Add native SIMD testing and benchmark module #708
+1,076
−57
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,118 @@ | ||
| /* | ||
| * Copyright DataStax, Inc. | ||
| * | ||
| * Licensed under the Apache License, Version 2.0 (the "License"); | ||
| * you may not use this file except in compliance with the License. | ||
| * You may obtain a copy of the License at | ||
| * | ||
| * http://www.apache.org/licenses/LICENSE-2.0 | ||
| * | ||
| * Unless required by applicable law or agreed to in writing, software | ||
| * distributed under the License is distributed on an "AS IS" BASIS, | ||
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| * See the License for the specific language governing permissions and | ||
| * limitations under the License. | ||
| */ | ||
|
|
||
| // Google Benchmark micro-benchmarks for the fp32 vector similarity kernels: | ||
| // cosine_f32, dot_product_f32, euclidean_f32 | ||
| // | ||
| // Parameterised over the realistic embedding dimensions used in production: | ||
| // 128, 256, 512, 1024, 1536, 3072 | ||
| // | ||
| // Build (requires google-benchmark installed or available via pkg-config): | ||
| // meson setup build && ninja -C build bench_simd_kernels | ||
|
Contributor
There was a problem hiding this comment. This doesn't work for me. I cloned the fork but this command fails with the following
Contributor
There was a problem hiding this comment. Oh it's |
||
| // | ||
| // Run: | ||
| // ./build/bench_simd_kernels [--benchmark_filter=<pattern>] | ||
|
|
||
| #include <benchmark/benchmark.h> | ||
| #include <cmath> | ||
| #include <vector> | ||
|
|
||
| #include "jvector_simd.h" | ||
|
|
||
| // --------------------------------------------------------------------------- | ||
| // Helpers | ||
| // --------------------------------------------------------------------------- | ||
|
|
||
| // Deterministic, non-zero float vector: avoids degenerate cosine=NaN cases. | ||
| static std::vector<float> make_vec(size_t n, float seed) | ||
| { | ||
| std::vector<float> v(n); | ||
| for (size_t i = 0; i < n; ++i) { | ||
| v[i] = seed * (1.0f + static_cast<float>(i % 7) * 0.13f); | ||
| if (i % 3 == 0) v[i] = -v[i]; | ||
| v[i] += 0.5f; | ||
| } | ||
| return v; | ||
| } | ||
|
|
||
| // Benchmark sizes matching production embedding dimensions. | ||
| static const std::vector<int64_t> kBenchSizes = {128, 256, 512, 1024, 1536, 3072}; | ||
|
|
||
| // --------------------------------------------------------------------------- | ||
| // dot_product_f32 | ||
| // --------------------------------------------------------------------------- | ||
|
|
||
| static void BM_dot_product_f32(benchmark::State& state) | ||
| { | ||
| const size_t n = static_cast<size_t>(state.range(0)); | ||
| auto a = make_vec(n, 0.7f); | ||
| auto b = make_vec(n, 1.3f); | ||
|
|
||
| for (auto _ : state) { | ||
| float result = dot_product_f32(a.data(), 0, b.data(), 0, n); | ||
| benchmark::DoNotOptimize(result); | ||
| } | ||
|
|
||
| state.SetItemsProcessed(state.iterations() * static_cast<int64_t>(n)); | ||
| state.SetBytesProcessed(state.iterations() * static_cast<int64_t>(n) * 2 * sizeof(float)); | ||
| } | ||
| BENCHMARK(BM_dot_product_f32)->ArgsProduct({kBenchSizes}); | ||
|
|
||
| // --------------------------------------------------------------------------- | ||
| // euclidean_f32 | ||
| // --------------------------------------------------------------------------- | ||
|
|
||
| static void BM_euclidean_f32(benchmark::State& state) | ||
| { | ||
| const size_t n = static_cast<size_t>(state.range(0)); | ||
| auto a = make_vec(n, 0.7f); | ||
| auto b = make_vec(n, 1.3f); | ||
|
|
||
| for (auto _ : state) { | ||
| float result = euclidean_f32(a.data(), 0, b.data(), 0, n); | ||
| benchmark::DoNotOptimize(result); | ||
| } | ||
|
|
||
| state.SetItemsProcessed(state.iterations() * static_cast<int64_t>(n)); | ||
| state.SetBytesProcessed(state.iterations() * static_cast<int64_t>(n) * 2 * sizeof(float)); | ||
| } | ||
| BENCHMARK(BM_euclidean_f32)->ArgsProduct({kBenchSizes}); | ||
|
|
||
| // --------------------------------------------------------------------------- | ||
| // cosine_f32 | ||
| // --------------------------------------------------------------------------- | ||
|
|
||
| static void BM_cosine_f32(benchmark::State& state) | ||
| { | ||
| const size_t n = static_cast<size_t>(state.range(0)); | ||
| auto a = make_vec(n, 0.7f); | ||
| auto b = make_vec(n, 1.3f); | ||
|
|
||
| for (auto _ : state) { | ||
| float result = cosine_f32(a.data(), 0, b.data(), 0, n); | ||
| benchmark::DoNotOptimize(result); | ||
| } | ||
|
|
||
| state.SetItemsProcessed(state.iterations() * static_cast<int64_t>(n)); | ||
| state.SetBytesProcessed(state.iterations() * static_cast<int64_t>(n) * 2 * sizeof(float)); | ||
| } | ||
| BENCHMARK(BM_cosine_f32)->ArgsProduct({kBenchSizes}); | ||
|
|
||
| // --------------------------------------------------------------------------- | ||
| // Entry point — benchmark::Initialize parses --benchmark_* flags. | ||
| // --------------------------------------------------------------------------- | ||
|
|
||
| BENCHMARK_MAIN(); | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
File renamed without changes.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
File renamed without changes.
File renamed without changes.
File renamed without changes.
File renamed without changes.
File renamed without changes.
File renamed without changes.
File renamed without changes.
File renamed without changes.
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this looks like it is probably fine, but for future reference it would be easier to verify if this was a branch and not coming from a fork. Since it's a fork I can't just run the gha and verify the results, which would make it easier to be comfortable with approval. Not a showstopper but something to keep in mind for future PRs.