You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "RatulDawar (via GitHub)" <gi...@apache.org> on 2026/04/14 22:01:26 UTC
[PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
RatulDawar opened a new pull request, #21632:
URL: https://github.com/apache/datafusion/pull/21632
## Summary
This PR fixes an indefinite wait situation in Hash Join when dynamic filtering is enabled and some partitions have zero rows.
### The Issue
The short-circuit optimization in `state_after_build_ready` was marking empty partitions as `Completed` immediately. However, when dynamic filtering is used, all partitions must report their build-side information to a shared accumulator (which uses a barrier). Empty partitions were skipping this report, causing the barrier to never be reached and leading to a deadlock.
Additionally, once the short-circuit was disabled to allow for coordination, a `debug_assert!` in `process_probe_batch` was being triggered for join types that produce empty results when the build side is empty.
### The Fix
1. Modified `state_after_build_ready` to only short-circuit if no `build_accumulator` is present. This ensures that empty partitions still report their (empty) filters to the shared accumulator before completing.
2. Updated `process_probe_batch` to handle empty build sides gracefully when coordination is enabled, bypassing the join logic and the `debug_assert!`.
## Test plan
1. Run a hash join with dynamic filtering enabled and a high number of partitions such that some partitions are empty (e.g. TPC-H Q18 with target_partitions=24).
2. Verify that the query completes instead of hanging or panicking.
Made with [Cursor](https://cursor.com)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248070285
run benchmark tpch tpcds
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangb (via GitHub)" <gi...@apache.org>.
adriangb commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4261636795
I *think* https://github.com/apache/datafusion/pull/21666 is the right solution but this is a non trivial problem so some more ๐ง on it would be appreciated
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4252343008
@adriangb GTG ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248084846
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21632#issuecomment-4248070285)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4248070285-1266-2vnm8 6.12.55+ #1 SMP Sun Feb 1 08:59:41 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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>
Comparing fix-hash-join-dynamic-filter-deadlock (2745dce8a57f354ce7173e9ed858bc63e377eb86) to 2818abb (merge-base) [diff](https://github.com/apache/datafusion/compare/2818abb8a10bf8b001669c15cf597f9ddfb86ed1..2745dce8a57f354ce7173e9ed858bc63e377eb86) using: tpcds
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4261176046
> Did the original PR get backported to 53.0? It was a performance improvement not a regression fix.
Let me try to reproduce the issue on `branch-53`
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "Omega359 (via GitHub)" <gi...@apache.org>.
Omega359 commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4257155536
> > @adriangb appreciate if you can have a second look and we can merge it and backport to 53.1.0
>
> Did the original PR get backported to 53.0? It was a performance improvement not a regression fix.
Ohhh ... that's a good point, I never actually checked before ๐คฆ๐ป It doesn't look like it was.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4253993479
Added a deterministic integration test in `datafusion/core/tests/repro_deadlock_integration.rs` that forces a `Partitioned` join with an empty partition to ensure the dynamic filter coordination deadlock does not regress.
The test uses `MemTable` with explicit partitioning to guarantee the short-circuit condition and includes a 1-minute timeout to detect potential deadlocks.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar closed pull request #21632: fix: prevent hash join deadlock when dynamic filtering is enabled
URL: https://github.com/apache/datafusion/pull/21632
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248086666
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21632#issuecomment-4248070285)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4248070285-1265-fbpvq 6.12.55+ #1 SMP Sun Feb 1 08:59:41 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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>
Comparing fix-hash-join-dynamic-filter-deadlock (2745dce8a57f354ce7173e9ed858bc63e377eb86) to 2818abb (merge-base) [diff](https://github.com/apache/datafusion/compare/2818abb8a10bf8b001669c15cf597f9ddfb86ed1..2745dce8a57f354ce7173e9ed858bc63e377eb86) using: tpch
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4256727602
@adriangb Gtg ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "ratuldawar11 (via GitHub)" <gi...@apache.org>.
ratuldawar11 commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4249430352
@Omega359 have fixed the failing CI checks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4249465321
@Omega359 have fixed the failing CI checks and tests
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4251764433
@Dandandan any action to be taken based on the benchmarks ? Seem fine to me
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4251868317
> @Dandandan any action to be taken based on the benchmarks ? Seem fine to me
Seems good to me as well.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "Omega359 (via GitHub)" <gi...@apache.org>.
Omega359 commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248744000
@RatulDawar can you look at the test failures and fix the cargo fmt issue.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248161829
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21632#issuecomment-4248070285)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and fix-hash-join-dynamic-filter-deadlock
--------------------
Benchmark tpcds_sf1.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ fix-hash-join-dynamic-filter-deadlock โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 1 โ 6.77 / 7.22 ยฑ0.80 / 8.82 ms โ 43.40 / 44.54 ยฑ0.81 / 45.91 ms โ 6.17x slower โ
โ QQuery 2 โ 145.02 / 145.56 ยฑ0.41 / 146.06 ms โ 146.23 / 146.90 ยฑ0.44 / 147.48 ms โ no change โ
โ QQuery 3 โ 113.70 / 114.39 ยฑ0.59 / 115.42 ms โ 114.15 / 114.77 ยฑ0.40 / 115.25 ms โ no change โ
โ QQuery 4 โ 1324.84 / 1355.85 ยฑ22.58 / 1385.83 ms โ 1345.11 / 1361.94 ยฑ15.31 / 1383.17 ms โ no change โ
โ QQuery 5 โ 173.36 / 174.59 ยฑ0.74 / 175.38 ms โ 173.72 / 174.95 ยฑ0.71 / 175.70 ms โ no change โ
โ QQuery 6 โ 824.26 / 861.57 ยฑ29.98 / 913.84 ms โ 837.64 / 868.71 ยฑ30.19 / 926.32 ms โ no change โ
โ QQuery 7 โ 345.07 / 347.22 ยฑ1.58 / 349.72 ms โ 343.77 / 347.50 ยฑ2.47 / 351.26 ms โ no change โ
โ QQuery 8 โ 116.57 / 117.85 ยฑ0.84 / 119.01 ms โ 115.96 / 117.41 ยฑ0.83 / 118.29 ms โ no change โ
โ QQuery 9 โ 102.02 / 109.11 ยฑ10.45 / 129.25 ms โ 100.67 / 105.80 ยฑ6.05 / 117.21 ms โ no change โ
โ QQuery 10 โ 107.22 / 108.07 ยฑ0.82 / 109.45 ms โ 106.66 / 108.08 ยฑ0.93 / 109.49 ms โ no change โ
โ QQuery 11 โ 932.88 / 946.09 ยฑ8.40 / 957.27 ms โ 951.19 / 978.85 ยฑ20.65 / 1014.16 ms โ no change โ
โ QQuery 12 โ 45.93 / 47.09 ยฑ1.00 / 48.85 ms โ 48.15 / 48.99 ยฑ1.08 / 51.05 ms โ no change โ
โ QQuery 13 โ 402.98 / 404.81 ยฑ2.64 / 409.99 ms โ 405.27 / 407.76 ยฑ2.27 / 411.86 ms โ no change โ
โ QQuery 14 โ 1013.56 / 1020.01 ยฑ4.06 / 1026.32 ms โ 1010.31 / 1020.62 ยฑ9.33 / 1035.15 ms โ no change โ
โ QQuery 15 โ 15.94 / 16.70 ยฑ0.98 / 18.53 ms โ 15.97 / 17.53 ยฑ0.95 / 18.31 ms โ no change โ
โ QQuery 16 โ 7.51 / 7.92 ยฑ0.36 / 8.45 ms โ 39.74 / 40.47 ยฑ0.77 / 41.85 ms โ 5.11x slower โ
โ QQuery 17 โ 229.60 / 232.98 ยฑ2.40 / 236.01 ms โ 232.35 / 235.11 ยฑ1.95 / 238.20 ms โ no change โ
โ QQuery 18 โ 128.87 / 129.60 ยฑ0.64 / 130.70 ms โ 128.50 / 129.81 ยฑ0.86 / 130.97 ms โ no change โ
โ QQuery 19 โ 155.60 / 156.61 ยฑ0.51 / 157.01 ms โ 157.25 / 159.78 ยฑ1.83 / 162.12 ms โ no change โ
โ QQuery 20 โ 13.41 / 14.44 ยฑ0.69 / 15.35 ms โ 14.06 / 14.70 ยฑ0.47 / 15.50 ms โ no change โ
โ QQuery 21 โ 19.29 / 19.78 ยฑ0.31 / 20.26 ms โ 20.37 / 21.01 ยฑ0.43 / 21.69 ms โ 1.06x slower โ
โ QQuery 22 โ 493.09 / 496.27 ยฑ3.59 / 503.13 ms โ 495.25 / 500.42 ยฑ3.97 / 504.92 ms โ no change โ
โ QQuery 23 โ 869.46 / 880.93 ยฑ10.20 / 894.89 ms โ 893.85 / 910.42 ยฑ13.08 / 930.32 ms โ no change โ
โ QQuery 24 โ 383.93 / 388.54 ยฑ4.63 / 395.50 ms โ 406.22 / 408.49 ยฑ2.58 / 412.76 ms โ 1.05x slower โ
โ QQuery 25 โ 340.98 / 344.70 ยฑ3.22 / 349.79 ms โ 345.09 / 349.60 ยฑ2.87 / 353.19 ms โ no change โ
โ QQuery 26 โ 83.15 / 83.85 ยฑ0.70 / 85.01 ms โ 82.01 / 83.45 ยฑ1.34 / 85.59 ms โ no change โ
โ QQuery 27 โ 7.01 / 7.48 ยฑ0.53 / 8.48 ms โ 339.02 / 341.41 ยฑ1.33 / 342.82 ms โ 45.67x slower โ
โ QQuery 28 โ 150.04 / 151.07 ยฑ1.17 / 153.33 ms โ 149.00 / 151.17 ยฑ1.58 / 153.84 ms โ no change โ
โ QQuery 29 โ 282.77 / 285.58 ยฑ2.44 / 289.07 ms โ 289.16 / 290.10 ยฑ0.62 / 290.94 ms โ no change โ
โ QQuery 30 โ 43.33 / 44.36 ยฑ1.16 / 46.58 ms โ 46.28 / 47.54 ยฑ1.08 / 49.47 ms โ 1.07x slower โ
โ QQuery 31 โ 171.36 / 172.40 ยฑ1.03 / 173.93 ms โ 172.74 / 174.13 ยฑ0.72 / 174.76 ms โ no change โ
โ QQuery 32 โ 56.57 / 58.59 ยฑ1.52 / 60.58 ms โ 58.00 / 58.74 ยฑ0.87 / 60.42 ms โ no change โ
