vortex-data
vortex
Blog
Docs
Changelog
Blog
Docs
Changelog
Overview
Branches
Benchmarks
Runs
Performance History
Latest Results
DecimalArray Signed-off-by: Andrew Duffy <andrew@a10y.dev>
aduffy/bufferhandle-more
4 hours ago
fix: canonicalize sliced VarBinArray failure Signed-off-by: Andrew Duffy <andrew@a10y.dev>
varbin-arrow-fix
4 hours ago
fix: canonicalize sliced VarBinArray failure Signed-off-by: Andrew Duffy <andrew@a10y.dev>
varbin-arrow-fix
4 hours ago
fix: canonicalize sliced VarBinArray failure Signed-off-by: Andrew Duffy <andrew@a10y.dev>
varbin-arrow-fix
5 hours ago
fix: FSST canonicalize when > 2GB of buffers Signed-off-by: Andrew Duffy <andrew@a10y.dev>
fsst-canonical
5 hours ago
Optimize BitBuffer::select with block8 + bidirectional search Key optimizations: - Block8: Process 8 chunks at a time for instruction-level parallelism - Bidirectional: Search from end if n > 50% of true_count (up to 60x speedup) - Fix select_in_word_reverse to correctly find nth bit from end Also adds select_with_true_count() for Mask::rank to use the cached true_count value, avoiding redundant popcount computation. Includes comprehensive test suite with 58 select tests covering: - Aligned and unaligned buffers - Block8 boundary conditions - Various offset scenarios - Forward/reverse consistency Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
claude/explain-mask-rank-function-9Zbxp
5 hours ago
Optimize BitBuffer::select with block8 + bidirectional search Key optimizations: - Block8: Process 8 chunks at a time for instruction-level parallelism - Bidirectional: Search from end if n > 50% of true_count (up to 60x speedup) - Fix select_in_word_reverse to correctly find nth bit from end Also adds select_with_true_count() for Mask::rank to use the cached true_count value, avoiding redundant popcount computation. Includes comprehensive test suite with 58 select tests covering: - Aligned and unaligned buffers - Block8 boundary conditions - Various offset scenarios - Forward/reverse consistency Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
claude/explain-mask-rank-function-9Zbxp
5 hours ago
Optimize BitBuffer::select with block8 + bidirectional search Key optimizations: - Block8: Process 8 chunks at a time for instruction-level parallelism - Bidirectional: Search from end if n > 50% of true_count (up to 60x speedup) - Fix select_in_word_reverse to correctly find nth bit from end Also adds select_with_true_count() for Mask::rank to use the cached true_count value, avoiding redundant popcount computation. Includes comprehensive test suite with 58 select tests covering: - Aligned and unaligned buffers - Block8 boundary conditions - Various offset scenarios - Forward/reverse consistency Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
claude/explain-mask-rank-function-9Zbxp
5 hours ago
Active Branches
more bufferhandle
last run
4 hours ago
#5988
CodSpeed Performance Gauge
0%
fix: canonicalize sliced VarBinArray failure
last run
4 hours ago
#5962
CodSpeed Performance Gauge
0%
fix: FSST canonicalize when > 2GB of buffers
last run
5 hours ago
#5961
CodSpeed Performance Gauge
0%
© 2026 CodSpeed Technology
Home
Terms
Privacy
Docs