PushdownFilter optimizations by joroKr21 · Pull Request #21668 · apache/datafusion

joroKr21 · 2026-04-16T06:59:23Z

Which issue does this PR close?

Addresses perf: push_down_filter is pathologically slow for some plans #20002
Includes Actually preserve predicate execution order in PushDownFilter #21643
Looks like we had a similar idea with perf(push_down_filter): skip filter revalidation #21667, cc @kumarUjjawal
- I like the name try_unchecked more so I would adopt that
- I'm not sure if we need to keep the unaliasing, the optimizer rule already does that
- I have additional changes for more aggressive optimizations, so I'm happy to combine the PRs or rebase after merge

Rationale for this change

What changes are included in this PR?

Add a hidden Filter::new_unchecked constructor that skips type-checking
Less allocations, more modification of mutable plan nodes
Less cloning, use references when possible

Are these changes tested?

Relying on existing tests mostly, added a few more tests.

Are there any user-facing changes?

make_filter is removed, probably wasn't meant to be a public function.

kumarUjjawal · 2026-04-16T10:33:52Z

@joroKr21 If the direction is agreed by the reviewer you should combine the pr, I will be happy to help.

- Add a hidden `Filter::new` constructor that skips type-checking - Less allocations, more modification of mutable plan nodes - Less cloning, use references when possible

kumarUjjawal · 2026-04-29T16:47:52Z

run benchmark sql_planner

adriangbot · 2026-04-29T16:52:58Z

🤖 Criterion benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4345734491-1918-pxl6d 6.12.55+ #1 SMP Sun Feb 1 08:59:41 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing opt/pushdown-filter (740cf2c) to 3aefba7 (merge-base) diff
BENCH_NAME=sql_planner
BENCH_COMMAND=cargo bench --features=parquet --bench sql_planner
BENCH_FILTER=
Results will be posted here when complete

File an issue against this benchmark runner

kumarUjjawal · 2026-04-29T17:08:18Z

show benchmark queue

adriangbot · 2026-04-29T17:08:20Z

Hi @kumarUjjawal, you asked to view the benchmark queue (#21668 (comment)).

Comment	Repo	PR	User	Benchmarks	Status
#4345685099	adriangb/datafusion	#11	adriangb	["clickbench_partitioned"]	running
#4345734491	apache/datafusion	#21668	kumarUjjawal	["sql_planner"]	running
#4345770652	apache/datafusion	#21931	adriangb	["tpch"]	running
#4345772857	apache/datafusion	#21931	adriangb	["tpch"]	running

File an issue against this benchmark runner

kumarUjjawal · 2026-04-29T17:38:45Z

run benchmark push_down_filter

adriangbot · 2026-04-29T17:42:05Z

🤖 Criterion benchmark running (GKE) | trigger
Instance: c4a-highmem-16 (12 vCPU / 65 GiB) | Linux bench-c4346084743-1922-s4rh4 6.12.55+ #1 SMP Sun Feb 1 08:59:41 UTC 2026 aarch64 GNU/Linux

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Comparing opt/pushdown-filter (740cf2c) to 3aefba7 (merge-base) diff
BENCH_NAME=push_down_filter
BENCH_COMMAND=cargo bench --features=parquet --bench push_down_filter
BENCH_FILTER=
Results will be posted here when complete