โ QQuery 33 โ 142.68 / 143.47 ยฑ0.45 / 144.02 ms โ 141.33 / 143.56 ยฑ1.67 / 145.99 ms โ no change โ
โ QQuery 34 โ 7.10 / 7.27 ยฑ0.24 / 7.75 ms โ 106.01 / 107.13 ยฑ0.80 / 108.45 ms โ 14.74x slower โ
โ QQuery 35 โ 104.86 / 109.79 ยฑ2.67 / 112.42 ms โ 106.83 / 107.96 ยฑ0.71 / 108.93 ms โ no change โ
โ QQuery 36 โ 6.47 / 6.84 ยฑ0.46 / 7.75 ms โ 215.42 / 219.74 ยฑ2.51 / 222.74 ms โ 32.11x slower โ
โ QQuery 37 โ 8.41 / 9.07 ยฑ0.76 / 10.53 ms โ 176.33 / 179.40 ยฑ2.45 / 183.82 ms โ 19.78x slower โ
โ QQuery 38 โ 84.01 / 87.65 ยฑ3.91 / 94.97 ms โ 84.19 / 88.18 ยฑ2.96 / 93.24 ms โ no change โ
โ QQuery 39 โ 129.15 / 130.26 ยฑ1.18 / 132.41 ms โ 127.48 / 131.04 ยฑ2.45 / 134.76 ms โ no change โ
โ QQuery 40 โ 111.29 / 116.67 ยฑ6.27 / 128.96 ms โ 112.65 / 117.72 ยฑ6.31 / 130.05 ms โ no change โ
โ QQuery 41 โ 14.22 / 14.88 ยฑ0.40 / 15.44 ms โ 14.40 / 15.98 ยฑ0.96 / 17.43 ms โ 1.07x slower โ
โ QQuery 42 โ 107.61 / 109.87 ยฑ1.17 / 110.82 ms โ 107.24 / 108.33 ยฑ1.23 / 110.31 ms โ no change โ
โ QQuery 43 โ 6.17 / 6.63 ยฑ0.62 / 7.79 ms โ 84.44 / 84.99 ยฑ0.43 / 85.61 ms โ 12.82x slower โ
โ QQuery 44 โ 12.08 / 12.18 ยฑ0.06 / 12.26 ms โ 11.90 / 12.34 ยฑ0.27 / 12.70 ms โ no change โ
โ QQuery 45 โ 50.92 / 51.87 ยฑ1.19 / 54.22 ms โ 51.54 / 52.54 ยฑ0.66 / 53.24 ms โ no change โ
โ QQuery 46 โ 8.65 / 9.01 ยฑ0.25 / 9.32 ms โ 230.98 / 234.57 ยฑ2.05 / 236.80 ms โ 26.05x slower โ
โ QQuery 47 โ 712.97 / 720.60 ยฑ4.71 / 725.62 ms โ 713.44 / 732.59 ยฑ15.18 / 748.95 ms โ no change โ
โ QQuery 48 โ 292.35 / 296.06 ยฑ3.10 / 300.08 ms โ 288.62 / 292.12 ยฑ2.83 / 295.62 ms โ no change โ
โ QQuery 49 โ 252.84 / 255.97 ยฑ2.52 / 260.41 ms โ 252.59 / 253.18 ยฑ0.75 / 254.57 ms โ no change โ
โ QQuery 50 โ 227.29 / 231.66 ยฑ4.46 / 238.77 ms โ 227.36 / 230.12 ยฑ2.60 / 233.56 ms โ no change โ
โ QQuery 51 โ 180.18 / 185.18 ยฑ3.29 / 190.18 ms โ 182.59 / 184.50 ยฑ2.21 / 188.42 ms โ no change โ
โ QQuery 52 โ 109.57 / 111.29 ยฑ1.30 / 113.61 ms โ 107.57 / 109.49 ยฑ1.33 / 111.57 ms โ no change โ
โ QQuery 53 โ 102.67 / 104.59 ยฑ1.09 / 106.08 ms โ 103.08 / 105.80 ยฑ2.32 / 109.28 ms โ no change โ
โ QQuery 54 โ 145.89 / 148.86 ยฑ1.90 / 151.42 ms โ 147.96 / 149.03 ยฑ1.01 / 150.68 ms โ no change โ
โ QQuery 55 โ 107.67 / 108.81 ยฑ1.44 / 111.32 ms โ 106.74 / 107.60 ยฑ0.77 / 108.69 ms โ no change โ
โ QQuery 56 โ 140.46 / 143.09 ยฑ2.45 / 146.39 ms โ 142.10 / 142.62 ยฑ0.46 / 143.24 ms โ no change โ
โ QQuery 57 โ 172.48 / 174.78 ยฑ2.03 / 178.10 ms โ 175.91 / 177.34 ยฑ0.85 / 178.39 ms โ no change โ
โ QQuery 58 โ 294.85 / 297.86 ยฑ2.28 / 301.09 ms โ 290.36 / 304.06 ยฑ8.26 / 312.18 ms โ no change โ
โ QQuery 59 โ 199.07 / 201.18 ยฑ1.22 / 202.78 ms โ 198.25 / 203.85 ยฑ5.85 / 214.76 ms โ no change โ
โ QQuery 60 โ 146.19 / 147.43 ยฑ0.90 / 148.68 ms โ 144.75 / 145.83 ยฑ0.77 / 146.69 ms โ no change โ
โ QQuery 61 โ 13.75 / 13.90 ยฑ0.20 / 14.30 ms โ 172.89 / 174.17 ยฑ1.01 / 175.91 ms โ 12.53x slower โ
โ QQuery 62 โ 892.72 / 940.92 ยฑ32.86 / 989.35 ms โ 886.94 / 948.18 ยฑ49.41 / 993.17 ms โ no change โ
โ QQuery 63 โ 103.93 / 106.79 ยฑ2.35 / 111.10 ms โ 106.11 / 108.08 ยฑ2.30 / 112.14 ms โ no change โ
โ QQuery 64 โ 692.48 / 695.62 ยฑ2.84 / 699.47 ms โ 684.00 / 691.39 ยฑ4.78 / 698.04 ms โ no change โ
โ QQuery 65 โ 253.06 / 256.67 ยฑ3.13 / 260.85 ms โ 256.33 / 258.47 ยฑ2.59 / 263.25 ms โ no change โ
โ QQuery 66 โ 231.97 / 258.29 ยฑ15.48 / 277.79 ms โ 243.06 / 256.13 ยฑ9.13 / 267.44 ms โ no change โ
โ QQuery 67 โ 312.40 / 320.54 ยฑ7.65 / 333.08 ms โ 305.35 / 316.61 ยฑ7.14 / 327.04 ms โ no change โ
โ QQuery 68 โ 9.66 / 10.88 ยฑ0.87 / 12.06 ms โ 276.70 / 282.29 ยฑ4.06 / 287.94 ms โ 25.95x slower โ
โ QQuery 69 โ 102.57 / 106.23 ยฑ2.74 / 110.53 ms โ 102.87 / 104.22 ยฑ0.73 / 105.04 ms โ no change โ
