Outputting to file #52
Labels
ready for review
ready for review
Comments
bencopl
added a commit
that referenced
this issue
Aug 24, 2022
bencopl
added a commit
that referenced
this issue
Aug 31, 2022
bencopl
added a commit
that referenced
this issue
Sep 1, 2022
|
Accepting suggestions for what to call both the default file name and the optional environment variable! |
bencopl
added a commit
that referenced
this issue
Sep 9, 2022
|
Note: MPI stuff currently hard-coded in and trying to compile a test case normally with just |
bencopl
added a commit
that referenced
this issue
Oct 4, 2022
bencopl
added a commit
that referenced
this issue
Oct 4, 2022
bencopl
added a commit
that referenced
this issue
Oct 4, 2022
bencopl
added a commit
that referenced
this issue
Oct 4, 2022
bencopl
added a commit
that referenced
this issue
Oct 13, 2022
Closed
bencopl
added a commit
that referenced
this issue
Nov 3, 2022
bencopl
added a commit
that referenced
this issue
Nov 9, 2022
Closed
Closed
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
bencopl
added a commit
that referenced
this issue
Nov 10, 2022
andrewcoughtrie
pushed a commit
that referenced
this issue
Apr 18, 2023
* Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Double stdfunction strategy pattern that includes ouput format (#72) * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Adding new headers to public header files list (#72) * Acting on review, slight change to where clock stuff lives (#72) * Reverting top_walltime calculation changes (#72) * Added some end-of-file blank lines (#72) * Removing erroneous include from profiler.h (#72)
andrewcoughtrie
added a commit
that referenced
this issue
Jun 23, 2023
* Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]>
andrewcoughtrie
added a commit
that referenced
this issue
Aug 1, 2023
* 'meto' namespace added into all source code (#49) * Bring fork up-to-date (#13) * 85 Fix failing workflows (#87) * Installing mpich and ammending final job step (#85) * Keeping action versions and ubuntu versions consistent (#85) * Checking clang version during build (#85) * Testing higher clang version (#85) * Testing new libomp and clang versions (#85) * Reverting previous tests and specifying mpich version (#85) * Restoring build workflow to original but with minor version tweaks (#85) * Testing ubuntu 20.04 (#85) * Update CHANGELOG.md * Better overhead measurement (#77) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]> --------- Co-authored-by: Andrew Coughtrie <[email protected]> Co-authored-by: Maff Glover <[email protected]> * Adding meto namespace to all files (#49) * Fixing erroneous lines resulting from merge (#49) * Removing unnecessary namespace prefixes within functions (#49) * Removing any unnecessary meto prefixes in header files (#49) * Missed profilercpp from prev commit (#49) * Unnecessary 'meto::' in C interface (#49) * Bring fork up-to-date (#13) * 85 Fix failing workflows (#87) * Installing mpich and ammending final job step (#85) * Keeping action versions and ubuntu versions consistent (#85) * Checking clang version during build (#85) * Testing higher clang version (#85) * Testing new libomp and clang versions (#85) * Reverting previous tests and specifying mpich version (#85) * Restoring build workflow to original but with minor version tweaks (#85) * Testing ubuntu 20.04 (#85) * Update CHANGELOG.md * Better overhead measurement (#77) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]> --------- Co-authored-by: Andrew Coughtrie <[email protected]> Co-authored-by: Maff Glover <[email protected]> * Adding meto namespace to all files (#49) * 'meto' namespace added into all source code (#49) * Fixing erroneous lines resulting from merge (#49) * Removing unnecessary namespace prefixes within functions (#49) * Unnecessary 'meto::' in C interface (#49) * Had to reimplement quite a few of the namespaces due to conflicts with a previous PR. * Some small bug fixes --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Maff Glover <[email protected]> Co-authored-by: MetBenjaminWent <[email protected]>
andrewcoughtrie
pushed a commit
that referenced
this issue
Aug 23, 2023
* Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Farting the current output into a hard-coded file (#52) * Farting the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Working byte copying, with appropriate changes to unit tests. * Tweaks to typecasting (remove build warning messages). * Use PROF_STRING_BUFFER_LENGTH. * Return overhead time pointers through argument list, not returned from functions. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) * Match integer types in assertion (new_chars_size == expected_size). * Patch string_view -> string. * Add @returns tag to compute_hash(). * Remove delimiters from hashtable hashes. These are not human-readable, so there's not much point incurring any overhead, however small, of adding them. * Add tid to comment blocks, and one int -> auto. * Patch missing include <cstddef> --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]>
