Skip to content

Actions: cwida/FastLanes

Actions

GALP

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
galp.yaml
102 workflow runs
102 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
GALP on dev (#102) — Add Five Benchmark Time Series Datasets (#60) * update dataset description file and benchmark * add emnist * add fashion mnist * add time series dataset * reload all the dataset * delete mnist related file * update result and restore local modification * delete mnist result * move timeseries dataset to FastLanes_Data * fix header check * recover benchmark * style: format timeseries_test.cpp * fix dir * wrong spell * fix dir * Bump base_seed to 13 * ci: retrigger ch... GALP #102: Commit f0edc10 pushed by peterboncz
1s dev
dev
1s
GALP on dev (#101) — Add MSVC and GCC support & add+improve CI Coverage (incl DEBUG runs) (#69) * Add MSVC support for Windows port ================================= Enable FastLanes to compile with both Clang (Linux/MacOS) and MSVC (Windows). (This is a first step towards vcpkg packaging, which will make including FastLanes in a project easier. We need Windows not only because it is such a modern and nice OS, but also because it is a native platform for DuckDB and we want to use the future vcpk... GALP #101: Commit 896e92c pushed by peterboncz
1s dev
dev
1s
GALP on add-mvsc+gcc-support (#100) — disabling windows11-arm clang GALP #100: Commit bc117d2 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#99) — try fix dll making on clang GALP #99: Commit 9129162 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#98) — Merge branch 'add-mvsc+gcc-support' of github.com:cwida/FastLanes into add-mvsc+gcc-support GALP #98: Commit ab004dd pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#97) — remove CI on ubuntu-arm/clang/Debug/static as it still times out GALP #97: Commit acec3cc pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#96) — fix tidy issues remove ubutu-arm clang debug static target (it is the only one still timing out) GALP #96: Commit a51b1af pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#95) — split physical_operator variant into enc/dec sub-variants to reduce compile times The 247-alternative physical_operator variant caused std::visit() to generate huge dispatch tables (28MB+ object files), making GCC Debug builds time out. Split into enc_physical_operator (~110 alternatives) and dec_physical_operator (~137 alternatives), with physical_operator as a thin wrapper variant over both. Added visit_enc/visit_dec/visit_physical helpers that unwrap the outer ... GALP #95: Commit 5388905 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#94) — Merge branch 'add-mvsc+gcc-support' of github.com:cwida/FastLanes into add-mvsc+gcc-support GALP #94: Commit 0276a95 pushed by peterboncz
2s add-mvsc+gcc-support
add-mvsc+gcc-support
2s
GALP on add-mvsc+gcc-support (#93) — null_map_arr[idx] out-of-bounds in rowgroup_equality_visitor when null_map_arr is empty but data has been padded to 1024 elements by FillMissingValues. Fixed by checking idx < null_map_arr.size() instead of !null_map_arr.empty(). Applied to both typed and string comparisons in src/table/rowgroup.cpp and rust/vendor/fastlanes/src/table/rowgroup.cpp. GALP #93: Commit 3e5e206 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#92) — - include fixes for tidy - fix histogram instantiation GALP #92: Commit 4c1a587 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#91) — make format GALP #91: Commit cb79623 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#90) — Two bugs, two fixes: 1. FSST12 memcpy overread — sizeof(u64) → sizeof(u32) for a u32 res variable (4 files) 2. ODR violation — two Histogram<T> classes in namespace fastlanes with different layouts Renamed the one in analyze_operator.hpp to AnalyzeHistogram<T> (4 files: header, impl, .cpp, plus rust vendor copies) GALP #90: Commit 59c4581 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#89) — one more forgotten file GALP #89: Commit 2d2f6e9 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#88) — add more forgotten files GALP #88: Commit 44e3859 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#87) — forgot to add files GALP #87: Commit c9009c6 pushed by peterboncz
6s add-mvsc+gcc-support
add-mvsc+gcc-support
6s
GALP on add-mvsc+gcc-support (#86) — split to reduce compilation effort split the four big operator files in encoder/decoder classes, such that: (1) we get more compilation paralellism, and (2) reduce the amount of template instantiations by half (in Debug this will count especially) GALP #86: Commit 0ebc7f4 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#85) — - shift ubuntu/Debig builds from gcc to clang (gcc times out) - hide all non-exported symbols since this may avoid certain bugs GALP #85: Commit b775c78 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#84) — The FSST12 encoder used memcpy(out, &res, sizeof(u64)) to speculatively write 8 bytes of packed 12-bit codes, but res was declared as u32 (4 bytes), causing a 4-byte overread from the stack variable. On x64 MSVC Release/static the /GS security cookie catches this and raises STATUS_STACK_BUFFER_OVERRUN (0xc0000409), crashing EQUALITY_STRPT and SINGLE_COLUMN_STRPT. Fix: widen res from u32 to u64 so the 8-byte memcpy reads from a valid 8-byte variable. Changed in ... GALP #84: Commit 04ecfcc pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#83) — fix frequency coding GALP #83: Commit 4f5e820 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#82) — fix bimap_frequency and min/max lost after Cast due to Finalize/Cast reorder Moving Finalize before Cast (needed for min/max in type-narrowing) left bimap_frequency empty on cast columns, crashing the encoder with "Value not found in BiMapFrequency." It also left min/max at defaults on newly created cast columns, causing silent data corruption on round-trip. Fix: split bimap_frequency population and min/max computation into a new PopulateBiMap() step that runs aft... GALP #82: Commit 78dfcac pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#81) — fix unsigned overflow in is_exception() causing all values to be marked as exceptions When base + 2^bw overflowed the unsigned type (e.g. base=64512, bw=10 for uint16_t: 64512+1024=65536 wraps to 0), the upper bound check val >= 0 became true for all values, producing 1024 exceptions per vector instead of the expected 0-15. This silently inflated compressed output in Release (assertions disabled) and caused assertion failures Fix: replace the overflow-prone upper ... GALP #81: Commit 6987a1a pushed by peterboncz
2s add-mvsc+gcc-support
add-mvsc+gcc-support
2s
GALP on add-mvsc+gcc-support (#80) — one more out of bounds error GALP #80: Commit 73d8533 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#79) — another turn of the wheel GALP #79: Commit 77d1ab2 pushed by peterboncz
1s add-mvsc+gcc-support
add-mvsc+gcc-support
1s
GALP on add-mvsc+gcc-support (#78) — bugs found by windows testing GALP #78: Commit bb719c7 pushed by peterboncz
2s add-mvsc+gcc-support
add-mvsc+gcc-support
2s