perf: zero copy validity export to duckdb#7371
Open
joseph-isaacs wants to merge 16 commits intodevelopfrom
Open
perf: zero copy validity export to duckdb#7371joseph-isaacs wants to merge 16 commits intodevelopfrom
joseph-isaacs wants to merge 16 commits intodevelopfrom
Conversation
joseph-isaacs
added
the
action/benchmark
github-actions
Bot
removed
the
action/benchmark
Contributor
Polar Signals Profiling ResultsLatest Run
Previous Runs (4)
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals ProfilingVortex (geomean): 0.993x ➖ datafusion / vortex-file-compressed (0.993x ➖, 3↑ 3↓)
|
Contributor
File Sizes: PolarSignals ProfilingNo file size changes detected. |
Contributor
Benchmarks: FineWeb NVMeVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.951x ➖, 1↑ 0↓)
datafusion / vortex-compact (1.199x ❌, 1↑ 7↓)
datafusion / parquet (0.991x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.888x ✅, 4↑ 0↓)
duckdb / vortex-compact (0.911x ➖, 2↑ 0↓)
duckdb / parquet (0.938x ➖, 2↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: FineWeb NVMeNo file size changes detected. |
Contributor
Benchmarks: TPC-H SF=1 on NVMEVerdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.923x ➖, 7↑ 1↓)
datafusion / vortex-compact (0.980x ➖, 0↑ 0↓)
datafusion / parquet (1.025x ➖, 0↑ 2↓)
datafusion / arrow (0.973x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (1.018x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.036x ➖, 0↑ 1↓)
duckdb / parquet (1.004x ➖, 1↑ 1↓)
duckdb / duckdb (1.011x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-H SF=1 on NVMENo file size changes detected. |
Contributor
Benchmarks: TPC-DS SF=1 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.942x ➖, 22↑ 0↓)
datafusion / vortex-compact (1.076x ➖, 3↑ 42↓)
datafusion / parquet (1.001x ➖, 1↑ 1↓)
duckdb / vortex-file-compressed (1.014x ➖, 0↑ 4↓)
duckdb / vortex-compact (1.005x ➖, 0↑ 2↓)
duckdb / parquet (1.006x ➖, 0↑ 1↓)
duckdb / duckdb (1.002x ➖, 2↑ 4↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-DS SF=1 on NVMENo file size changes detected. |
Contributor
Benchmarks: TPC-H SF=1 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.983x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.212x ➖, 0↑ 5↓)
datafusion / parquet (0.975x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.949x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.005x ➖, 0↑ 0↓)
duckdb / parquet (1.027x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: TPC-H SF=10 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.965x ➖, 1↑ 0↓)
datafusion / vortex-compact (0.987x ➖, 0↑ 0↓)
datafusion / parquet (0.994x ➖, 0↑ 0↓)
datafusion / arrow (0.981x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.994x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.996x ➖, 0↑ 0↓)
duckdb / parquet (1.003x ➖, 0↑ 0↓)
duckdb / duckdb (0.997x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: TPC-H SF=10 on NVMENo file size changes detected. |
Contributor
Benchmarks: FineWeb S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.919x ➖, 1↑ 0↓)
datafusion / vortex-compact (1.115x ➖, 0↑ 3↓)
datafusion / parquet (0.940x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.982x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.048x ➖, 0↑ 0↓)
duckdb / parquet (1.010x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: Statistical and Population GeneticsVerdict: No clear signal (low confidence) duckdb / vortex-file-compressed (1.037x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.000x ➖, 0↑ 0↓)
duckdb / parquet (1.000x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: Statistical and Population GeneticsNo file size changes detected. |
Contributor
Benchmarks: TPC-H SF=10 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.880x ➖, 1↑ 1↓)
datafusion / vortex-compact (1.052x ➖, 0↑ 0↓)
datafusion / parquet (0.912x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (1.019x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.037x ➖, 0↑ 0↓)
duckdb / parquet (0.994x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: Clickbench on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.868x ✅, 14↑ 1↓)
datafusion / parquet (0.995x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.958x ➖, 3↑ 0↓)
duckdb / parquet (0.999x ➖, 0↑ 0↓)
duckdb / duckdb (0.955x ➖, 5↑ 0↓)
Full attributed analysis
|
Contributor
File Sizes: Clickbench on NVMEFile Size Changes (3 files changed, -0.0% overall, 0↑ 3↓)
Totals:
|
Contributor
Benchmarks: CompressionVortex (geomean): 1.009x ➖ unknown / unknown (0.993x ➖, 1↑ 2↓)
|
Contributor
Benchmarks: Random AccessVortex (geomean): 1.295x ❌ unknown / unknown (1.173x ❌, 2↑ 47↓)
|
joseph-isaacs
added
the
action/benchmark
github-actions
Bot
removed
the
action/benchmark
myrrc
reviewed
Apr 10, 2026
|
|
||
| // Set the validity pointer for the vector to external data, and store the buffer in auxiliary | ||
| // to keep it alive. This enables zero-copy export of validity masks. | ||
| void duckdb_vx_vector_set_validity_data(duckdb_vector ffi_vector, void *validity_ptr, idx_t capacity, |
Contributor
There was a problem hiding this comment.
If validity_ptr points to buffer, just pass the buffer
Contributor
Author
There was a problem hiding this comment.
validity_ptr is not the buffer is something a few levels of ptr deep. We could fix, but would also want do change Primitive/Decimal Export at once
myrrc
reviewed
Apr 10, 2026
| // Same hack for ValidityMask: access protected fields via inheritance. | ||
| class ExternalValidityMask : public ValidityMask { | ||
| public: | ||
| inline void SetExternal(validity_t *ptr, idx_t cap, |
Contributor
There was a problem hiding this comment.
Same here, pass just the buffer and derive ptr from it
myrrc
reviewed
Apr 10, 2026
| *ext_buf, reinterpret_cast<TemplatedValidityData<validity_t> *>(ext_buf->get())); | ||
|
|
||
| // Set validity_mask, capacity, and validity_data (which keeps the buffer alive). | ||
| ext_validity->SetExternal(reinterpret_cast<validity_t *>(validity_ptr), capacity, |
Contributor
There was a problem hiding this comment.
Technically this will slice the class to base's validity, but as derived class doesn't have any members, it's fine. Worth adding a comment
myrrc
reviewed
Apr 10, 2026
…ty-export Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> # Conflicts: # vortex-duckdb/src/exporter/constant.rs # vortex-duckdb/src/exporter/vector.rs
joseph-isaacs
added
action/benchmark
github-actions
Bot
removed
the
action/benchmark
joseph-isaacs
added
the
action/benchmark
github-actions
Bot
removed
the
action/benchmark
myrrc
reviewed
Apr 21, 2026
joseph-isaacs
added
action/benchmark
github-actions
Bot
removed
the
action/benchmark
joseph-isaacs
changed the title
…ty-export Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Zero copy export validity similarly to how we export data for Primitive or Decimal.