andrewcoughtrie
pushed a commit
that referenced
this issue
Jun 7, 2024
* Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Double stdfunction strategy pattern that includes ouput format (#72) * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Adding new headers to public header files list (#72) * Acting on review, slight change to where clock stuff lives (#72) * Reverting top_walltime calculation changes (#72) * Added some end-of-file blank lines (#72) * Removing erroneous include from profiler.h (#72)
andrewcoughtrie
added a commit
that referenced
this issue
Jun 7, 2024
* Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]>
andrewcoughtrie
added a commit
that referenced
this issue
Jun 7, 2024
* 'meto' namespace added into all source code (#49) * Bring fork up-to-date (#13) * 85 Fix failing workflows (#87) * Installing mpich and ammending final job step (#85) * Keeping action versions and ubuntu versions consistent (#85) * Checking clang version during build (#85) * Testing higher clang version (#85) * Testing new libomp and clang versions (#85) * Reverting previous tests and specifying mpich version (#85) * Restoring build workflow to original but with minor version tweaks (#85) * Testing ubuntu 20.04 (#85) * Update CHANGELOG.md * Better overhead measurement (#77) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]> --------- Co-authored-by: Andrew Coughtrie <[email protected]> Co-authored-by: Maff Glover <[email protected]> * Adding meto namespace to all files (#49) * Fixing erroneous lines resulting from merge (#49) * Removing unnecessary namespace prefixes within functions (#49) * Removing any unnecessary meto prefixes in header files (#49) * Missed profilercpp from prev commit (#49) * Unnecessary 'meto::' in C interface (#49) * Bring fork up-to-date (#13) * 85 Fix failing workflows (#87) * Installing mpich and ammending final job step (#85) * Keeping action versions and ubuntu versions consistent (#85) * Checking clang version during build (#85) * Testing higher clang version (#85) * Testing new libomp and clang versions (#85) * Reverting previous tests and specifying mpich version (#85) * Restoring build workflow to original but with minor version tweaks (#85) * Testing ubuntu 20.04 (#85) * Update CHANGELOG.md * Better overhead measurement (#77) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Andrew Coughtrie <[email protected]> --------- Co-authored-by: Andrew Coughtrie <[email protected]> Co-authored-by: Maff Glover <[email protected]> * Adding meto namespace to all files (#49) * 'meto' namespace added into all source code (#49) * Fixing erroneous lines resulting from merge (#49) * Removing unnecessary namespace prefixes within functions (#49) * Unnecessary 'meto::' in C interface (#49) * Had to reimplement quite a few of the namespaces due to conflicts with a previous PR. * Some small bug fixes --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]> Co-authored-by: Maff Glover <[email protected]> Co-authored-by: MetBenjaminWent <[email protected]>
andrewcoughtrie
pushed a commit
that referenced
this issue
Jun 7, 2024
* Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Writing the current output into a hard-coded file (#52) * Writing the current output into a hard-coded file (#52) * Support for setting the outfile name via an environment variable (#52) * MPI support, one output file per rank (#52) * Tidying up profiler's write function (#52) * Initial workflow [checking gcov works before messing with lcov] (#45) * Cosmetic changes to yml file * Fixing error messsage from actions tab * Typo in yml file * Disabling doxygen, explicitly declaring compilers using matrix method * Including GTest * Disabling fortran tests in cmake config * Altering gcov flags * Attempting conda install of gcov and lcov * Merging into single job * Fixed typo * Removing environment file line from conda setup * Removing conda install of gcov * Including conda forge channel * Attempting to fix lcov command not working by including shell -l {0} in workflow step * Also adding shell command to genhtml step, workflow should now run all the way through * Trying default workspace option part-way through job * Removing steps from line56 * Added default shell * Removing defaults option... * Added step that updates github pages * Resolving unknown branch issue * Typo in final step * Removing git remote update line * Fixing error msg when github tries to remove folder that isnt there * Removing git prefix from rm -rf * Altering docs folder * Filtering out STL headers * Fixed typo in shell option * Attempt to include fortran tests * Scrapping fortran tests * Getting Doxygen AND coverage docs to appear * Fixed error in documentation.yml * Changing where doxygen documentation is put * Testing putting both workflows straight into docs * Separating them into documentation and coverage files * Typo in coverage.yml * Another typo in coverage.yml * Hopefully final error with coverage.yml * Overhead times reported from all threads. * Draft of new IO structure contained within the two new files (#52) * Formatting mostly there but not working with multiple threads (#70) * Draft solution for combining all hashtables together (#70) * Added HashVec class that collects all hashtables, then sorts & writes entries (#70) * Adding banners and fixing includes (#70) * Switching to getter that returns a reference (instead of using pointers) (#70) * Improved comments (#70) * Capture Fortran time (appending null character). * Adding in PLACEHOLDER block of info above table (#70) * Switching to functions rather than single-method objects (#52) * Move post final gettime time updates to pointer / ref increments. * Attempting to debug mpi issue in workflow (#52) * Installing mpich in workflow (#52) * Changing workflow step name to reflect mpich installation (#52) * Slight comment changes (#52) * Printing mpich version within workflow (#52) * Trying to print OpenMP version (#52) * Updating readme with correct OpenMP version number (#52) * Change the traceback from std::vector to std::array. * Catch the case where a stop calliper is called before a start calliper. * Double stdfunction strategy pattern that includes ouput format (#72) * Correct handling of child time for top-level calls. * Draft of visitor pattern solution (#72) * Better comment blocks in visitor pattern solution (#72) * Slight change in how the visitor method is called (#72) * Combination of std::function and virtual strategy methods (#72) * Updating comment blocks (#72) * Tidying up and updating call count unit test (#72) * Adding some const's where appropriate and slight comment changes (#72) * Modifying string concatenation (#72) * Persistent entries for each region stored in vector; use iterators. * Fixing mistake in string concatenation (#72) * Erroneous line in HashTable::append_to() (#72) * Better timings - don't rehash the hashes (plus a few consts). * Changing final loop in call count unit test (#72) * Removing accidental addition of new coverage worklow (#72) * Fixing documentation workflow based on removal of coverage one (#72) * Hopeful solution to call count unit test error (#72) * Modifications to query_insert() * Altering call count test - removing pair from vector (#72) * Final tidying up (#72) * Index-based addressing for the vector. Sorting needs attention. * Correct lookup_table's view of hashvec indices after sort. * Separate sort method; TracebackEntry; other tweaks. * Modified comment blocks, plus other tweaks. * Remove unused variable. * Don't include asserts that might trap application-level coding faults. * Add test for case where traceback array exhausted. * Add [[maybe_unused]] in traceback entry exhaustion test. * Tidy up around start parts and namespacing threadprivate storage. * Minor correction to commenting. * Merging suggested changes from MG branch (#72) * Minor cosmetic/comment additions or adjustments (#72) * Accidentally removed required include in last commit (#72) * Corrections following merge of formatting changes on main. * Add __profiler__ region name test, alongside tests for other regions. * Move RegionRecord constructor. * Working byte copying, with appropriate changes to unit tests. * Tweaks to typecasting (remove build warning messages). * Use PROF_STRING_BUFFER_LENGTH. * Return overhead time pointers through argument list, not returned from functions. * Forward declare HashVecHandler; use hashvec_t in one place; TracebackEntry() = default. * Corrections to doxygen comments. * Make start_part1 and start_part2 methods private in Profiler class. * Return overhead time pointers through argument list. * Remove extraneous hash of __profiler__. * add_child_time -> add_child_time_to_parent. * reference -> pointer in comments in one place. * File-scope variables become static data members. * Overloaded const and non-const hash2record. (Can have same name.) * Match integer types in assertion (new_chars_size == expected_size). * Patch string_view -> string. * Add @returns tag to compute_hash(). * Remove delimiters from hashtable hashes. These are not human-readable, so there's not much point incurring any overhead, however small, of adding them. * Add tid to comment blocks, and one int -> auto. * Patch missing include <cstddef> --------- Co-authored-by: ben.copley <[email protected]> Co-authored-by: Ben Copley <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This is potentially a duplicate of Issue #9 but less broad, and more specifically focused on simply taking the current output and putting it in a default or user-defined file using
<fstream>.The text was updated successfully, but these errors were encountered: