Bump toolchain to 1.97.1

Cargo.toml

@@ -87,7 +87,7 @@ dashmap = "6.1.0"
 futures-util = { version = "0.3.31", default-features = false }
 half = "2.6.0"
 hashbrown = "0.16.0"
-iai-callgrind = "0.14.0"
+iai-callgrind = "0.16.1"
 itertools = "0.13.0"
 num-traits = "0.2.15"
 num_cpus = "1.16.0"

diskann-benchmark/src/inputs/graph_index.rs

@@ -617,7 +617,7 @@ impl IndexLoad {
             save_and_load::get_graph_max_observed_degree(&storage_provider, &self.load_path)?;
 
         let metadata =
-            load_metadata_from_file(&storage_provider, &format!("{}.data", &self.load_path))?;
+            load_metadata_from_file(&storage_provider, &format!("{}.data", self.load_path))?;
 
         let distance: diskann_vector::distance::Metric = self.distance.into();
         let config = config::Builder::new(

diskann-disk/benches/bench_main_iai.rs

@@ -5,16 +5,13 @@
 
 use benchmarks_iai::aligned_file_reader_bench_iai::aligned_file_reader_bench_iai;
 use benchmarks_iai::kmeans_bench_iai::kmeans_bench_iai;
-use iai_callgrind::{main, EventKind, LibraryBenchmarkConfig, RegressionConfig};
+use iai_callgrind::{main, Callgrind, EventKind, LibraryBenchmarkConfig};
 
 mod benchmarks_iai;
 
 main!(
     config = LibraryBenchmarkConfig::default()
-        .regression(
-            RegressionConfig::default()
-                .limits([(EventKind::Ir, 5.0), (EventKind::EstimatedCycles, 5.0)])
-        );
+        .tool(Callgrind::default().soft_limits([(EventKind::Ir, 5.0), (EventKind::EstimatedCycles, 5.0)]));
     library_benchmark_groups =
         aligned_file_reader_bench_iai,
         kmeans_bench_iai,

diskann-label-filter/src/parser/query_parser.rs

@@ -84,10 +84,8 @@ fn parse_query_filter_with_depth(
         if let Some(arr) = obj.get("$and").and_then(|v| v.as_array()) {
             let mut exprs = Vec::new();
             for v in arr {
-                match parse_query_filter_with_depth(v, depth + 1) {
-                    Ok(expr) => exprs.push(expr),
-                    Err(e) => return Err(e),
-                }
+                let expr = parse_query_filter_with_depth(v, depth + 1)?;
+                exprs.push(expr)
             }
             if exprs.is_empty() {
                 return Err(QueryFilterError::ParseFailure(
@@ -100,10 +98,8 @@ fn parse_query_filter_with_depth(
         if let Some(arr) = obj.get("$or").and_then(|v| v.as_array()) {
             let mut exprs = Vec::new();
             for v in arr {
-                match parse_query_filter_with_depth(v, depth + 1) {
-                    Ok(expr) => exprs.push(expr),
-                    Err(e) => return Err(e),
-                }
+                let expr = parse_query_filter_with_depth(v, depth + 1)?;
+                exprs.push(expr)
             }
             if exprs.is_empty() {
                 return Err(QueryFilterError::ParseFailure(

diskann-quantization/src/minmax/quantizer.rs

@@ -504,9 +504,9 @@ mod minmax_quantizer_tests {
         assert!(
             (reconstruction_error / norm) <= relative_err,
             "Expected vector : {:?} to be reconstructed within error {} but instead got : {:?}, with error {} for dim : {}",
-            &vector,
+            vector,
             relative_err,
-            &reconstructed,
+            reconstructed,
             reconstruction_error / norm,
             dim,
         );
@@ -522,7 +522,7 @@ mod minmax_quantizer_tests {
             "Encoded vector with dim : {dim} is {:?}, got error : {} for vector : {:?}",
             encoded.reborrow(),
             (code_sum - expected_code_sum).abs(),
-            &vector,
+            vector,
         );
         let recon_norm_sq = reconstructed.iter().map(|x| x * x).sum::<f32>();
         assert!((encoded.reborrow().meta().norm_squared - recon_norm_sq).abs() <= 1e-3);