โ QQuery 70 โ 325.19 / 346.99 ยฑ14.77 / 370.53 ms โ 345.01 / 349.93 ยฑ6.14 / 360.92 ms โ no change โ
โ QQuery 71 โ 135.56 / 137.77 ยฑ2.00 / 140.17 ms โ 136.15 / 136.91 ยฑ0.65 / 137.78 ms โ no change โ
โ QQuery 72 โ 621.79 / 632.45 ยฑ8.96 / 645.48 ms โ 617.57 / 626.42 ยฑ7.72 / 637.37 ms โ no change โ
โ QQuery 73 โ 7.69 / 8.57 ยฑ0.90 / 9.88 ms โ 103.00 / 104.42 ยฑ1.07 / 105.72 ms โ 12.18x slower โ
โ QQuery 74 โ 595.10 / 606.10 ยฑ8.29 / 618.02 ms โ 580.76 / 588.59 ยฑ6.56 / 600.04 ms โ no change โ
โ QQuery 75 โ 276.16 / 279.95 ยฑ2.88 / 283.63 ms โ 277.72 / 279.39 ยฑ1.44 / 281.58 ms โ no change โ
โ QQuery 76 โ 132.77 / 135.89 ยฑ1.93 / 138.86 ms โ 131.94 / 133.20 ยฑ0.91 / 134.47 ms โ no change โ
โ QQuery 77 โ 188.56 / 191.28 ยฑ2.79 / 196.52 ms โ 188.09 / 190.06 ยฑ1.67 / 193.13 ms โ no change โ
โ QQuery 78 โ 345.33 / 348.31 ยฑ1.73 / 350.71 ms โ 343.33 / 346.81 ยฑ3.92 / 353.70 ms โ no change โ
โ QQuery 79 โ 231.30 / 237.04 ยฑ4.02 / 243.24 ms โ 232.87 / 234.74 ยฑ1.24 / 236.31 ms โ no change โ
โ QQuery 80 โ 325.32 / 328.01 ยฑ2.43 / 331.77 ms โ 326.70 / 328.86 ยฑ2.24 / 332.53 ms โ no change โ
โ QQuery 81 โ 27.11 / 28.06 ยฑ1.01 / 29.95 ms โ 25.93 / 27.55 ยฑ1.31 / 29.72 ms โ no change โ
โ QQuery 82 โ 202.50 / 204.20 ยฑ1.51 / 206.04 ms โ 202.23 / 204.51 ยฑ2.34 / 208.65 ms โ no change โ
โ QQuery 83 โ 39.50 / 40.28 ยฑ0.64 / 41.42 ms โ 38.99 / 39.70 ยฑ0.93 / 41.45 ms โ no change โ
โ QQuery 84 โ 49.18 / 49.80 ยฑ0.81 / 51.23 ms โ 48.88 / 49.57 ยฑ0.53 / 50.49 ms โ no change โ
โ QQuery 85 โ 150.70 / 152.66 ยฑ1.59 / 155.58 ms โ 149.06 / 151.21 ยฑ2.43 / 155.72 ms โ no change โ
โ QQuery 86 โ 39.30 / 40.76 ยฑ1.28 / 42.94 ms โ 38.99 / 40.13 ยฑ0.84 / 41.12 ms โ no change โ
โ QQuery 87 โ 87.36 / 91.10 ยฑ2.68 / 94.25 ms โ 85.14 / 88.98 ยฑ3.72 / 96.07 ms โ no change โ
โ QQuery 88 โ 101.36 / 102.08 ยฑ0.50 / 102.63 ms โ 100.88 / 101.17 ยฑ0.21 / 101.48 ms โ no change โ
โ QQuery 89 โ 119.77 / 121.10 ยฑ1.47 / 123.25 ms โ 119.12 / 121.03 ยฑ1.47 / 122.94 ms โ no change โ
โ QQuery 90 โ 23.87 / 25.00 ยฑ1.11 / 26.98 ms โ 23.92 / 24.98 ยฑ0.70 / 25.87 ms โ no change โ
โ QQuery 91 โ 63.62 / 66.64 ยฑ2.00 / 69.05 ms โ 63.41 / 64.89 ยฑ2.02 / 68.67 ms โ no change โ
โ QQuery 92 โ 58.96 / 60.02 ยฑ1.08 / 61.63 ms โ 57.79 / 59.14 ยฑ1.11 / 61.03 ms โ no change โ
โ QQuery 93 โ 190.87 / 192.13 ยฑ1.03 / 193.75 ms โ 188.99 / 190.21 ยฑ1.07 / 191.50 ms โ no change โ
โ QQuery 94 โ 61.91 / 62.62 ยฑ0.42 / 63.10 ms โ 61.68 / 62.65 ยฑ0.63 / 63.21 ms โ no change โ
โ QQuery 95 โ 130.64 / 132.08 ยฑ1.18 / 133.92 ms โ 128.38 / 129.45 ยฑ0.95 / 130.73 ms โ no change โ
โ QQuery 96 โ 71.52 / 74.28 ยฑ1.64 / 76.53 ms โ 73.48 / 74.48 ยฑ0.94 / 76.19 ms โ no change โ
โ QQuery 97 โ 125.18 / 127.35 ยฑ1.90 / 130.64 ms โ 127.25 / 128.18 ยฑ0.82 / 129.50 ms โ no change โ
โ QQuery 98 โ 154.81 / 157.15 ยฑ1.23 / 158.29 ms โ 151.71 / 156.15 ยฑ2.46 / 158.78 ms โ no change โ
โ QQuery 99 โ 10737.94 / 10792.58 ยฑ40.43 / 10845.93 ms โ 10746.24 / 10768.09 ยฑ25.09 / 10813.61 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 31676.16ms โ
โ Total Time (fix-hash-join-dynamic-filter-deadlock) โ 33463.26ms โ
โ Average Time (HEAD) โ 319.96ms โ
โ Average Time (fix-hash-join-dynamic-filter-deadlock) โ 338.01ms โ
โ Queries Faster โ 0 โ
โ Queries Slower โ 15 โ
โ Queries with No Change โ 84 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 158.7s |
| Peak memory | 5.7 GiB |
| Avg memory | 4.8 GiB |
| CPU user | 262.1s |
| CPU sys | 17.6s |
| Peak spill | 0 B |
**tpcds โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 167.6s |
| Peak memory | 5.8 GiB |
| Avg memory | 4.7 GiB |
| CPU user | 272.5s |
| CPU sys | 17.9s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4247493688
run bechmarks tpch tpcds
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "Dandandan (via GitHub)" <gi...@apache.org>.
