CoreNeuron Versions Save

#796 altered the versioning scheme of CoreNEURON: now the NEURON and CoreNEURON versions are pegged to one another.

Bugfix:

• CMake v3.20+ no longer labels NVHPC as PGI. #543

• First stable release integrated with NEURON 8.0 release
• Fix GPU compilation issue when nrn_ghk is used in the MOD files
• Various GPU bugfixes
• Add OpenACC data cleanup on GPU
• Add GitLab CI pipeline and various CI improvements and fixes
• Various reporting enhancements for BBP use cases
• Enable CoreNEURON test in NEURON
• Cmake-format and clang-format improvements across project
• Build release branches with Github actions

• Cmake-format and clang-format improvements across project
• Fix synapse report issue
• Various fixes related to reportinglib and its CI
• Use enum to represent memory layout
• Build release branches with Github actions
• Print memory use from model_size
• Added codecov tests
• Updated credits and license text
• Publish documentation in ReadTheDocs
• Fixes related to NMODL integration
• (Minimum) Changes to avoid floating-point exceptions
• Enable CoreNEURON test in NEURON
• Prevent nrnivmodl running on every build
• Fix GPU compilation issue when nrn_ghk is used in the MOD files
• LFP Calculator
• Various GPU bugfixes
• Add GitLab CI pipeline
• Add OpenACC data cleanup
• nrnran123.h : change which to char to match neuron signature (breaks compatibility with old mod files)
• Add summation reports
• Various CI improvements and fixes

Todo

• Add SEClamp support in CoreNEURON (#381)
• Add support for inserting nmodl flags in nrnivmodl-core calls (#398)
• Allow NetSendBuffer_t to grow (#395)
• Implement support for version from last git commit and bugfix for PR #390 (#392)
• Improve nrnivmodl-core workflow and integration with nmodl/mod2c (#388)
• Bug fix for PR #385 : add missing const for virtual method (#394)
• If a netcon_srcgid is negative, need to determine the thread. (#390)
• Fix warning by returning true in try_lock and remove byteswap1.dat code (#387)
• Copies to NEURON the voltage, i_membrane_, and mechanism data. (#382)
• Removing NRNSTATS and ifdef OPENMP (#385)
• Update README with current execution workflow with NEURON (#374)
• Clear spike vectors after writing or transferring to NEURON (#380)
• Minor fixes : memory leak, free -> delete (#378)
• Added documentaiton for BBCOREPOINTER (#379)
• Avoid linking to C++ MPI libraries (#372)
• GPU build improvements for NEURON integration (#373)
• Remove queue statistics related (dead) code (#375)
• Remove unused code (#376)
• Unitialized pointer, later segmentation fault after freeing. (#377)
• Use coreneuron module from NEURON instead of pc.nrncore_run() (#370)
• CMake improvement : clone hpc-coding-convention module automatically (#368)
• Support for FOR_NETCON (#355)
• Integrate hpc coding conventions and reformat CMake files (#361)
• Remove extracon and multiple, fix #291 (#360)
• Whitelist master for after-merge build to avoid long travis build queue (#362)
• Add docs for CoreNEURON (#359)
• Refactoring of nrn_setup.cpp in separate phases (phase1.cpp / phase2.cpp) (#283)

• Update voltage vector from GPU to CPU if trajectories are recorded (#357)
• Patternstim fix : allow restore phases to provide new pattern-stim (#121)
• Add --verbose option Closes #338 (#341)
• Return spikes vectors to NEURON with in-memory mode transfer (#354)
• Override global seed with user provided one in CLI (#350)
• Fixes for multi-execution in-memory transfer with NEURON (#349)