[Mark Hildebrand (hildebrandmw)]

DiskANN v0.61.0

Highlights

• Quantization: Added 3-bit spherical quantization with AVX2 acceleration, faster
  spherical encoding, and NEON float32 MinMax kernels.
• In-memory indexing (experimental): Added product quantization and 1-8-bit scalar
  quantization, unified quantized representations, and AArch64 prefetching.
• Multi-vector search: Added i8 x i8 and MinMax8 x MinMax4 MaxSim kernels.
• Fixes: Preserve virtual start-point out-edges when writing disk indexes, correctly
  round-trip scalar quantization configurations without standard deviation, and support
  zero-column mutable matrix iterators.

Breaking changes and upgrading

Changes are concentrated in quantization APIs and experimental/tooling crates.

• PQ training ([diskann-quantization] Teach PQ training to return a BasicTable. #1467) - Training now returns BasicTable. SimplePivots was removed.
  Replace flatten() with view_pivots().as_slice() and use pivots_for(chunk) for
  per-chunk access.
• PQ distance tables ([diskann-quantization] PQ Infrastructure for diskann-inmem #1458) - Change ProcessInto bounds to ProcessInto<f32> and calls
  from process_into::<Metric>(...) to process_into::<Metric, f32>(...).
• MaxSim kernels (i8 x i8 MaxSim Kernels #1429) - Generic kernel/erasure code now requires T: MaxSimElement and
  score buffers of T::Score (f32 for floats, i32 for i8). Update build_max_sim
  error handling from NotSupported to BuildMaxSimError.
• diskann-inmem ([diskann-inmem] Spherical Quantization Round 2 #1485) - Replace repr::Spherical with repr::Quantized. Wrap the
  quantizer in repr::quantized::spherical::Spherical::new(...) when configuring it.
  Configuration and Rerank types now live under repr::quantized.
• diskann-benchmark-runner ([benchmark-runner] Simplify Input representation #1475) - input::Registered and public Registry::input were
  removed. Use the concrete Input implementation's tag() and example() methods instead.

