「view this page in B3 βῆτα server」

Revisions №61031

branch: master 「№61031」
Commited by: Vikram K. Mulligan
GitHub commit link: 「868bf3f9db0e20a0」 「№4345」
Difference from previous tested commit:  code diff
Commit date: 2019-11-08 13:01:12
linux.clang linux.gcc linux.srlz mac.clang
debug
release
unit
PyRosetta4.notebook gcc-9.gcc.python37.PyRosetta4.unit linux.clang.cxx11thread.serialization.python37.PyRosetta4.unit linux.gcc.python36.PyRosetta4.unit mac.PyRosetta.unit build.clean.debug cppcheck alpine.gcc.build.debug gcc-9.gcc.build.debug mysql postgres linux.clang.python36.build.debug linux.zeromq.debug mpi mpi.serialization linux.icc.build.debug OpenCL mac.clang.python36.build.debug build.header build.levels ninja graphics static linux.ui mac.ui build.xcode beautification code_quality.clang_analysis serialization integration.mpi integration.release_debug integration.tensorflow integration.thread integration performance profile release.PyRosetta4.Release release.source scientific.antibody_grafting linux.clang.score linux.gcc.score mac.clang.score linux.scripts.pyrosetta scripts.rosetta.parse scripts.rosetta.validate scripts.rosetta.verify linux.clang.unit.release linux.gcc.unit.release util.apps

Merge pull request #4345 from RosettaCommons/vmullig/make_FASTER_ig_threadsafe Add unit tests confirming that the FASTER interaction graph is threadsafe, and update test to use ref2015 instead of pre_talaris_2013_standard The FASTER interaction graph unit test fails stochastically in the multi-threaded build. This is likely due to a thread safety issue. This PR attempts to fix the issue. EDIT: The FASTER ig is, as far as I can tell, totally threadsafe. (This PR does add a pretty stringent test for this, computing the graph 100 times in threads and comparing each result to the graph computed in a single thread.) The score12 scorefunction, however, is not. I get concurrency issues ONLY when using score12, and these vanish with ref2015 (which is fully threadsafe). For now, I'm switching the FASTER ig unit test to use ref2015 instead of score12, though at some point I'll track down the concurrency issues in score12.