This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datafusion.git
The following commit(s) were added to refs/heads/main by this push:
new 37c1b75263 test: scale remaining sort-merge join (SMJ) benchmark
queries (#21200)
37c1b75263 is described below
commit 37c1b75263aad44d17f3c30259c5b81b79e5722f
Author: Matt Butrovich <[email protected]>
AuthorDate: Fri Mar 27 11:36:11 2026 -0400
test: scale remaining sort-merge join (SMJ) benchmark queries (#21200)
## Which issue does this PR close?
<!--
We generally require a GitHub issue to be filed for all bug fixes and
enhancements and this helps us generate change logs for our releases.
You can link an issue to this PR using the GitHub syntax. For example
`Closes #123` indicates that this PR will close issue #123.
-->
- Closes #.
## Rationale for this change
<!--
Why are you proposing this change? If this is already explained clearly
in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand
your changes and offer better suggestions for fixes.
-->
Our SMJ benchmark queries finish too quickly to demonstrate improvements
that aren't massive. For example, I am working on an optimization that
introduces `DynComparator` (part of #20910) and it's about a 10%
improvement, but only when you actually make the queries run long
enough. The new queries for #21184 are scaled enough to see
improvements, but we need to scale the older queries.
I am also continuing to see SMJ issues with Comet when running joins
with billions (sometimes trillions) of rows. We can't do that for
microbenchmarks, but we can at least start hitting millions of rows to
look at more than a handful of batches.
## What changes are included in this PR?
<!--
There is no need to duplicate the description in the issue here but it
is sometimes worth providing a summary of the individual changes in this
PR.
-->
Bring our SMJ queries into alignment with some of the newer ones
(Q21-23) to demonstrate further performance wins.
## Are these changes tested?
<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code
If tests are not included in your PR, please explain why (for example,
are they covered by existing tests)?
-->
I ran the benchmark. On my M3 Max, here's how long it takes:
| Query | Join Type | Rows | Keys | Filter | Median (ms) |
|-------|-----------|------|------|--------|-------------|
| Q1 | INNER | 1M×1M | 1:1 | — | 16.3 |
| Q2 | INNER | 1M×10M | 1:10 | — | 117.4 |
| Q3 | INNER | 1M×1M | 1:100 | — | 74.2 |
| Q4 | INNER | 1M×10M | 1:10 | 1% | 17.1 |
| Q5 | INNER | 1M×1M | 1:100 | 10% | 18.4 |
| Q6 | LEFT | 1M×10M | 1:10 | — | 129.3 |
| Q7 | LEFT | 1M×10M | 1:10 | 50% | 150.2 |
| Q8 | FULL | 1M×1M | 1:10 | — | 16.6 |
| Q9 | FULL | 1M×10M | 1:10 | 10% | 153.5 |
| Q10 | LEFT SEMI | 1M×10M | 1:10 | — | 53.1 |
| Q11 | LEFT SEMI | 1M×10M | 1:10 | 1% | 15.5 |
| Q12 | LEFT SEMI | 1M×10M | 1:10 | 50% | 65.0 |
| Q13 | LEFT SEMI | 1M×10M | 1:10 | 90% | 105.7 |
| Q14 | LEFT ANTI | 1M×10M | 1:10 | — | 54.3 |
| Q15 | LEFT ANTI | 1M×10M | 1:10 | partial | 51.5 |
| Q16 | LEFT ANTI | 1M×1M | 1:1 | — | 10.3 |
| Q17 | INNER | 1M×50M | 1:50 | 5% | 75.9 |
| Q18 | LEFT SEMI | 1M×50M | 1:50 | 2% | 50.2 |
| Q19 | LEFT ANTI | 1M×50M | 1:50 | partial | 336.4 |
| Q20 | INNER | 1M×10M | 1:100 | GROUP BY | 763.7 |
| Q21 | INNER | 10M×10M | 1:1 | 50% | 186.1 |
| Q22 | LEFT | 10M×10M | 1:1 | 50% | 10,193.8 |
| Q23 | FULL | 10M×10M | 1:1 | 50% | 10,194.7 |
Note that Q22 and Q23 will be about 20x faster when #21184 merges, so
taking 10 seconds to run is just a short-term issue.
## Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->
<!--
If there are any breaking changes to public APIs, please add the `api
change` label.
-->
No.
---
benchmarks/src/smj.rs | 166 +++++++++++++++++++++++++-------------------------
1 file changed, 83 insertions(+), 83 deletions(-)
diff --git a/benchmarks/src/smj.rs b/benchmarks/src/smj.rs
index d782762a1b..1829b77a3d 100644
--- a/benchmarks/src/smj.rs
+++ b/benchmarks/src/smj.rs
@@ -60,27 +60,27 @@ pub struct RunOpt {
/// - Key cardinality (rows per key)
/// - Filter selectivity (if applicable)
const SMJ_QUERIES: &[&str] = &[
- // Q1: INNER 100K x 100K | 1:1
+ // Q1: INNER 1M x 1M | 1:1
r#"
WITH t1_sorted AS (
- SELECT value as key FROM range(100000) ORDER BY value
+ SELECT value as key FROM range(1000000) ORDER BY value
),
t2_sorted AS (
- SELECT value as key FROM range(100000) ORDER BY value
+ SELECT value as key FROM range(1000000) ORDER BY value
)
SELECT t1_sorted.key as k1, t2_sorted.key as k2
FROM t1_sorted JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
"#,
- // Q2: INNER 100K x 1M | 1:10
+ // Q2: INNER 1M x 10M | 1:10
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
@@ -101,16 +101,16 @@ const SMJ_QUERIES: &[&str] = &[
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
FROM t1_sorted JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
"#,
- // Q4: INNER 100K x 1M | 1:10 | 1%
+ // Q4: INNER 1M x 10M | 1:10 | 1%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
@@ -133,63 +133,63 @@ const SMJ_QUERIES: &[&str] = &[
FROM t1_sorted JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
WHERE t1_sorted.data <> t2_sorted.data AND t2_sorted.data % 10 = 0
"#,
- // Q6: LEFT 100K x 1M | 1:10
+ // Q6: LEFT 1M x 10M | 1:10
r#"
WITH t1_sorted AS (
- SELECT value % 10500 as key, value as data
- FROM range(100000)
+ SELECT value % 105000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
FROM t1_sorted LEFT JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
"#,
- // Q7: LEFT 100K x 1M | 1:10 | 50%
+ // Q7: LEFT 1M x 10M | 1:10 | 50%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
FROM t1_sorted LEFT JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
WHERE t2_sorted.data IS NULL OR t2_sorted.data % 2 = 0
"#,
- // Q8: FULL 100K x 100K | 1:10
+ // Q8: FULL 1M x 1M | 1:10
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 12500 as key, value as data
- FROM range(100000)
+ SELECT value % 125000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
)
SELECT t1_sorted.key as k1, t1_sorted.data as d1,
t2_sorted.key as k2, t2_sorted.data as d2
FROM t1_sorted FULL JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
"#,
- // Q9: FULL 100K x 1M | 1:10 | 10%
+ // Q9: FULL 1M x 10M | 1:10 | 10%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key as k1, t1_sorted.data as d1,
@@ -199,16 +199,16 @@ const SMJ_QUERIES: &[&str] = &[
OR t1_sorted.data <> t2_sorted.data)
AND (t1_sorted.data IS NULL OR t1_sorted.data % 10 = 0)
"#,
- // Q10: LEFT SEMI 100K x 1M | 1:10
+ // Q10: LEFT SEMI 1M x 10M | 1:10
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key
- FROM range(1000000)
+ SELECT value % 100000 as key
+ FROM range(10000000)
ORDER BY key
)
SELECT t1_sorted.key, t1_sorted.data
@@ -218,16 +218,16 @@ const SMJ_QUERIES: &[&str] = &[
WHERE t2_sorted.key = t1_sorted.key
)
"#,
- // Q11: LEFT SEMI 100K x 1M | 1:10 | 1%
+ // Q11: LEFT SEMI 1M x 10M | 1:10 | 1%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data
@@ -239,16 +239,16 @@ const SMJ_QUERIES: &[&str] = &[
