q-pix / qpixrtd Goto Github PK

setup.sh is inefficient. Has the tclsh line that can be removed. also has cmake and make that run each time, when both only need to be ran once. Break setup.sh into build.sh and setup.sh where setup.sh is just the export command and build is the rest.

Concerning source/ROOTFileManager.cpp

We can condense the T_Drift calculation to only use 1 variable as opposed to 2

terminate called after throwing an instance of 'std::out_of_range' what(): vector::_M_range_check: __n (which is 3) >= this->size() (which is 3)

Error thrown when running on single Alpha particles generated with the latest QPixG4.

Seems to be related to number_hits being changed from integer to long (#8 ).

It appears that there is an issue with Boost, at least when using C++17. This seems to be causing issues in building the repository and the example. It appears that we can fix this using

target_compile_options(QPixRTD PRIVATE -DBOOST_NO_CXX11_SCOPED_ENUMS)

in source/CMakeLists.txt.

Also need to change CMakeLists.txt back to include FIND_PACKAGE(Boost 1.61 COMPONENTS system filesystem REQUIRED)

Need to include all the build steps in ./build.sh

the variable number_hits_ in ROOTFileManager is currently of type int, however with beam neutrinos, events could have nHits that exceed the available range of a normal int. Need to change to long long to match new qpixg4 format

Downsampling factors less than 0.1 are causing a systematic shift in the number of pixel resets. This may be due to the fact that the charge variable in /source/PixelResponse.cpp is an integer but needs to be a double to account for downsampling factors.

the scripts are set up poorly, with redundancies that convolute what is actually being done. Need to neaten them up.

Need an executable that can take in values of parameters to be more versatile.

Need to be able to turn on or turn off the ROOTFileManager.cpp:259 for background studies

directory named 'develop'. Need to either remove or rename, so as not to confuse git.

thnsparse branch introduces THnSparse as a way to reduce the amount of time and memory necessary to run qpixrtd. Using PixFunc::Reset_THnSparse does not seem to show much improvement. Potentially because it is still creating a new THnSparse for every pixel and still using a giant vector of electrons.

Proposing:

1. changing this to be one 2D THnSparse (dimensions are pixel_ID and time) created in the example so there is only one THnSparse per event.
2. the vector of electrons is only ever the size of a single hit. electrons are filled into the THnSparse and then reset for next hit
3. The reset function will change to a function where it just integrates to a threshold N and marks the time of the Nth electron, then resets and continues

Concerning ROOTFileManager.cpp, lines 290-334

Problem:
A T_drift is declared and used to calculate the transverse and longitudinal diffusion of each electron, but then when calculating the (x,y,z,t) to convert to pixel and time of arrival to be added to hit_e, the same T_drift is used instead of calculating a new T_drift from the new z of the diffused electron, leading to an extremely small rms of the resulting charge swarm.