Repository navigation

Release 0.6.1 with non-search packed performance improvements - #22

Merged

jcfangc merged 51 commits into

mainfrom

integration/pre-search

Oct 20, 2026

Merged

Release 0.6.1 with non-search packed performance improvements#22
jcfangc merged 51 commits into
mainfrom
integration/pre-search

Commits on Sep 14, 2026
- Optimize packed code extraction
  jcfangc
  committed
  7af2ede
Commits on Sep 16, 2026
- Optimize packed view access
  jcfangc
  committed
  bbf04d9
- Share packed code extraction
  jcfangc
  committed
  151e8a5
- Optimize packed construction
  jcfangc
  committed
  215e980
Commits on Sep 19, 2026
- Adopt owned bit-string words
  jcfangc
  committed
  2ec5fcf
- Add packed word block backend
  jcfangc
  committed
  09a7759
- Specialize scalar packing for eight-bit codes
  jcfangc
  committed
  b154885
- Specialize scalar packing for four-bit codes
  jcfangc
  committed
  7c45c3c
- Specialize scalar packing for two-bit codes
  jcfangc
  committed
  b5d4324
- Optimize irregular packed widths in blocks
  jcfangc
  committed
  42d82eb
- Specialize scalar packing for one-bit codes
  jcfangc
  committed
  50884f6
- Batch packed construction before SIMD
  jcfangc
  committed
  be42264
- Prepare packed word packing backends
  jcfangc
  committed
  e9cbfb8
- Finalize packed word packing contract
  jcfangc
  committed
  eb3b172
Commits on Sep 20, 2026
- Add AVX2 nibble packing backend
  jcfangc
  committed
  9476d3a
- Add AVX2 bit-mask packing backend
  jcfangc
  committed
  3d6db1f
- Document AVX2 packing support matrix
  jcfangc
  committed
  028c8c0
- Add AVX2 six-bit packing backend
  jcfangc
  committed
  da342d3
- Add AVX2 three-bit packing backend
  jcfangc
  committed
  58b63d7
- Add AVX2 five-bit packing backend
  jcfangc
  committed
  6e08105
- Add AVX2 seven-bit packing backend
  jcfangc
  committed
  a10ad6f
- Freeze AVX2 packing thresholds
  jcfangc
  committed
  cc36fbf
- Organize AVX2 packing kernels
  jcfangc
  committed
  1d65434
Commits on Sep 26, 2026
- Optimize sequential packed iteration
  jcfangc
  committed
  854cb1a
Commits on Sep 29, 2026
- Add packed benchmarks for widths five and six
  jcfangc
  committed
  e8f7d60
- Optimize packed iteration with a rolling window
  jcfangc
  committed
  5aaa220
Commits on Sep 30, 2026
- Add AVX2 packed bulk unpack for conversion
  jcfangc
  committed
  ae9d107
- Avoid slow random access for short packed conversion
  jcfangc
  committed
  8ab8113
Commits on Oct 1, 2026
- Unpack complete layout blocks in conversion tail
  jcfangc
  committed
  bfd3f12
- Clarify conversion layout length name
  jcfangc
  committed
  3803ddb
- Expand AVX2 unpack width support
  jcfangc
  committed
  d4f689d
- Strengthen AVX2 unpack regression coverage
  jcfangc
  committed
  6d3e690
- Use bulk unpack only with accelerated backend
  jcfangc
  committed
  5ce7012
Commits on Oct 7, 2026
- Clarify bulk unpack capability naming
  jcfangc
  committed
  8ab293f
- Use unchecked word refill in packed iterator
  jcfangc
  committed
  b31ec03
- Use unchecked word loads for packed get
  jcfangc
  committed
  b9f3b5f
- Document packed view alignment invariants
  jcfangc
  committed
  eeef026
- Optimize packed view iteration with rolling cursor
  jcfangc
  committed
  621718c
- Cover aligned packed view iterator starts
  jcfangc
  committed
  d9296dd
- Use packed bit append for character push
  jcfangc
  committed
  70c5de7
Commits on Oct 11, 2026
- Specialize byte-width packed push
  jcfangc
  committed
  51df6fa
Commits on Oct 14, 2026
- Optimize packed insertion and removal with fixed-width shifts
  jcfangc
  committed
  c6cdbf0
Commits on Oct 15, 2026
- Optimize in-place word shifts with a carry stream
  jcfangc
  committed
  390d58f
Commits on Oct 16, 2026
- Optimize AVX2 in-place right shifts with word lookahead
  jcfangc
  committed
  3cf8fa6
Commits on Oct 17, 2026
- Document SSE2 carry prototype codegen and negative results
  jcfangc
  committed
  30591e9
- Experiment with SSE2 vector lookahead for in-place right shifts
  jcfangc
  committed
  09ac225
- Reuse packed character push in Extend
  jcfangc
  committed
  2dafe6a
- Batch packed codes when extending PackedString
  jcfangc
  committed
  431179e
- Document and benchmark batched packed extension
  jcfangc
  committed
  6e3c688
Commits on Oct 20, 2026
- Bump workspace version to 0.6.1
  jcfangc
  committed
  dc8e55a
- Distinguish packing and unpacking batch constants
  jcfangc
  committed
  a180923

Commits

Commits on Sep 14, 2026

Commits on Sep 16, 2026

Commits on Sep 19, 2026

Commits on Sep 20, 2026

Commits on Sep 26, 2026

Commits on Sep 29, 2026

Commits on Sep 30, 2026

Commits on Oct 1, 2026

Commits on Oct 7, 2026

Commits on Oct 11, 2026

Commits on Oct 14, 2026

Commits on Oct 15, 2026

Commits on Oct 16, 2026

Commits on Oct 17, 2026

Commits on Oct 20, 2026