Repository navigation

Impement advanced CFL reachability algorithm and optimization matrix - #5

Open

homka122 wants to merge 172 commits into

stablefrom

homka122/CFL_optimizations

Open

Impement advanced CFL reachability algorithm and optimization matrix#5
homka122 wants to merge 172 commits into
stablefrom
homka122/CFL_optimizations

Commits on Oct 16, 2025
- Feat: intial commit with draft version of CFL algorithm with
  homka122
  committed
  ade1e99
- Feat: add algorithm with optimization to header
  homka122
  committed
  8b16e59
- Feat: add measure time macro
  homka122
  committed
  2cdcee7
- Feat: add null check to mxm
  homka122
  committed
  9548c1e
- Refactor: replace apply to ewiseAdd
  homka122
  committed
  4660261
- Feat: replace apply to dup method
  homka122
  committed
  6bd45f2
- Fix: type in printf while iteration
  homka122
  committed
  76bc80a
- Refactor: apply mask now seperate function
  homka122
  committed
  4115a05
- Refactor: now bench information output only with flag
  homka122
  committed
  838ad5b
- Feat: add ISO check to rsub
  homka122
  committed
  7d113dd
- Refactor: iteration info now print if flag set
  homka122
  committed
  0433361
- Feat: add null check in first mxm
  homka122
  committed
  155e609
- Feat: now empty optimization is seperate
  homka122
  committed
  8dc7d1d
- Feat: add rsub, wise and dup with format optimization
  homka122
  committed
  08ab76c
- Feat: add mxm format optimization
  homka122
  committed
  4668c88
- Refactor: delete unused comments
  homka122
  committed
  45aaf82
- Fix: change NULL to GrB_NULL
  homka122
  committed
  fe16527
- Fix: change dup method according to format optimization
  homka122
  committed
  b198dd3
- Feat: add clear methods with optimization
  homka122
  committed
  54ba2b8
- Feat: add change format methods
  homka122
  committed
  9c7c923
- Feat: process result of matrix_mxm
  homka122
  committed
  067fb3a
- Fix: now all methods in empty wise use format methods
  homka122
  committed
  ba335af
- Refactor: change rsub method with format optim
  homka122
  committed
  8877469
- Refactor: change wise method with format optimization
  homka122
  committed
  a5b9f0e
- Fix: fix wrong type in for cycle
  homka122
  committed
  0a59d20
- Fix: change GrB clear method to mine
  homka122
  committed
  0bc599f
- Feat: add rmxm method
  homka122
  committed
  135ba92
- Fix: incompitible type of for cycle
  homka122
  committed
  e85a7c1
- Refactor: delete unused variables
  homka122
  committed
  e95dbad
- Feat: delete unused sync format method
  homka122
  committed
  3756ed9
- Feat: use more optimize methods in wise method
  homka122
  committed
  539260d
- Feat: manually choose format in wise and dup method
  homka122
  committed
  007c6b7
- Fix: wrong type in for cycle
  homka122
  committed
  36ffc72
- Feat: now update Matrix object in place
  homka122
  committed
  a180624
- Feat: delete seperate rmxm method
  homka122
  committed
  cf89d6a
- Refactor: matrix mxm with swap operation
  homka122
  committed
  901323d
- Feat: add lazy optimization
  homka122
  committed
  61f90aa
- Fix: change < to <=
  homka122
  committed
  fd2502e
- Feat: use lazy optimization
  homka122
  committed
  7b0a5e1
- Feat: add flags for turn on optimizations on runtime
  homka122
  committed
  935baef
- Feat: matrix_print_lazy now accepts pointer
  homka122
  committed
  aa0dd4a
- Feat: add naming for matrices in methods
  homka122
  committed
  7184953
- Feat: lazy optimizations now works
  homka122
