://gitlab.ras.byu.edu/ras-devel/flag/-/issues/4Translate beamformer weight code to python command-line tools2019-04-23T21:30:25ZMitch BurnettTranslate beamformer weight code to python command-line toolsCurrently the beamformer weight calibration process is implemented in Matlab. This should be converted to Python fir integration into the system.Currently the beamformer weight calibration process is implemented in Matlab. This should be converted to Python fir integration into the system.https://gitlab.ras.byu.edu/ras-devel/flag/-/issues/3Gracefully manage hangs2019-05-31T16:56:16ZMitch BurnettGracefully manage hangsResolving processing pipeline hangs (#1) is a critical task. Our hashpipe plugin implementing a state machine has an unknown transition resulting in a hang. The idea is to add logic to detect that we are hanging and to notify the user (o...Resolving processing pipeline hangs (#1) is a critical task. Our hashpipe plugin implementing a state machine has an unknown transition resulting in a hang. The idea is to add logic to detect that we are hanging and to notify the user (or log the event) and gracefully catch up and continue processing.
As mentioned in the reported hanging issue (#1) one case is identified in the net thread and signaled with the print statement `NET: HANGING HERE!!!`. An approach might be to exam that loop where the statement is first printed and then determine the ways we can enter that loop and identify if we can exit and begin processing again.https://gitlab.ras.byu.edu/ras-devel/flag/-/issues/1Processing pipeline hangs2019-05-31T16:54:37ZMitch BurnettProcessing pipeline hangsOne of the major problems we have in the processing pipeline are the hangs in our hashpipe plugin.
The reason why it hangs is unknown. We are unsure if it is limits in the network, memory limits being reached, or if we are to slow on th...One of the major problems we have in the processing pipeline are the hangs in our hashpipe plugin.
The reason why it hangs is unknown. We are unsure if it is limits in the network, memory limits being reached, or if we are to slow on the I/O transfers on and off the GPU device (i.e., I/O bandwidth). More work needs to be done to identify the stalls.
Currently there is one case of stalling that is reported by the net thread (it may be all cases that this occurs, but not sure). This is a print statement saying `NET: HANGING HERE!!!`. This is in the net thread when waiting for either the input or output buffer to be free. To get into this state we missed are hanging up stream somewhere and the semaphore is not triggering to release the buffers.