File an issue against this benchmark runner

adriangbot · 2026-04-29T17:47:59Z

🤖 Criterion benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Details

group                                                 main                                   opt_pushdown-filter
-----                                                 ----                                   -------------------
logical_aggregate_with_join                           1.00    409.0±1.35µs        ? ?/sec    1.01    412.5±0.94µs        ? ?/sec
logical_plan_struct_join_agg_sort                     1.00    168.6±1.29µs        ? ?/sec    1.02    171.4±1.36µs        ? ?/sec
logical_select_all_from_1000                          1.00      8.0±0.01ms        ? ?/sec    1.00      8.0±0.02ms        ? ?/sec
logical_select_one_from_700                           1.01    273.4±6.71µs        ? ?/sec    1.00    270.4±1.57µs        ? ?/sec
logical_trivial_join_high_numbered_columns            1.00    249.1±0.95µs        ? ?/sec    1.00    249.5±0.98µs        ? ?/sec
logical_trivial_join_low_numbered_columns             1.00    237.4±0.83µs        ? ?/sec    1.00    236.8±1.07µs        ? ?/sec
physical_intersection                                 1.02    573.2±2.07µs        ? ?/sec    1.00    564.7±1.47µs        ? ?/sec
physical_join_consider_sort                           1.01   1006.4±6.97µs        ? ?/sec    1.00    995.4±3.41µs        ? ?/sec
physical_join_distinct                                1.00    231.2±0.68µs        ? ?/sec    1.00    230.8±1.05µs        ? ?/sec
physical_many_self_joins                              1.01      7.5±0.02ms        ? ?/sec    1.00      7.5±0.03ms        ? ?/sec
physical_plan_clickbench_all                          1.00    128.3±0.32ms        ? ?/sec    1.01    129.6±0.32ms        ? ?/sec
physical_plan_clickbench_q1                           1.01   1378.8±4.14µs        ? ?/sec    1.00  1362.5±11.94µs        ? ?/sec
physical_plan_clickbench_q10                          1.00      2.0±0.01ms        ? ?/sec    1.00      2.0±0.01ms        ? ?/sec
physical_plan_clickbench_q11                          1.01      2.2±0.01ms        ? ?/sec    1.00      2.2±0.01ms        ? ?/sec
physical_plan_clickbench_q12                          1.01      2.3±0.01ms        ? ?/sec    1.00      2.3±0.03ms        ? ?/sec
physical_plan_clickbench_q13                          1.01      2.1±0.01ms        ? ?/sec    1.00      2.0±0.01ms        ? ?/sec
physical_plan_clickbench_q14                          1.01      2.2±0.14ms        ? ?/sec    1.00      2.2±0.02ms        ? ?/sec
physical_plan_clickbench_q15                          1.00      2.1±0.03ms        ? ?/sec    1.00      2.1±0.01ms        ? ?/sec
physical_plan_clickbench_q16                          1.00   1741.2±5.29µs        ? ?/sec    1.00   1740.6±8.70µs        ? ?/sec
physical_plan_clickbench_q17                          1.01   1807.0±6.19µs        ? ?/sec    1.00  1792.2±24.57µs        ? ?/sec
physical_plan_clickbench_q18                          1.00   1628.5±4.98µs        ? ?/sec    1.00   1632.5±6.38µs        ? ?/sec
physical_plan_clickbench_q19                          1.00      2.0±0.01ms        ? ?/sec    1.01      2.1±0.01ms        ? ?/sec
physical_plan_clickbench_q2                           1.02   1793.3±9.36µs        ? ?/sec    1.00  1752.5±41.17µs        ? ?/sec
physical_plan_clickbench_q20                          1.00   1552.8±5.47µs        ? ?/sec    1.00   1552.0±8.87µs        ? ?/sec
physical_plan_clickbench_q21                          1.02  1815.1±10.12µs        ? ?/sec    1.00  1786.7±17.36µs        ? ?/sec
physical_plan_clickbench_q22                          1.03      2.3±0.01ms        ? ?/sec    1.00      2.2±0.01ms        ? ?/sec
physical_plan_clickbench_q23                          1.02      2.5±0.01ms        ? ?/sec    1.00      2.4±0.01ms        ? ?/sec
physical_plan_clickbench_q24                          1.00      5.7±0.01ms        ? ?/sec    1.00      5.7±0.01ms        ? ?/sec
physical_plan_clickbench_q25                          1.02   1934.7±5.73µs        ? ?/sec    1.00   1905.6±5.88µs        ? ?/sec
physical_plan_clickbench_q26                          1.02   1762.6±6.71µs        ? ?/sec    1.00   1729.3±5.45µs        ? ?/sec
physical_plan_clickbench_q27                          1.01  1949.9±12.04µs        ? ?/sec    1.00   1926.5±7.16µs        ? ?/sec
physical_plan_clickbench_q28                          1.01      2.4±0.00ms        ? ?/sec    1.00      2.4±0.01ms        ? ?/sec