AND t2_sorted.data % 100 = 0
)
"#,
- // Q12: LEFT SEMI 100K x 1M | 1:10 | 50%
+ // Q12: LEFT SEMI 1M x 10M | 1:10 | 50%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data
@@ -260,16 +260,16 @@ const SMJ_QUERIES: &[&str] = &[
AND t2_sorted.data % 2 = 0
)
"#,
- // Q13: LEFT SEMI 100K x 1M | 1:10 | 90%
+ // Q13: LEFT SEMI 1M x 10M | 1:10 | 90%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(1000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(10000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data
@@ -281,16 +281,16 @@ const SMJ_QUERIES: &[&str] = &[
AND t2_sorted.data % 10 <> 0
)
"#,
- // Q14: LEFT ANTI 100K x 1M | 1:10
+ // Q14: LEFT ANTI 1M x 10M | 1:10
r#"
WITH t1_sorted AS (
- SELECT value % 10500 as key, value as data
- FROM range(100000)
+ SELECT value % 105000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key
- FROM range(1000000)
+ SELECT value % 100000 as key
+ FROM range(10000000)
ORDER BY key
)
SELECT t1_sorted.key, t1_sorted.data
@@ -300,16 +300,16 @@ const SMJ_QUERIES: &[&str] = &[
WHERE t2_sorted.key = t1_sorted.key
)
"#,
- // Q15: LEFT ANTI 100K x 1M | 1:10 | partial match
+ // Q15: LEFT ANTI 1M x 10M | 1:10 | partial match
r#"
WITH t1_sorted AS (
- SELECT value % 12000 as key, value as data
- FROM range(100000)
+ SELECT value % 120000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key
- FROM range(1000000)
+ SELECT value % 100000 as key
+ FROM range(10000000)
ORDER BY key
)
SELECT t1_sorted.key, t1_sorted.data
@@ -319,16 +319,16 @@ const SMJ_QUERIES: &[&str] = &[
WHERE t2_sorted.key = t1_sorted.key
)
"#,
- // Q16: LEFT ANTI 100K x 100K | 1:1 | stress
+ // Q16: LEFT ANTI 1M x 1M | 1:1 | stress
r#"
WITH t1_sorted AS (
- SELECT value % 11000 as key, value as data
- FROM range(100000)
+ SELECT value % 110000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key
- FROM range(100000)
+ SELECT value % 100000 as key
+ FROM range(1000000)
ORDER BY key
)
SELECT t1_sorted.key, t1_sorted.data
@@ -338,32 +338,32 @@ const SMJ_QUERIES: &[&str] = &[
WHERE t2_sorted.key = t1_sorted.key
)
"#,
- // Q17: INNER 100K x 5M | 1:50 | 5%
+ // Q17: INNER 1M x 50M | 1:50 | 5%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(5000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(50000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data as d1, t2_sorted.data as d2
FROM t1_sorted JOIN t2_sorted ON t1_sorted.key = t2_sorted.key
WHERE t2_sorted.data <> t1_sorted.data AND t2_sorted.data % 20 = 0
"#,
- // Q18: LEFT SEMI 100K x 5M | 1:50 | 2%
+ // Q18: LEFT SEMI 1M x 50M | 1:50 | 2%
r#"
WITH t1_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(100000)
+ SELECT value % 100000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key, value as data
- FROM range(5000000)
+ SELECT value % 100000 as key, value as data
+ FROM range(50000000)
ORDER BY key, data
)
SELECT t1_sorted.key, t1_sorted.data
@@ -375,16 +375,16 @@ const SMJ_QUERIES: &[&str] = &[
AND t2_sorted.data % 50 = 0
)
"#,
- // Q19: LEFT ANTI 100K x 5M | 1:50 | partial match
+ // Q19: LEFT ANTI 1M x 50M | 1:50 | partial match
r#"
WITH t1_sorted AS (
- SELECT value % 15000 as key, value as data
- FROM range(100000)
+ SELECT value % 150000 as key, value as data
+ FROM range(1000000)
ORDER BY key, data
),
t2_sorted AS (
- SELECT value % 10000 as key
- FROM range(5000000)
+ SELECT value % 100000 as key
+ FROM range(50000000)
ORDER BY key
)
SELECT t1_sorted.key, t1_sorted.data
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