Dandandan commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4249542586
run benchmark tpcds
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4263323905
> Let me try to reproduce the issue on `branch-53`
The deadlock test passes on `branch-53`, means there is no issue in 53.1.0, Thanks @mbutrovich for helping realize we solving non existing problem. Moving forward with the release
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4253411800
@comphead for a deadlock situation a failing test with say a minute of timeout should suffice ? I wanted to add a test but I think that timeout based test cases make the testcases flaky.
Another way to make thsi testable is I can add a unit test for the join state machine to manually check for correct expected states (poor coverage).
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "ratuldawar11 (via GitHub)" <gi...@apache.org>.
ratuldawar11 commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4252340041
@adriangb GTG ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangb (via GitHub)" <gi...@apache.org>.
adriangb commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4261205179
Please lets wait to merge this. I don't think this is the right fix.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4247502692
```
---- joins::hash_join::exec::tests::test_hash_join_skips_probe_on_empty_build_after_partition_bounds_report stdout ----
Error: Execution("bad data error")
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "mbutrovich (via GitHub)" <gi...@apache.org>.
mbutrovich commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4261556034
> Please lets wait to merge this. I'm not sure this is the right fix. But a good question is what the urgency is (depends on if the bug has been released or not, which it seems it has not).
Agreed, let's slow down and get this right.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4248812289
Sure checking
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4249551906
๐ค Benchmark running (GKE) | [trigger](https://github.com/apache/datafusion/pull/21632#issuecomment-4249542586)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB) | `Linux bench-c4249542586-1272-c4ncf 6.12.55+ #1 SMP Sun Feb 1 08:59:41 UTC 2026 aarch64 GNU/Linux`
<details><summary>CPU Details (lscpu)</summary>
```
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>
Comparing fix-hash-join-dynamic-filter-deadlock (1f5644dcbe5ff6d34c333f16dfca98266e4c62fb) to 2818abb (merge-base) [diff](https://github.com/apache/datafusion/compare/2818abb8a10bf8b001669c15cf597f9ddfb86ed1..1f5644dcbe5ff6d34c333f16dfca98266e4c62fb) using: tpcds
Results will be posted here when complete
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4253215228
benchmarks are fine, it would be nice to add a test to prevent regression, @RatulDawar WDYT is there a way to add a quick test with small repro you already mentioned in https://github.com/apache/datafusion/issues/21625#issuecomment-4246606398 ?
@adriangb for the #21641 should we also wait for it?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "adriangbot (via GitHub)" <gi...@apache.org>.
adriangbot commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4249632392
๐ค Benchmark completed (GKE) | [trigger](https://github.com/apache/datafusion/pull/21632#issuecomment-4249542586)
**Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
<details><summary>CPU Details (lscpu)</summary>
```
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>
<details><summary>Details</summary>
<p>
```
Comparing HEAD and fix-hash-join-dynamic-filter-deadlock
--------------------
Benchmark tpcds_sf1.json
--------------------
โโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโโโโ
โ Query โ HEAD โ fix-hash-join-dynamic-filter-deadlock โ Change โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ QQuery 1 โ 7.01 / 7.48 ยฑ0.75 / 8.98 ms โ 6.75 / 7.25 ยฑ0.88 / 9.01 ms โ no change โ
โ QQuery 2 โ 144.95 / 145.59 ยฑ0.45 / 146.34 ms โ 144.32 / 146.20 ยฑ1.57 / 148.69 ms โ no change โ