diskann-tools/src/utils/compute_bitmap.rs

@@ -79,7 +79,7 @@ impl QueryAccelerator for InvertedIndexAccelerator {
 
     fn universe(&self) -> BitSet {
         let mut result = BitSet::new();
-        for (_, bits) in self.map.iter() {
+        for bits in self.map.values() {
             result.extend(bits);
         }
         result
@@ -121,7 +121,7 @@ impl QueryAccelerator for BTreeAccelerator {
 
     fn universe(&self) -> BitSet {
         let mut result = BitSet::new();
-        for (_, ids) in self.map.iter() {
+        for ids in self.map.values() {
             result.extend(ids.iter().cloned());
         }
         result

diskann-vector/benches/bench_main_vector_iai.rs

@@ -4,16 +4,14 @@
  */
 
 use benchmarks_iai::{contains_bench_iai::*, cosine_iai::*, cosine_normalized_iai::*, l2_iai::*};
-use iai_callgrind::{main, EventKind, LibraryBenchmarkConfig, RegressionConfig};
+use iai_callgrind::{main, Callgrind, EventKind, LibraryBenchmarkConfig};
 mod benchmarks_iai;
 pub(crate) mod utils;
 
 main!(
     config = LibraryBenchmarkConfig::default()
-        .regression(
-            RegressionConfig::default()
-                .limits([(EventKind::Ir, 5.0), (EventKind::EstimatedCycles, 5.0), (EventKind::TotalRW, 5.0), (EventKind::L1hits, 5.0)])
-        );
+        .tool(Callgrind::default().soft_limits([(EventKind::Ir, 5.0), (EventKind::EstimatedCycles, 5.0), (EventKind::TotalRW, 5.0), (EventKind::L1hits, 5.0)]));
+
     library_benchmark_groups =
         benchmark_contains_bench_iai,
         cosine_f16_iai,

diskann-vector/benches/benchmarks_iai/contains_bench_iai.rs

@@ -3,8 +3,8 @@
  * Licensed under the MIT license.
  */
 use diskann_vector::contains::ContainsSimd;
-use iai_callgrind::black_box;
 use rand::{distr::Uniform, prelude::Distribution, rngs::StdRng, Rng, SeedableRng};
+use std::hint::black_box;
 
 iai_callgrind::library_benchmark_group!(
     name = benchmark_contains_bench_iai;

diskann-vector/benches/benchmarks_iai/cosine_iai.rs

@@ -7,9 +7,9 @@ use std::num::NonZeroUsize;
 
 use diskann_vector::{distance::Cosine, DistanceFunction};
 use half::f16;
-use iai_callgrind::black_box;
 use rand::{rngs::StdRng, SeedableRng};
 use rand_distr::{Distribution, Normal, StandardUniform};
+use std::hint::black_box;
 
 use crate::utils::{AlignedDataset, AlignedVector, DatasetArgs, InlineBarrier, Static};
 iai_callgrind::library_benchmark_group!(

diskann-vector/benches/benchmarks_iai/cosine_normalized_iai.rs

@@ -7,9 +7,9 @@ use std::num::NonZeroUsize;
 
 use diskann_vector::distance::CosineNormalized;
 use half::f16;
-use iai_callgrind::black_box;
 use rand::{rngs::StdRng, SeedableRng};
 use rand_distr::{Distribution, Normal};
+use std::hint::black_box;
 
 use crate::utils::{AlignedDataset, AlignedVector, DatasetArgs, InlineBarrier, Static};
 

diskann-wide/src/test_utils/distribution.rs

@@ -103,7 +103,11 @@ macro_rules! finite {
                     // infinity/NaN).
                     //
                     // The mantissa is allowed to be all zeros.
-                    (<$T>::EXPONENT_MASK | <$T>::MANTISSA_MASK, false, true)
+                    (
+                        <$T as Layout>::EXPONENT_MASK | <$T as Layout>::MANTISSA_MASK,
+                        false,
+                        true,
+                    )
                 } else if kind < NORMAL_WEIGHT + SUBNORMAL_WEIGHT {
                     // Generate a subnormal floating point number.
                     //

rust-toolchain.toml

@@ -1,3 +1,3 @@
 [toolchain]
-channel = "1.95"
+channel = "1.97.1"
 components = ["rust-src"]