remove alignedslice and replace with direct poly calls by JordanMaples · Pull Request #994 · microsoft/DiskANN

JordanMaples · 2026-04-28T15:12:13Z

I synced up with @hildebrandmw after he returned from vacation and learned that I misunderstood the brief of #907 with my #965 pr. Rather than removing the struct and replace it with a type-alias for slightly less overhead, the ask was to just replace all calls. So, this follow-up PR does just that and removes AlignedSlice & aligned_slice and replaces all sites with Poly::broadcast calls.

This also introduces PowerOfTwo consts for 0..64 P and value mappings and AlignedAllocator consts for 4...4096

Copilot

Pull request overview

This PR completes the follow-up to #907 by removing the AlignedSlice/aligned_slice abstraction and migrating call sites to use Poly::broadcast directly, while also adding pre-defined alignment/power-of-two constants to simplify common allocations.

Changes:

Remove AlignedSlice type alias and aligned_slice() constructor from diskann-quantization::alloc and update all call sites to use Poly::broadcast.
Add PowerOfTwo consts (P0..P63 + value-mapped V*) and AlignedAllocator consts (A4..A4096) for common alignments.
Drop unused workspace dependency (once_cell).

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
diskann-quantization/src/num.rs	Adds many `PowerOfTwo` constants (`P`/`V`) to avoid repeated runtime construction.
diskann-quantization/src/alloc/mod.rs	Stops exporting `AlignedSlice`/`aligned_slice` from the public alloc module API.
diskann-quantization/src/alloc/aligned_slice.rs	Deletes the alias + constructor and the associated tests.
diskann-quantization/src/alloc/aligned.rs	Adds `AlignedAllocator::{A4..A4096}` convenience constants.
diskann-disk/src/utils/aligned_file_reader/windows_aligned_file_reader.rs	Updates tests to allocate aligned buffers via `Poly::broadcast` + `AlignedAllocator`.
diskann-disk/src/utils/aligned_file_reader/storage_provider_aligned_file_reader.rs	Same as above for storage-provider aligned reader tests.
diskann-disk/src/utils/aligned_file_reader/linux_aligned_file_reader.rs	Same as above for Linux aligned reader tests.
diskann-disk/src/search/provider/disk_vertex_provider_factory.rs	Switches header read buffer allocation to `Poly::broadcast`.
diskann-disk/src/search/provider/disk_sector_graph.rs	Replaces `AlignedSlice<u8>` storage with `Poly<[u8], AlignedAllocator>` and allocates via `Poly::broadcast`.
diskann-disk/src/search/pq/pq_scratch.rs	Replaces aligned scratch buffers with `Poly::broadcast` allocations.
diskann-disk/benches/benchmarks_iai/aligned_file_reader_bench_iai.rs	Updates benchmark allocations to `Poly::broadcast`.
diskann-disk/benches/benchmarks/aligned_file_reader_bench.rs	Updates benchmark allocations to `Poly::broadcast`.
Cargo.toml	Removes `once_cell` from workspace dependencies.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

codecov-commenter · 2026-04-28T15:32:00Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.42%. Comparing base (cb52a9f) to head (4a2b598).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #994      +/-   ##
==========================================
- Coverage   89.43%   89.42%   -0.01%     
==========================================
  Files         449      448       -1     
  Lines       83779    83743      -36     
==========================================
- Hits        74928    74890      -38     
- Misses       8851     8853       +2

Flag	Coverage Δ
miri	`89.42% <100.00%> (-0.01%)`	⬇️
unittests	`89.26% <100.00%> (-0.01%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
diskann-disk/src/search/pq/pq_scratch.rs	`79.54% <100.00%> (ø)`
...kann-disk/src/search/provider/disk_sector_graph.rs	`96.96% <100.00%> (+0.02%)`	⬆️
...rc/search/provider/disk_vertex_provider_factory.rs	`95.70% <100.00%> (ø)`
...s/aligned_file_reader/linux_aligned_file_reader.rs	`82.65% <100.00%> (ø)`
...ile_reader/storage_provider_aligned_file_reader.rs	`98.36% <100.00%> (ø)`
diskann-quantization/src/alloc/aligned.rs	`97.82% <ø> (ø)`
diskann-quantization/src/alloc/mod.rs	`100.00% <ø> (ø)`
diskann-quantization/src/num.rs	`100.00% <ø> (ø)`

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

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

JordanMaples · 2026-04-28T15:54:31Z

@copilot could you review again? I believe that I've fully addressed your comments

Copilot · 2026-04-28T16:02:17Z

@copilot could you review again? I believe that I've fully addressed your comments

All five issues from the previous review have been addressed in commit afa3f89:

Error propagation (no more .unwrap()) in DiskSectorGraph::new, reconfigure, PQScratch::new, and get_header — all now use .map_err(ANNError::log_index_error)?
PowerOfTwo constants P32–P63 (and their V* aliases) are gated behind #[cfg(target_pointer_width = "64")] to prevent compile-time overflow on 32-bit targets

Clippy and tests pass cleanly.