โ QQuery 3 โ 113.50 / 113.99 ยฑ0.37 / 114.63 ms โ 113.13 / 113.80 ยฑ0.41 / 114.31 ms โ no change โ
โ QQuery 4 โ 1372.72 / 1422.25 ยฑ25.68 / 1442.53 ms โ 1282.75 / 1309.27 ยฑ15.96 / 1327.32 ms โ +1.09x faster โ
โ QQuery 5 โ 173.15 / 175.64 ยฑ1.45 / 177.54 ms โ 172.03 / 173.13 ยฑ0.83 / 174.60 ms โ no change โ
โ QQuery 6 โ 840.29 / 849.20 ยฑ7.04 / 860.40 ms โ 833.37 / 857.69 ยฑ20.55 / 889.85 ms โ no change โ
โ QQuery 7 โ 342.82 / 345.82 ยฑ1.98 / 347.96 ms โ 344.32 / 345.76 ยฑ0.89 / 346.85 ms โ no change โ
โ QQuery 8 โ 115.93 / 117.49 ยฑ0.80 / 118.20 ms โ 116.89 / 117.82 ยฑ0.51 / 118.42 ms โ no change โ
โ QQuery 9 โ 101.63 / 106.82 ยฑ2.67 / 109.36 ms โ 101.10 / 105.10 ยฑ2.98 / 108.60 ms โ no change โ
โ QQuery 10 โ 107.65 / 108.67 ยฑ0.85 / 109.59 ms โ 108.10 / 108.34 ยฑ0.20 / 108.60 ms โ no change โ
โ QQuery 11 โ 961.14 / 983.83 ยฑ16.93 / 1009.65 ms โ 894.63 / 919.72 ยฑ18.11 / 945.86 ms โ +1.07x faster โ
โ QQuery 12 โ 47.08 / 47.82 ยฑ1.06 / 49.92 ms โ 45.23 / 46.55 ยฑ1.20 / 48.48 ms โ no change โ
โ QQuery 13 โ 400.24 / 403.81 ยฑ2.97 / 409.18 ms โ 401.04 / 404.69 ยฑ1.96 / 406.36 ms โ no change โ
โ QQuery 14 โ 1012.54 / 1020.46 ยฑ5.04 / 1027.08 ms โ 1001.35 / 1012.76 ยฑ8.64 / 1021.87 ms โ no change โ
โ QQuery 15 โ 16.23 / 17.04 ยฑ0.96 / 18.73 ms โ 15.85 / 16.42 ยฑ0.65 / 17.66 ms โ no change โ
โ QQuery 16 โ 7.31 / 7.81 ยฑ0.58 / 8.93 ms โ 7.23 / 7.59 ยฑ0.27 / 7.97 ms โ no change โ
โ QQuery 17 โ 228.18 / 231.17 ยฑ2.00 / 233.71 ms โ 228.75 / 230.78 ยฑ1.58 / 233.36 ms โ no change โ
โ QQuery 18 โ 127.32 / 128.18 ยฑ0.74 / 129.09 ms โ 128.55 / 129.65 ยฑ0.61 / 130.43 ms โ no change โ
โ QQuery 19 โ 155.42 / 156.06 ยฑ0.68 / 157.22 ms โ 154.08 / 155.61 ยฑ1.21 / 157.81 ms โ no change โ
โ QQuery 20 โ 14.02 / 14.39 ยฑ0.42 / 15.20 ms โ 13.61 / 14.41 ยฑ0.51 / 14.96 ms โ no change โ
โ QQuery 21 โ 19.45 / 19.78 ยฑ0.22 / 20.13 ms โ 19.68 / 20.17 ยฑ0.35 / 20.52 ms โ no change โ
โ QQuery 22 โ 493.32 / 498.14 ยฑ4.44 / 503.55 ms โ 489.61 / 492.50 ยฑ3.17 / 498.14 ms โ no change โ
โ QQuery 23 โ 879.94 / 909.00 ยฑ21.77 / 943.54 ms โ 873.45 / 884.47 ยฑ6.65 / 893.47 ms โ no change โ
โ QQuery 24 โ 389.52 / 393.41 ยฑ3.06 / 398.82 ms โ 381.35 / 385.36 ยฑ2.76 / 388.85 ms โ no change โ
โ QQuery 25 โ 344.76 / 346.18 ยฑ1.34 / 348.09 ms โ 339.19 / 343.42 ยฑ2.43 / 345.95 ms โ no change โ
โ QQuery 26 โ 81.48 / 83.78 ยฑ1.78 / 85.67 ms โ 81.20 / 82.81 ยฑ1.29 / 84.36 ms โ no change โ
โ QQuery 27 โ 7.25 / 7.69 ยฑ0.60 / 8.89 ms โ 7.04 / 7.70 ยฑ0.95 / 9.56 ms โ no change โ
โ QQuery 28 โ 150.23 / 151.57 ยฑ1.80 / 155.08 ms โ 148.28 / 149.34 ยฑ0.92 / 150.70 ms โ no change โ
โ QQuery 29 โ 284.82 / 287.49 ยฑ1.60 / 289.19 ms โ 277.78 / 281.59 ยฑ2.25 / 284.52 ms โ no change โ
โ QQuery 30 โ 44.33 / 45.68 ยฑ1.04 / 47.00 ms โ 43.82 / 45.23 ยฑ1.52 / 47.73 ms โ no change โ
โ QQuery 31 โ 173.39 / 176.19 ยฑ1.84 / 178.39 ms โ 169.21 / 172.29 ยฑ1.89 / 174.78 ms โ no change โ
โ QQuery 32 โ 58.69 / 59.19 ยฑ0.50 / 60.15 ms โ 56.96 / 58.03 ยฑ0.69 / 58.92 ms โ no change โ
โ QQuery 33 โ 143.30 / 145.32 ยฑ1.34 / 147.40 ms โ 141.19 / 142.46 ยฑ0.92 / 143.94 ms โ no change โ
โ QQuery 34 โ 7.33 / 7.58 ยฑ0.26 / 8.07 ms โ 7.00 / 7.47 ยฑ0.39 / 7.93 ms โ no change โ
โ QQuery 35 โ 106.96 / 109.50 ยฑ1.45 / 110.92 ms โ 108.12 / 109.62 ยฑ1.64 / 112.78 ms โ no change โ
โ QQuery 36 โ 6.73 / 7.24 ยฑ0.46 / 8.06 ms โ 6.74 / 6.96 ยฑ0.19 / 7.20 ms โ no change โ
โ QQuery 37 โ 8.75 / 9.01 ยฑ0.25 / 9.48 ms โ 8.43 / 9.19 ยฑ0.77 / 10.67 ms โ no change โ
โ QQuery 38 โ 87.12 / 89.51 ยฑ1.87 / 91.96 ms โ 83.87 / 88.73 ยฑ3.12 / 92.15 ms โ no change โ
โ QQuery 39 โ 127.99 / 130.68 ยฑ2.15 / 134.08 ms โ 127.80 / 130.97 ยฑ1.80 / 133.26 ms โ no change โ
โ QQuery 40 โ 114.51 / 118.70 ยฑ6.05 / 130.36 ms โ 110.74 / 114.82 ยฑ5.29 / 125.28 ms โ no change โ
โ QQuery 41 โ 15.06 / 15.66 ยฑ0.62 / 16.45 ms โ 14.38 / 15.81 ยฑ1.09 / 16.88 ms โ no change โ
โ QQuery 42 โ 107.17 / 109.51 ยฑ1.67 / 112.07 ms โ 106.66 / 108.78 ยฑ1.73 / 110.63 ms โ no change โ
โ QQuery 43 โ 6.22 / 6.35 ยฑ0.11 / 6.54 ms โ 6.25 / 6.36 ยฑ0.15 / 6.66 ms โ no change โ