physical_plan_clickbench_q29                          1.03      2.7±0.01ms        ? ?/sec    1.00      2.6±0.01ms        ? ?/sec
physical_plan_clickbench_q3                           1.03  1680.2±14.11µs        ? ?/sec    1.00   1623.9±4.85µs        ? ?/sec
physical_plan_clickbench_q30                          1.00     16.3±0.04ms        ? ?/sec    1.00     16.2±0.03ms        ? ?/sec
physical_plan_clickbench_q31                          1.02      2.5±0.01ms        ? ?/sec    1.00      2.5±0.01ms        ? ?/sec
physical_plan_clickbench_q32                          1.02      2.5±0.01ms        ? ?/sec    1.00      2.5±0.01ms        ? ?/sec
physical_plan_clickbench_q33                          1.01      2.1±0.01ms        ? ?/sec    1.00      2.1±0.01ms        ? ?/sec
physical_plan_clickbench_q34                          1.00   1755.0±5.49µs        ? ?/sec    1.02   1793.8±6.13µs        ? ?/sec
physical_plan_clickbench_q35                          1.00   1839.2±5.35µs        ? ?/sec    1.01   1859.4±5.40µs        ? ?/sec
physical_plan_clickbench_q36                          1.00      2.2±0.01ms        ? ?/sec    1.01      2.2±0.03ms        ? ?/sec
physical_plan_clickbench_q37                          1.00      2.6±0.01ms        ? ?/sec    1.00      2.6±0.01ms        ? ?/sec
physical_plan_clickbench_q38                          1.00      2.6±0.01ms        ? ?/sec    1.00      2.6±0.01ms        ? ?/sec
physical_plan_clickbench_q39                          1.00      2.7±0.01ms        ? ?/sec    1.01      2.7±0.01ms        ? ?/sec
physical_plan_clickbench_q4                           1.03  1476.3±12.04µs        ? ?/sec    1.00   1432.4±5.17µs        ? ?/sec
physical_plan_clickbench_q40                          1.00      3.5±0.01ms        ? ?/sec    1.01      3.5±0.01ms        ? ?/sec
physical_plan_clickbench_q41                          1.00      3.0±0.01ms        ? ?/sec    1.01      3.0±0.01ms        ? ?/sec
physical_plan_clickbench_q42                          1.00      3.1±0.01ms        ? ?/sec    1.02      3.2±0.10ms        ? ?/sec
physical_plan_clickbench_q43                          1.02      3.3±0.02ms        ? ?/sec    1.00      3.3±0.01ms        ? ?/sec
physical_plan_clickbench_q44                          1.02   1574.9±8.48µs        ? ?/sec    1.00   1537.1±5.57µs        ? ?/sec
physical_plan_clickbench_q45                          1.02   1574.2±6.15µs        ? ?/sec    1.00   1539.9±6.10µs        ? ?/sec
physical_plan_clickbench_q46                          1.02  1923.1±18.78µs        ? ?/sec    1.00   1890.5±7.58µs        ? ?/sec
physical_plan_clickbench_q47                          1.01      2.7±0.01ms        ? ?/sec    1.00      2.6±0.01ms        ? ?/sec
physical_plan_clickbench_q48                          1.00      2.9±0.01ms        ? ?/sec    1.00      2.9±0.01ms        ? ?/sec
physical_plan_clickbench_q49                          1.00      2.9±0.01ms        ? ?/sec    1.00      2.9±0.01ms        ? ?/sec
physical_plan_clickbench_q5                           1.02  1594.1±10.86µs        ? ?/sec    1.00  1563.8±36.72µs        ? ?/sec
physical_plan_clickbench_q50                          1.00      2.9±0.00ms        ? ?/sec    1.00      2.9±0.01ms        ? ?/sec
physical_plan_clickbench_q51                          1.00   1991.3±6.29µs        ? ?/sec    1.01      2.0±0.01ms        ? ?/sec
physical_plan_clickbench_q6                           1.03  1615.7±14.54µs        ? ?/sec    1.00  1574.8±17.49µs        ? ?/sec
physical_plan_clickbench_q7                           1.01   1630.3±5.54µs        ? ?/sec    1.00   1615.1±6.04µs        ? ?/sec
physical_plan_clickbench_q8                           1.02   1935.2±5.85µs        ? ?/sec    1.00   1904.6±6.54µs        ? ?/sec
physical_plan_clickbench_q9                           1.01   1914.5±9.95µs        ? ?/sec    1.00   1899.5±5.49µs        ? ?/sec
physical_plan_struct_join_agg_sort                    1.01   1385.4±2.25µs        ? ?/sec    1.00   1373.2±8.47µs        ? ?/sec
physical_plan_tpcds_all                               1.01    727.2±1.67ms        ? ?/sec    1.00    718.4±8.38ms        ? ?/sec
physical_plan_tpch_all                                1.03     48.7±0.42ms        ? ?/sec    1.00     47.2±0.51ms        ? ?/sec
physical_plan_tpch_q1                                 1.01   1580.5±5.13µs        ? ?/sec    1.00   1571.1±2.03µs        ? ?/sec