  committed
  66d55a8
- Fix: erase matrix when accum true in mxm
  homka122
  committed
  6d97d82
- Feat: add sort and combine function to lazy optimization
  homka122
  committed
  146d58c
- Fix: some memory leaks
  homka122
  committed
  bd2c142
- Feat: now matrix has not just mxm size
  homka122
  committed
  c0eae50
- Feat: Add fields to Matrix object for block matrices
  homka122
  committed
  bc0af0d
- Feat: change rule parsing
  homka122
  committed
  eb313af
- Feat: implement block mxm operation
  homka122
  committed
  9be8a15
- Fix: first mxm need accum operator for correct result
  homka122
  committed
  cc23b24
- Feat: implement wise block operation
  homka122
  committed
  e6a77bf
- Feat: add create and free methods for Matrix structure
  homka122
  committed
  8249fc1
- Feat: implement rsub block operation
  homka122
  committed
  6225f49
- Feat: add block hyper rotate method
  homka122
  committed
  3595a8f
- Feat: add block matrix to diag method
  homka122
  committed
  c0adbab
- Feat: add block matrix reduce method
  homka122
  committed
  caddbf1
- Fix: wrong type input
  homka122
  committed
  b7e68b8
- Feat: add method matrix -> block vector
  homka122
  committed
  3359d61
- Fix: fix bug when dublicate adj matrices
  homka122
  committed
  c79861f
- Fix: compiler warnings
  homka122
  committed
  5f96e31
- Fix: copy graph matrix from ajd
  homka122
  committed
  6b25fa4
- Feat: add block optimization choice
  homka122
  committed
  c030cca
- Feat: now implicitly declare that graph matrices are lazy
  homka122
  committed
  3aea853
- Fix: is lazy check in rsub function
  homka122
  committed
  68301f9
- Fix: update base matrix nnz before calc in lazy matrix
  homka122
  committed
  d734683
- Refactor: algorithm now accepth symbol count
  homka122
  committed
  7c17e47
- Feat: sort matrices before find acc
  homka122
  committed
  8440287
- Feat: new way checking changing graph
  homka122
  committed
  9cf80c2
- Refactor: make more useful matrix printing
  homka122
  committed
  b99cc1b
- Feat: add simple rules adding to matrices
  homka122
  committed
  ee4ffa4
- Refactor: move custom methods definition
  homka122
  committed
  e166fc1
- Feat: now delta matrices dublicates from adj matrices
  homka122
  committed
  5b73708
- Refactor: rework print lazy method
  homka122
  committed
  5502c89
- Fix: check is lazy status in mxm lazy method
  homka122
  committed
  6ce43b0
- Feat: add iadd interpration to wise methods
  homka122
  committed
  b0d3f53
- Fix: problems with sorting in mxm lazy
  homka122
  committed
  5100773
- Fix: update matrix after combine base matrices
  homka122
  committed
  2e63e5f
- Refactor: now use TRY macro
  homka122
  committed
  08a19c9
- Fix: return from block mxm is both matrices are cell
  homka122
  committed
  df5cdea
- Fix: use lazy matrices with lazy optimization and higher
  homka122
  committed
  8a1b404
- Fix: wise_block
  homka122
  committed
  88f25bd
- Fix: forget update matrix after changing
  homka122
  committed
  c05d6c0
- Fix: now if symbol is indexed matrix is block
  homka122
  committed
  e0841c0
- Feat: add print_graph_info function
  homka122
  committed
  61ada93
- Fix: add method for dup block matrices
  homka122
  committed
  22e4c4c
- Fix: bug that lazy matrix didn't create
  homka122
  committed
  8c85e4e
- Fix: matrix mxm now can multiply not only square matrices
  homka122
  committed
  6b59c0f
- Feat: create temp matrix in mxm block
  homka122
  committed
  46608e1
