Files
faiss/perf_tests/bench_scalar_quantizer_distance.cpp
Mulugeta Mammo c000190dc0 Accelerate ScalarQuantizer::QT_bf16 with AVX512-BF16. (#4889)
Summary:
This PR accelerates `ScalarQuantizer::QT_bf16` by vectorizing the BF16 encode/decode path with AVX-512 BF16 and integer instructions, targeting Intel® Sapphire Rapids and newer processors.

Distance computations are also optimized to operate directly on BF16 data, avoiding the overhead of FP32 conversion:
- **Inner product:** The query is encoded to BF16 once, and dot products are computed natively in BF16.
- **L2 distance:** Leverages the norm expansion identity, with all terms evaluated using the native BF16 dot-product instruction.

Scalar quantizer benchmarks demonstrate speedups of **1.42×** for encoding, **1.32×** for decoding, and **1.16×** for distance computation.

 ---

`perf_tests/bench_scalar_quantizer_encode --d=768 --n=2000 --iterations=20 --benchmark_filter="QT_bf16"`

```
Existing impl:
--------------------------------------------------------------------------------
Benchmark                      Time             CPU   Iterations UserCounters...
--------------------------------------------------------------------------------
QT_bf16/iterations:20     689149 ns       689010 ns           20 code_size=1.536k

This PR:
--------------------------------------------------------------------------------
Benchmark                      Time             CPU   Iterations UserCounters...
--------------------------------------------------------------------------------
QT_bf16/iterations:20     485337 ns       484511 ns           20 code_size=1.536k
```

`perf_tests/bench_scalar_quantizer_decode --d=768 --n=2000 --iterations=20 --benchmark_filter="QT_bf16"`

```
Existing impl:
--------------------------------------------------------------------------------
Benchmark                      Time             CPU   Iterations UserCounters...
--------------------------------------------------------------------------------
QT_bf16/iterations:20     498607 ns       498539 ns           20 code_size=1.536k

This PR:
--------------------------------------------------------------------------------
Benchmark                      Time             CPU   Iterations UserCounters...
--------------------------------------------------------------------------------
QT_bf16/iterations:20     376614 ns       374200 ns           20 code_size=1.536k
```

`perf_tests/bench_scalar_quantizer_distance --d=768 --n=2000 --iterations=20 --benchmark_filter="QT_bf16"`

```
Existing impl:
-----------------------------------------------------------------------------------
Benchmark                         Time             CPU   Iterations UserCounters...
-----------------------------------------------------------------------------------
QT_bf16/L2/iterations:20  254155481 ns    253585062 ns           20 code_size=1.536k
QT_bf16/IP/iterations:20  221556582 ns    221156894 ns           20 code_size=1.536k

This PR:
-----------------------------------------------------------------------------------
Benchmark                         Time             CPU   Iterations UserCounters...
-----------------------------------------------------------------------------------
QT_bf16/L2/iterations:20  226742244 ns    226270273 ns           20 code_size=1.536k
QT_bf16/IP/iterations:20  190916826 ns    190591678 ns           20 code_size=1.536k
```

Pull Request resolved: https://github.com/facebookresearch/faiss/pull/4889

Reviewed By: mdouze

Differential Revision: D106149331

Pulled By: mnorris11

fbshipit-source-id: 114097d0597028c43c19727b677dcc494250be41
2026-06-02 13:50:33 -07:00

98 lines
2.5 KiB
C++

/*
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* This source code is licensed under the MIT license found in the
* LICENSE file in the root directory of this source tree.
*/
#include <gflags/gflags.h>
#include <omp.h>
#include <cstdio>
#include <map>
#include <benchmark/benchmark.h>
#include <faiss/MetricType.h>
#include <faiss/impl/ScalarQuantizer.h>
#include <faiss/utils/random.h>
#include "utils.h"
using namespace faiss;
DEFINE_uint32(d, 128, "dimension");
DEFINE_uint32(n, 2000, "number of vectors");
DEFINE_uint32(iterations, 20, "iterations");
static void bench_distance(
benchmark::State& state,
ScalarQuantizer::QuantizerType type,
MetricType metric,
int d,
int n) {
std::vector<float> x(d * n);
float_rand(x.data(), d * n, 12345);
ScalarQuantizer sq(d, type);
omp_set_num_threads(1);
sq.train(n, x.data());
size_t code_size = sq.code_size;
state.counters["code_size"] = sq.code_size;
// encode
std::vector<uint8_t> codes(code_size * n);
sq.compute_codes(x.data(), codes.data(), n);
std::unique_ptr<ScalarQuantizer::SQDistanceComputer> dc(
sq.get_distance_computer(metric));
dc->codes = codes.data();
dc->code_size = sq.code_size;
for (auto _ : state) {
float sum_dis = 0;
for (int i = 0; i < n; i++) {
dc->set_query(&x[i * d]);
for (int j = 0; j < n; j++) {
benchmark::DoNotOptimize(sum_dis += (*dc)(j));
}
}
benchmark::ClobberMemory();
}
}
int main(int argc, char** argv) {
benchmark::Initialize(&argc, argv);
gflags::AllowCommandLineReparsing();
gflags::ParseCommandLineFlags(&argc, &argv, true);
int iterations = FLAGS_iterations;
int d = FLAGS_d;
int n = FLAGS_n;
auto benchs = ::perf_tests::sq_types();
for (auto& [bench_name, quantizer_type] : benchs) {
benchmark::RegisterBenchmark(
(bench_name + "/L2").c_str(),
bench_distance,
quantizer_type,
METRIC_L2,
d,
n)
->Iterations(iterations);
benchmark::RegisterBenchmark(
(bench_name + "/IP").c_str(),
bench_distance,
quantizer_type,
METRIC_INNER_PRODUCT,
d,
n)
->Iterations(iterations);
}
benchmark::RunSpecifiedBenchmarks();
benchmark::Shutdown();
}