physical_plan_tpch_q10                                1.02      3.0±0.00ms        ? ?/sec    1.00      3.0±0.00ms        ? ?/sec
physical_plan_tpch_q11                                1.00      2.2±0.00ms        ? ?/sec    1.04      2.3±0.22ms        ? ?/sec
physical_plan_tpch_q12                                1.04   1337.1±2.13µs        ? ?/sec    1.00   1290.5±2.47µs        ? ?/sec
physical_plan_tpch_q13                                1.00   1000.7±2.72µs        ? ?/sec    1.00   995.7±16.31µs        ? ?/sec
physical_plan_tpch_q14                                1.03   1476.4±5.40µs        ? ?/sec    1.00   1426.6±3.92µs        ? ?/sec
physical_plan_tpch_q16                                1.03   1687.8±5.94µs        ? ?/sec    1.00   1646.3±2.97µs        ? ?/sec
physical_plan_tpch_q17                                1.04   1841.2±3.34µs        ? ?/sec    1.00   1774.0±3.23µs        ? ?/sec
physical_plan_tpch_q18                                1.03   1935.8±3.05µs        ? ?/sec    1.00   1886.4±1.70µs        ? ?/sec
physical_plan_tpch_q19                                1.30      2.5±0.00ms        ? ?/sec    1.00   1942.9±3.42µs        ? ?/sec
physical_plan_tpch_q2                                 1.01      4.3±0.00ms        ? ?/sec    1.00      4.3±0.01ms        ? ?/sec
physical_plan_tpch_q20                                1.02      2.4±0.00ms        ? ?/sec    1.00      2.4±0.00ms        ? ?/sec
physical_plan_tpch_q21                                1.01      3.2±0.00ms        ? ?/sec    1.00      3.1±0.06ms        ? ?/sec
physical_plan_tpch_q22                                1.02   1581.4±5.49µs        ? ?/sec    1.00   1544.2±5.25µs        ? ?/sec
physical_plan_tpch_q3                                 1.02      2.0±0.00ms        ? ?/sec    1.00      2.0±0.00ms        ? ?/sec
physical_plan_tpch_q4                                 1.00   1174.1±1.45µs        ? ?/sec    1.00   1177.8±1.96µs        ? ?/sec
physical_plan_tpch_q5                                 1.01      2.6±0.00ms        ? ?/sec    1.00      2.6±0.00ms        ? ?/sec
physical_plan_tpch_q6                                 1.01    673.6±1.76µs        ? ?/sec    1.00    666.1±1.80µs        ? ?/sec
physical_plan_tpch_q7                                 1.02      3.2±0.00ms        ? ?/sec    1.00      3.2±0.03ms        ? ?/sec
physical_plan_tpch_q8                                 1.01      4.2±0.00ms        ? ?/sec    1.00      4.1±0.06ms        ? ?/sec
physical_plan_tpch_q9                                 1.01      2.9±0.00ms        ? ?/sec    1.00      2.9±0.00ms        ? ?/sec
physical_select_aggregates_from_200                   1.00     14.7±0.02ms        ? ?/sec    1.00     14.7±0.03ms        ? ?/sec
physical_select_all_from_1000                         1.00     17.4±0.04ms        ? ?/sec    1.00     17.4±0.04ms        ? ?/sec
physical_select_one_from_700                          1.01    714.1±5.21µs        ? ?/sec    1.00    705.7±3.27µs        ? ?/sec
physical_sorted_union_order_by_10_int64               1.00      4.8±0.01ms        ? ?/sec    1.00      4.8±0.00ms        ? ?/sec
physical_sorted_union_order_by_10_uint64              1.01     11.7±0.06ms        ? ?/sec    1.00     11.6±0.02ms        ? ?/sec
physical_sorted_union_order_by_50_int64               1.01    114.4±0.40ms        ? ?/sec    1.00    113.6±0.60ms        ? ?/sec
physical_sorted_union_order_by_50_uint64              1.00    600.3±1.63ms        ? ?/sec    1.00    600.7±3.30ms        ? ?/sec
physical_theta_join_consider_sort                     1.01   1042.6±6.59µs        ? ?/sec    1.00   1030.5±6.33µs        ? ?/sec
physical_unnest_to_join                               1.01    615.6±1.49µs        ? ?/sec    1.00    608.1±1.86µs        ? ?/sec
physical_window_function_partition_by_12_on_values    1.03    757.4±1.70µs        ? ?/sec    1.00    738.1±2.43µs        ? ?/sec
physical_window_function_partition_by_30_on_values    1.01   1492.8±2.77µs        ? ?/sec    1.00   1479.9±6.19µs        ? ?/sec
physical_window_function_partition_by_4_on_values     1.03    458.5±1.28µs        ? ?/sec    1.00    443.0±2.14µs        ? ?/sec
physical_window_function_partition_by_7_on_values     1.03    567.7±2.17µs        ? ?/sec    1.00    550.9±2.39µs        ? ?/sec
physical_window_function_partition_by_8_on_values     1.03    609.0±1.73µs        ? ?/sec    1.00    592.5±2.64µs        ? ?/sec
with_param_values_many_columns                        1.00    457.5±3.20µs        ? ?/sec    1.00    457.2±1.75µs        ? ?/sec