All Changes

• Fix serialization round trip for QuantizationType::SQ without standard deviation by Alex Razumov (@arrayka) in Fix serialization round trip for QuantizationType::SQ without standard deviation #1461
• i8 x i8 MaxSim Kernels by Suryansh Gupta (@suri-kumkaran) in i8 x i8 MaxSim Kernels #1429
• Adding subview_mut to MatrixMut by Aditya Krishnan (@arkrishn94) in Adding subview_mut to MatrixMut  #1463
• Add InsertResult::SuccessUpdate by Jack Moffitt (@metajack) in Add InsertResult::SuccessUpdate #1449
• [diskann-quantization] PQ Infrastructure for diskann-inmem by Mark Hildebrand (@hildebrandmw) in [diskann-quantization] PQ Infrastructure for diskann-inmem #1458
• [diskann-quantization] Teach PQ training to return a BasicTable. by Mark Hildebrand (@hildebrandmw) in [diskann-quantization] Teach PQ training to return a BasicTable. #1467
• [deps] Run cargo-update by Mark Hildebrand (@hildebrandmw) in [deps] Run cargo-update #1452
• Support zero-column mutable matrix iterators by Alex Razumov (@arrayka) in Support zero-column mutable matrix iterators #1462
• [diskann-quantization] Use f32::NAN instead of core::f32::NAN. by Mark Hildebrand (@hildebrandmw) in [diskann-quantization] Use f32::NAN instead of core::f32::NAN. #1471
• [diskann-wide] Add V4::new_checked_miri_uncached. by Mark Hildebrand (@hildebrandmw) in [diskann-wide] Add V4::new_checked_miri_uncached. #1480
• [diskann-wide] Add maddubs intrinsic wrappers to V3 and V4 by juchen-ms (@partychen) in [diskann-wide] Add maddubs intrinsic wrappers to V3 and V4 #1479
• Bump the github-actions group across 1 directory with 5 updates by Dependabot (@dependabot)[bot] in Bump the github-actions group across 1 directory with 5 updates #1482
• [benchmark-runner] Simplify Input representation by Mark Hildebrand (@hildebrandmw) in [benchmark-runner] Simplify Input representation #1475
• [diskann-quantization] Optimize spherical encoding scale search by xinyuwen2 in [diskann-quantization] Optimize spherical encoding scale search #1478
• [diskann-garnet] Implement XVCREATE, XVIMPORT by Jack Moffitt (@metajack) in [diskann-garnet] Implement XVCREATE, XVIMPORT #1455
• [diskann-garnet] Support configurable start points and allow importing start points by Jack Moffitt (@metajack) in [diskann-garnet] Support configurable start points and allow importing start points #1464
• [diskann-garnet] Add spherical 2-bit and 4-bit quantization by Jack Moffitt (@metajack) in [diskann-garnet] Add spherical 2-bit and 4-bit quantization #1465
• [deps] Address depandabot issue 32 by Mark Hildebrand (@hildebrandmw) in [deps] Address depandabot issue 32 #1496
• [diskann-disk] Preserve virtual start point out-edges when writing the disk index by Dongliang Wu (@doliawu) in [diskann-disk] Preserve virtual start point out-edges when writing the disk index #1483
• [diskann-quantization] Add Neon F32 Minmax Kernels by pfoxARM in [diskann-quantization] Add Neon F32 Minmax Kernels #1431
• [diskann-garnet] Add macOS arm64 support to NuGet publishing by Jack Moffitt (@metajack) in [diskann-garnet] Add macOS arm64 support to NuGet publishing #1495
• [diskann-disk] Require start points at index initialization. by Mark Hildebrand (@hildebrandmw) in [diskann-disk] Require start points at index initialization. #1497
• Expose metric-aware quantizer centroid distance by Namrata Mohanty (@cherryrhea) in Expose metric-aware quantizer centroid distance #1395
• [diskann-inmem] Add support for Product Quantization by Mark Hildebrand (@hildebrandmw) in [diskann-inmem] Add support for Product Quantization #1469
• Adds aarch64 prefetching into diskann-vector and diskann-inmem by pfoxARM in Adds aarch64 prefetching into diskann-vector and diskann-inmem #1408
• [diskann-inmem] Add scalar quantization by Mark Hildebrand (@hildebrandmw) in [diskann-inmem] Add scalar quantization #1484
• Bump the github-actions group with 3 updates by Dependabot (@dependabot)[bot] in Bump the github-actions group with 3 updates #1505
• [diskann-inmem] Spherical Quantization Round 2 by Mark Hildebrand (@hildebrandmw) in [diskann-inmem] Spherical Quantization Round 2 #1485
• [diskann-quantization] Add SphericalQ 3-bit support and AVX2 kernel by xinyuwen2 in [diskann-quantization] Add SphericalQ 3-bit support and AVX2 kernel #1474
• [muilt-vetor] Add standalone MinMax8 x MinMax4 MaxSim kernel by juchen-ms (@partychen) in [muilt-vetor] Add standalone MinMax8 x MinMax4 MaxSim kernel #1494

New Contributors

• Namrata Mohanty (@cherryrhea) made their first contribution in Expose metric-aware quantizer centroid distance #1395

Full Changelog: v0.60.0...v0.61.0

[Copilot]

🟢 Approval recommended

The version-only changes are consistent across the workspace manifest and lockfile, with no blocking issues identified.

No open issues 🎉

What changed in this PR

Prepares the DiskANN workspace for v0.61.0 by updating release metadata. This PR changes versions only; the features described in the release notes were implemented separately.

Changes:

• Bumps the shared workspace version and internal dependency requirements from 0.60.0 to 0.61.0.
• Updates all 17 DiskANN package entries in the lockfile to match.

File	Description
Cargo.toml	Sets the workspace and internal dependency versions to 0.61.0.
Cargo.lock	Aligns DiskANN package versions with the release.

_{Review effort: Balanced}

[Codecov Comments Bot (codecov-commenter)]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.12%. Comparing base (8a9ff2d) to head (d96bb89).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1511      +/-   ##
==========================================
- Coverage   92.16%   91.12%   -1.05%     
==========================================
  Files         592      592              
  Lines      123582   123582              
==========================================
- Hits       113901   112610    -1291     
- Misses       9681    10972    +1291     

Flag	Coverage Δ
miri	91.12% <ø> (-1.05%)	⬇️
unittests	90.82% <ø> (-1.29%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.
see 53 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.