- Refactor: matrix dup now uses a different method
  homka122
  committed
  513aeca
- Fix: hyper rotate method
  homka122
  committed
  f49a6b2
- Feat: add aditional check in matrix update
  homka122
  committed
  697d23c
- Fix: assert condition in rsub block
  homka122
  committed
  6865893
- Feat: check result of rsub methods
  homka122
  committed
  3bf3ba4
- Fix: fix order of flags
  homka122
  committed
  595a2cb
- Refactor: delete unused T matrix from prev algo
  homka122
  committed
  af498f8
- Refactor: delete whitespaces after if statement
  homka122
  committed
  2b8337b
- Feat: move functions with optimized matrices in seperate file
  homka122
  committed
  8f57591
- Refactor: move optimizations methods to seperate file
  homka122
  committed
  b87cead
- Refactor: delete unused code and fix types of structures
  homka122
  committed
  2334b27
- test: add tests for adv algorithm
  homka122
  committed
  c7ba5c0
- Fix: delete files with algorithm
  homka122
  committed
  9941af1
Commits on Oct 17, 2025
- Fix: delete unused files
  homka122
  committed
  28c6696
Commits on Oct 31, 2025
- fix: update nvals before reading it
  spisladqo
  committed
  94aecfb
Commits on Nov 3, 2025
- test: add tests for optimized matrix
  homka122
  committed
  89756bf
- Test: add tests for matrix_opt: wise, rsub, format optimization
  homka122
  committed
  aa2d64a
- Test: add tests for matrix_opt: empty optimization
  homka122
  committed
  ae1daff
- Refactor: now matrix_to_base don't modify input matrix
  homka122
  committed
  39bf9da
- Feat: add dup_lazy
  homka122
  committed
  9608e2b
- Fix: delete unused sort of matrices
  homka122
  committed
  f1b4a5c
- Fix: delete unreacheable block of code in rsub
  homka122
  committed
  098bc8c
- Test: add tests for lazy optimization
  homka122
  committed
  a6e0148
Commits on Nov 4, 2025
- Refactor: reformat matrix_opt file
  homka122
  committed
  15bdb6b
- Refactor: comment print debug functions
  homka122
  committed
  eb151f3
- Test: add block tests for matrix_opt
  homka122
  committed
  d13785c
- Test: refactor tests for matrix_opt for coverage
  homka122
  committed
  29bfba9
- Fix: memory leaks in matrix_opt and in tests for opts
  homka122
  committed
  cd90d5e
Commits on Mar 7, 2026
- Feat: change API with opt methods
  homka122
  committed
  92f5cbd
- Fix: algorithm change input adj_matrices
  homka122
  committed
  793ed1e
- Refactor: apply clang-format to code
  homka122
  committed
  d3623c6
- Fix: change matrix_to_base function
  homka122
  committed
  d5ffb75
- Fix: add adv algorithm signature to LAGraphX.h
  homka122
  committed
  37df68c
- Test: add tests for adv CFL reachability algorithm
  homka122
  committed
  e948f0d
- Feature: add LAGraph_rule_EWCNF
  homka122
  committed
  4c65734
- Feat: now adv algorithm can process EWCNF rules
  homka122
  committed
  1b58a08
- Fix: fix adv algo tests with EWCNF rules
  homka122
  committed
  4dd5617
- Feat: add to print_matrix utility func optimization flag
  homka122
  committed
  edefac4
- Fix: fix memory leaks in adv algorithm
  homka122
  committed
  ebb46aa
- Fix: fix memory leak with symbols
  homka122
  committed
  2b9eee3
- Fix: fix dimension dismatch to output part of algorithm
  homka122
  committed
  d1c8b6c
- Fix: rewrite opt operaions for memory leaks fix
  homka122
  committed
  7936f15
- Fix: bug with unupdate matrix
  homka122
  committed
  dd9fc51
- Feat: add TRY_I macros
  homka122
  committed
  d6cdc3e
- Feat: add BENCH_CFL_REACHBILITY macro for debug purpose