Resource Usage

base (merge-base)

Metric	Value
Wall time	1240.3s
Peak memory	19.7 GiB
Avg memory	19.7 GiB
CPU user	1482.6s
CPU sys	1.7s
Peak spill	0 B

branch

Metric	Value
Wall time	1245.3s
Peak memory	19.7 GiB
Avg memory	19.7 GiB
CPU user	1486.3s
CPU sys	1.2s
Peak spill	0 B

File an issue against this benchmark runner

adriangbot · 2026-04-29T17:51:59Z

🤖 Criterion benchmark completed (GKE) | trigger

Instance: c4a-highmem-16 (12 vCPU / 65 GiB)

CPU Details (lscpu)

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  16
On-line CPU(s) list:                     0-15
Vendor ID:                               ARM
Model name:                              Neoverse-V2
Model:                                   1
Thread(s) per core:                      1
Core(s) per cluster:                     16
Socket(s):                               -
Cluster(s):                              1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti
L1d cache:                               1 MiB (16 instances)
L1i cache:                               1 MiB (16 instances)
L2 cache:                                32 MiB (16 instances)
L3 cache:                                80 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-15
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

Details

group               main                                   opt_pushdown-filter
-----               ----                                   -------------------
push_down_filter    1.00     53.0±0.27µs        ? ?/sec    1.01     53.6±0.17µs        ? ?/sec

Resource Usage

base (merge-base)

Metric	Value
Wall time	15.0s
Peak memory	5.5 GiB
Avg memory	5.4 GiB
CPU user	12.2s
CPU sys	0.8s
Peak spill	0 B

branch

Metric	Value
Wall time	10.0s
Peak memory	5.5 GiB
Avg memory	5.5 GiB
CPU user	11.8s
CPU sys	0.2s
Peak spill	0 B

File an issue against this benchmark runner

kumarUjjawal · 2026-04-29T18:30:45Z

@joroKr21 The benchmarks are slower, can you investigate this.

github-actions Bot added logical-expr Logical plan and expressions optimizer Optimizer rules sqllogictest SQL Logic Tests (.slt) labels Apr 16, 2026

joroKr21 force-pushed the opt/pushdown-filter branch 2 times, most recently from 8db87e3 to 7aee0a0 Compare April 16, 2026 08:04

joroKr21 marked this pull request as ready for review April 16, 2026 08:52

joroKr21 force-pushed the opt/pushdown-filter branch from 7aee0a0 to 6ce2a70 Compare April 16, 2026 09:57

joroKr21 added 4 commits April 29, 2026 19:05

Actually preserve predicate execution order in PushDownFilter

ea4b6b7

PushdownFilter optimizations

9be5b52

- Add a hidden `Filter::new` constructor that skips type-checking - Less allocations, more modification of mutable plan nodes - Less cloning, use references when possible

Use Filter::new_unchecked

d837e3d

Add expr_columns helper function

740cf2c

joroKr21 force-pushed the opt/pushdown-filter branch from 6ce2a70 to 740cf2c Compare April 29, 2026 16:10

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

PushdownFilter optimizations #21668

PushdownFilter optimizations #21668
joroKr21 wants to merge 4 commits intoapache:mainfrom
coralogix:opt/pushdown-filter

joroKr21 commented Apr 16, 2026 •

edited

Loading

Uh oh!

kumarUjjawal commented Apr 16, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

joroKr21 commented Apr 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Which issue does this PR close?

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Uh oh!

kumarUjjawal commented Apr 16, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

adriangbot commented Apr 29, 2026

Uh oh!

kumarUjjawal commented Apr 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

joroKr21 commented Apr 16, 2026 •

edited

Loading