โ QQuery 44 โ 12.50 / 13.25 ยฑ0.88 / 14.90 ms โ 12.07 / 12.65 ยฑ0.71 / 14.01 ms โ no change โ
โ QQuery 45 โ 51.71 / 53.29 ยฑ1.11 / 54.46 ms โ 51.09 / 52.03 ยฑ0.79 / 53.28 ms โ no change โ
โ QQuery 46 โ 8.93 / 9.33 ยฑ0.46 / 10.22 ms โ 8.79 / 9.17 ยฑ0.31 / 9.73 ms โ no change โ
โ QQuery 47 โ 742.42 / 753.18 ยฑ7.26 / 765.07 ms โ 692.77 / 706.00 ยฑ8.64 / 720.03 ms โ +1.07x faster โ
โ QQuery 48 โ 287.15 / 295.75 ยฑ5.11 / 302.76 ms โ 286.14 / 292.34 ยฑ5.33 / 301.96 ms โ no change โ
โ QQuery 49 โ 255.12 / 257.39 ยฑ2.30 / 261.59 ms โ 250.42 / 253.27 ยฑ2.61 / 257.47 ms โ no change โ
โ QQuery 50 โ 230.23 / 236.05 ยฑ4.03 / 239.71 ms โ 218.99 / 225.34 ยฑ5.08 / 231.41 ms โ no change โ
โ QQuery 51 โ 183.79 / 186.30 ยฑ2.57 / 190.96 ms โ 178.45 / 182.51 ยฑ3.10 / 187.83 ms โ no change โ
โ QQuery 52 โ 107.57 / 109.51 ยฑ1.26 / 111.11 ms โ 106.42 / 108.81 ยฑ1.88 / 111.09 ms โ no change โ
โ QQuery 53 โ 103.49 / 104.19 ยฑ0.58 / 105.25 ms โ 102.89 / 103.47 ยฑ0.47 / 104.25 ms โ no change โ
โ QQuery 54 โ 149.44 / 150.02 ยฑ0.54 / 150.66 ms โ 145.38 / 146.57 ยฑ1.20 / 148.52 ms โ no change โ
โ QQuery 55 โ 107.26 / 109.30 ยฑ1.53 / 111.22 ms โ 106.91 / 108.01 ยฑ0.91 / 109.59 ms โ no change โ
โ QQuery 56 โ 143.78 / 144.54 ยฑ0.60 / 145.33 ms โ 139.67 / 141.80 ยฑ2.33 / 146.14 ms โ no change โ
โ QQuery 57 โ 176.01 / 178.52 ยฑ1.96 / 181.02 ms โ 171.23 / 176.16 ยฑ2.79 / 179.09 ms โ no change โ
โ QQuery 58 โ 292.73 / 301.60 ยฑ5.51 / 308.25 ms โ 289.50 / 299.35 ยฑ6.39 / 306.97 ms โ no change โ
โ QQuery 59 โ 198.44 / 200.95 ยฑ1.73 / 203.23 ms โ 197.65 / 199.30 ยฑ1.33 / 201.28 ms โ no change โ
โ QQuery 60 โ 143.54 / 145.33 ยฑ1.46 / 146.94 ms โ 143.41 / 144.75 ยฑ0.99 / 146.01 ms โ no change โ
โ QQuery 61 โ 13.76 / 14.05 ยฑ0.18 / 14.25 ms โ 13.15 / 13.45 ยฑ0.25 / 13.87 ms โ no change โ
โ QQuery 62 โ 914.30 / 942.48 ยฑ28.63 / 995.49 ms โ 885.28 / 929.18 ยฑ53.81 / 1032.32 ms โ no change โ
โ QQuery 63 โ 104.02 / 107.41 ยฑ3.02 / 112.89 ms โ 101.10 / 103.46 ยฑ1.62 / 104.95 ms โ no change โ
โ QQuery 64 โ 697.69 / 702.85 ยฑ3.51 / 708.11 ms โ 674.15 / 685.89 ยฑ6.30 / 691.53 ms โ no change โ
โ QQuery 65 โ 262.20 / 264.19 ยฑ2.51 / 269.09 ms โ 249.87 / 254.25 ยฑ2.97 / 258.97 ms โ no change โ
โ QQuery 66 โ 248.25 / 253.87 ยฑ4.51 / 260.70 ms โ 238.00 / 246.61 ยฑ8.11 / 256.63 ms โ no change โ
โ QQuery 67 โ 309.10 / 319.22 ยฑ7.07 / 331.03 ms โ 303.33 / 310.32 ยฑ6.30 / 321.27 ms โ no change โ
โ QQuery 68 โ 9.03 / 11.35 ยฑ1.42 / 13.20 ms โ 8.63 / 9.87 ยฑ0.96 / 10.98 ms โ +1.15x faster โ
โ QQuery 69 โ 101.69 / 104.32 ยฑ1.78 / 106.57 ms โ 103.27 / 104.77 ยฑ1.06 / 106.10 ms โ no change โ
โ QQuery 70 โ 328.44 / 341.88 ยฑ10.97 / 361.52 ms โ 331.26 / 359.27 ยฑ17.84 / 381.18 ms โ 1.05x slower โ
โ QQuery 71 โ 135.55 / 137.28 ยฑ1.83 / 140.74 ms โ 133.15 / 134.96 ยฑ1.39 / 136.97 ms โ no change โ
โ QQuery 72 โ 607.26 / 619.12 ยฑ6.13 / 624.86 ms โ 619.75 / 632.34 ยฑ12.08 / 652.53 ms โ no change โ
โ QQuery 73 โ 7.45 / 8.61 ยฑ1.48 / 11.51 ms โ 6.81 / 7.70 ยฑ0.81 / 8.90 ms โ +1.12x faster โ
โ QQuery 74 โ 560.15 / 567.44 ยฑ5.50 / 575.16 ms โ 544.53 / 554.73 ยฑ5.48 / 558.87 ms โ no change โ
โ QQuery 75 โ 275.35 / 277.87 ยฑ1.62 / 280.36 ms โ 277.49 / 278.47 ยฑ0.89 / 279.77 ms โ no change โ
โ QQuery 76 โ 131.21 / 132.26 ยฑ0.94 / 133.71 ms โ 130.71 / 133.04 ยฑ1.57 / 135.02 ms โ no change โ
โ QQuery 77 โ 188.88 / 189.88 ยฑ1.42 / 192.53 ms โ 187.37 / 190.42 ยฑ1.99 / 193.47 ms โ no change โ
โ QQuery 78 โ 339.56 / 342.90 ยฑ2.49 / 346.96 ms โ 342.72 / 346.00 ยฑ3.08 / 350.52 ms โ no change โ
โ QQuery 79 โ 229.72 / 231.37 ยฑ1.26 / 233.42 ms โ 235.30 / 237.74 ยฑ1.46 / 239.29 ms โ no change โ
โ QQuery 80 โ 323.14 / 324.30 ยฑ1.07 / 325.75 ms โ 322.13 / 324.43 ยฑ1.84 / 327.69 ms โ no change โ
โ QQuery 81 โ 26.89 / 27.49 ยฑ0.83 / 29.04 ms โ 26.08 / 27.31 ยฑ1.12 / 29.15 ms โ no change โ
โ QQuery 82 โ 196.43 / 200.54 ยฑ3.31 / 205.91 ms โ 196.70 / 197.80 ยฑ1.70 / 201.20 ms โ no change โ
โ QQuery 83 โ 38.95 / 41.03 ยฑ2.05 / 44.94 ms โ 38.32 / 39.74 ยฑ0.87 / 40.68 ms โ no change โ
โ QQuery 84 โ 49.03 / 49.93 ยฑ0.72 / 51.17 ms โ 48.09 / 49.30 ยฑ0.97 / 50.43 ms โ no change โ