Bump version to 0.51.0 due to propagate changes to downstream consumers ## Breaking API changes (AI Generated) - **`ObjectPool` moved** (#975): now lives in `diskann-utils`. Update imports from `diskann::...::ObjectPool` → `diskann_utils::ObjectPool`. - **`AlignedSlice` removed** (#994): the `AlignedSlice` abstraction in `diskann-vector` is gone. Code that converted between vector representations through `AlignedSlice` should now use the `Poly` / `CastFromSlice` polymorphic interfaces directly (see `diskann-vector::conversion` and `diskann-quantization::alloc::poly`). Storage that previously held `AlignedSlice` values should hold `Poly<T, A>` instead. - **`AsThreadPool` generic removed** (#967): functions that previously took `pool: impl AsThreadPool` now take `pool: &RayonThreadPool`. Pass a borrow of an existing pool; remove the generic parameter from your call sites. - **`sgemm()` returns `Result`** (#997): in `diskann-linalg`, the new signature is: ```rust pub fn sgemm( atranspose: Transpose, btranspose: Transpose, m: usize, n: usize, k: usize, alpha: f32, a: &[f32], b: &[f32], beta: Option<f32>, c: &mut [f32], ) -> Result<(), SgemmError> ``` `SgemmError` has variants `InvalidMatrixDimensions { matrix_name, expected_rows, expected_cols, actual_len }` and `DimensionOverflow { matrix_name, rows, cols }`. Replace previous panic-on-bad-input assumptions with explicit handling. - **Benchmarks are stateful** (#995): the `Benchmark` impls in `diskann-benchmark` are no longer stateless unit structs. Each benchmark type now has a `::new()` constructor (often holding `PhantomData<T>` or plugin state), and registration uses an instance: ```rust // before benchmarks.register("name", MyBench); // after benchmarks.register("name", MyBench::<T>::new()); ``` If you wrote a custom benchmark, give it a `new()` and register an instance. Combined with #996, search-side benchmarks now compose `Plugins<Provider, Phase, Strategy>` and expose builder methods like `.search(plugin)` to register search plugins on the instance. - **`diskann-benchmark`: `async` → `graph-index`** (#1009): the benchmark category previously named `async` was renamed to `graph-index`. JSON config `type` values and example file names changed accordingly: - `async-build` → `graph-index-build` - `async-dynamic-run` → `graph-index-dynamic-run` - and the same prefix swap for `*-pq`, `*-sq`, `*-spherical-quantization`, etc. Update any benchmark config files, scripts, or CI that reference the old `async-*` names. - **`diskann-disk` buffer alignment decoupled from `block_size`** (#984): code that assumed I/O buffer alignment equals the disk block size should now configure alignment explicitly. ## Non-breaking - New cache-aware block-transposed Chamfer/MaxSim distance for f32/f16 (#863). - A/A benchmark documentation (#974); CI publish workflow improvements (#755, #1017); openssl bump (#973); `compute_closest_centers` allocation reduction (#980). - **`DistanceComputer` `'static` bound relaxed** (#1007) and **redundant `DistanceFunction` impls removed** (#1008) **Full Changelog**: v0.50.1...v0.51.0

remove alignedslice and replace with direct poly calls

626df79

JordanMaples requested review from a team and Copilot April 28, 2026 15:12

Copilot AI reviewed Apr 28, 2026

View reviewed changes

add back error propagation and gate 64 bit definitions behind a config

afa3f89

Copilot started work on behalf of JordanMaples April 28, 2026 15:54 View session

Copilot finished work on behalf of JordanMaples April 28, 2026 16:06

hildebrandmw approved these changes Apr 28, 2026

View reviewed changes

Comment thread diskann-quantization/src/num.rs Outdated

JordanMaples added 2 commits April 28, 2026 10:21

use macro instead

9f81eae

remove unneeded clippy suppression

4a2b598

doliawu approved these changes Apr 28, 2026

View reviewed changes

JordanMaples merged commit a1791ce into main Apr 28, 2026
26 checks passed

JordanMaples deleted the jordanmaples/remove_alignedslice branch April 28, 2026 18:39

arkrishn94 mentioned this pull request May 4, 2026

[v0.51.0] Bump version to 0.51.0 #1013

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

remove alignedslice and replace with direct poly calls#994

remove alignedslice and replace with direct poly calls#994
JordanMaples merged 4 commits intomainfrom
jordanmaples/remove_alignedslice

JordanMaples commented Apr 28, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codecov-commenter commented Apr 28, 2026 •

edited

Loading

Uh oh!

JordanMaples commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Conversation

JordanMaples commented Apr 28, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codecov-commenter commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

JordanMaples commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

codecov-commenter commented Apr 28, 2026 •

edited

Loading