  homka122
  committed
  c79a9c4
- Refactor: now use TRY macro for every GrB\LG call
  homka122
  committed
  e2fa0b7
- Refactor: change calls to opt methods because of changing API
  homka122
  committed
  84a32c8
- Test: add new grammars for indexed rules
  homka122
  committed
  6996a34
- Test: add two new graphs for tests
  homka122
  committed
  b78137a
- Test: add new indexed tests for adv CFL algo
  homka122
  committed
  cf51642
- Fix: fix matrix_update call with new API
  homka122
  committed
  682a23c
- Fix: fix wrong free symbols in LG_FREE_WORK
  homka122
  committed
  4162dd8
- Test: fix memory leaks and heap overflow in tests
  homka122
  committed
  8c8f7e4
- Refactor: delete unuserd debug prints
  homka122
  committed
  1cc6cef
- Refactor: use new CFL_create functions
  homka122
  committed
  101421b
- Refactor: simplify getting outputs matrices
  homka122
  committed
  812fd21
- Refactor: add documentations of inner functions in adv CFL algo
  homka122
  committed
  945e98f
- Refactor: add TRY_INNER macro for helper functions
  homka122
  committed
  04cbe89
- Feat: add CFL matrix split method
  homka122
  committed
  1c832d5
- Docs: adjust description of adv CFL reachability algorithm
  homka122
  committed
  f388754
- Fix: change initial capacity for dynamic arrays
  homka122
  committed
  b7c047a
- Test: add new indexed test
  homka122
  committed
  adde8fd
- Fix: revert CMakeLists.txt changes
  homka122
  committed
  1c4072e
- Refactor: now making new symbols is seperate function
  homka122
  committed
  bb634eb
- Refactor: now making new adj matrices in seperate function
  homka122
  committed
  ff5b8fa
- Refactor: now making new rules in seperate function
  homka122
  committed
  f1cdd8b
Commits on Mar 8, 2026
- Refactor: move CFL methods to seperate header file
  homka122
  committed
  ddd29f5
- Refactor: fix TODO in matrix opt methods
  homka122
  committed
  20e27c9
- Refactor: refactor LG_FREE_ALL macro in adv CFL algo
  homka122
  committed
  3b8afe5
- Refactor: comment unused code in matrix opt
  homka122
  committed
  3b3c7c7
- Tests: add tests for block format hyper rotate
  homka122
  committed
  1c90854
- Tests: fix test and add new for adv algo
  homka122
  committed
  e050117
Commits on Sep 29, 2026
- Fix: validate rules before symbol grouping in adv CFL algo
  homka122
  committed
  bc51b0b
- Fix: support cell multiplication into block vector
  homka122
  committed
  b8ad6ee
- Fix: handle null message in adv CFL algo
  homka122
  committed
  ad09795
- Fix: free workspace on allocation failure in adv CFL algo
  homka122
  committed
  246593a
- Fix: preserve error message during cleanup in adv CFL algo
  homka122
  committed
  475b0f8
- Fix: prevent overflow in lazy CFL matrix components
  homka122
  committed
  aa72ae6
Commits on Sep 30, 2026
- Feat: skip symbols unused by rules in CFL_reachability_adv
  homka122
  committed
  7b9236c
Commits on Oct 1, 2026
- Feat: move nonterminal matrices to outputs in CFL_reachability_adv
  homka122
  committed
  57c54e4
Commits on Oct 11, 2026
- Fix: keep the column format after rotating a block matrix
  homka122
  committed
  7c8160b
Commits on Oct 13, 2026
- Fix: clear the output of a block mxm without accumulation
  homka122
  committed
  66f9145
- Fix: drop the stale format copy after a single-format write
  homka122
  committed
  9bd84d0

Commits

Commits on Oct 16, 2025

Commits on Oct 17, 2025

Commits on Oct 31, 2025

Commits on Nov 3, 2025

Commits on Nov 4, 2025

Commits on Mar 7, 2026