โ QQuery 85 โ 148.60 / 149.77 ยฑ0.81 / 151.10 ms โ 148.75 / 150.63 ยฑ1.28 / 152.28 ms โ no change โ
โ QQuery 86 โ 40.09 / 40.73 ยฑ0.54 / 41.43 ms โ 37.95 / 39.69 ยฑ1.18 / 41.49 ms โ no change โ
โ QQuery 87 โ 85.74 / 89.41 ยฑ3.14 / 94.86 ms โ 85.74 / 89.28 ยฑ2.15 / 92.01 ms โ no change โ
โ QQuery 88 โ 100.88 / 102.29 ยฑ1.45 / 105.01 ms โ 100.69 / 101.85 ยฑ0.78 / 102.83 ms โ no change โ
โ QQuery 89 โ 118.34 / 120.98 ยฑ1.62 / 123.29 ms โ 117.18 / 119.83 ยฑ1.52 / 121.89 ms โ no change โ
โ QQuery 90 โ 23.05 / 23.92 ยฑ0.57 / 24.76 ms โ 23.59 / 23.97 ยฑ0.33 / 24.49 ms โ no change โ
โ QQuery 91 โ 62.10 / 64.06 ยฑ1.27 / 65.54 ms โ 63.18 / 64.57 ยฑ1.92 / 68.36 ms โ no change โ
โ QQuery 92 โ 57.15 / 57.78 ยฑ0.75 / 59.07 ms โ 57.59 / 58.47 ยฑ0.74 / 59.73 ms โ no change โ
โ QQuery 93 โ 186.95 / 187.93 ยฑ0.73 / 188.71 ms โ 186.44 / 188.09 ยฑ1.00 / 189.52 ms โ no change โ
โ QQuery 94 โ 61.55 / 62.22 ยฑ0.54 / 62.83 ms โ 60.45 / 61.58 ยฑ0.70 / 62.50 ms โ no change โ
โ QQuery 95 โ 127.48 / 128.86 ยฑ1.05 / 130.58 ms โ 129.12 / 130.19 ยฑ0.92 / 131.64 ms โ no change โ
โ QQuery 96 โ 71.12 / 73.15 ยฑ1.14 / 74.61 ms โ 70.16 / 73.37 ยฑ2.03 / 75.36 ms โ no change โ
โ QQuery 97 โ 126.52 / 128.23 ยฑ1.07 / 129.46 ms โ 124.77 / 125.74 ยฑ0.97 / 127.44 ms โ no change โ
โ QQuery 98 โ 154.57 / 155.83 ยฑ1.28 / 158.19 ms โ 155.52 / 158.98 ยฑ2.06 / 161.31 ms โ no change โ
โ QQuery 99 โ 10736.01 / 10752.91 ยฑ12.49 / 10769.04 ms โ 10756.36 / 10853.63 ยฑ79.75 / 10996.85 ms โ no change โ
โโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโณโโโโโโโโโโโโโ
โ Benchmark Summary โ โ
โกโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฉ
โ Total Time (HEAD) โ 31728.88ms โ
โ Total Time (fix-hash-join-dynamic-filter-deadlock) โ 31459.08ms โ
โ Average Time (HEAD) โ 320.49ms โ
โ Average Time (fix-hash-join-dynamic-filter-deadlock) โ 317.77ms โ
โ Queries Faster โ 5 โ
โ Queries Slower โ 1 โ
โ Queries with No Change โ 93 โ
โ Queries with Failure โ 0 โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโ
```
</p>
</details>
<details><summary>Resource Usage</summary>
**tpcds โ base (merge-base)**
| Metric | Value |
|--------|-------|
| Wall time | 159.0s |
| Peak memory | 5.6 GiB |
| Avg memory | 4.6 GiB |
| CPU user | 263.0s |
| CPU sys | 17.1s |
| Peak spill | 0 B |
**tpcds โ branch**
| Metric | Value |
|--------|-------|
| Wall time | 157.6s |
| Peak memory | 5.5 GiB |
| Avg memory | 4.4 GiB |
| CPU user | 261.3s |
| CPU sys | 16.4s |
| Peak spill | 0 B |
</details>
---
[File an issue](https://github.com/adriangb/datafusion-benchmarking/issues) against this benchmark runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "comphead (via GitHub)" <gi...@apache.org>.
comphead commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4253802851
> @comphead for a deadlock situation a failing test with say a minute of timeout should suffice ? I wanted to add a test but I think that timeout based test cases make the testcases flaky.
That was an initial idea, using `nextest`
https://nexte.st/docs/features/slow-tests/#per-test-settings
just set a timeout for a min or 2, it if deadlock happens the timeout will trigger
and the test can be part of `extended` tests
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "mbutrovich (via GitHub)" <gi...@apache.org>.
mbutrovich commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4256828645
> @adriangb appreciate if you can have a second look and we can merge it and backport to 53.1.0
Did the original PR get backported to 53.0? It was a performance improvement not a regression fix.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] fix: prevent hash join deadlock when dynamic filtering is enabled [datafusion]
Posted by "RatulDawar (via GitHub)" <gi...@apache.org>.
RatulDawar commented on PR #21632:
URL: https://github.com/apache/datafusion/pull/21632#issuecomment-4258247989
@Omega359 so in this case would we want to revert the performance improvement or merge the fix ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org