Re: [PR] feat(arrow-row): add MSD radix sort kernel for row-encoded keys [arrow-rs]

Mon, 13 Apr 2026 17:33:00 -0700 


mbutrovich commented on PR #9683:
URL: https://github.com/apache/arrow-rs/pull/9683#issuecomment-4240479489

   It took a [rewrite of 
ExternalSorter](https://github.com/apache/datafusion/pull/21600), but we do see 
the benefit of this sort approach with TPCH SF10:
   
   > ๐Ÿค– Benchmark completed (GKE) | 
[trigger](https://github.com/apache/datafusion/pull/21600#issuecomment-4239957729)
   > 
   > **Instance:** `c4a-highmem-16` (12 vCPU / 65 GiB)
   > 
   > CPU Details (lscpu)
   > Details
   > ```
   > Comparing HEAD and sort_redesign
   > --------------------
   > Benchmark tpch_sf10.json
   > --------------------
   > 
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   > โ”ƒ Query     โ”ƒ                                   HEAD โ”ƒ                     
    sort_redesign โ”ƒ        Change โ”ƒ
   > 
โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   > โ”‚ QQuery 1  โ”‚      368.58 / 370.85 ยฑ2.16 / 373.78 ms โ”‚     370.08 / 370.87 
ยฑ0.69 / 371.91 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 2  โ”‚      473.79 / 489.17 ยฑ9.04 / 502.16 ms โ”‚     503.45 / 511.22 
ยฑ4.26 / 514.75 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 3  โ”‚     657.78 / 680.49 ยฑ20.00 / 714.83 ms โ”‚     554.40 / 567.89 
ยฑ9.15 / 578.92 ms โ”‚ +1.20x faster โ”‚
   > โ”‚ QQuery 4  โ”‚     396.01 / 481.34 ยฑ49.21 / 529.24 ms โ”‚     375.96 / 378.87 
ยฑ1.95 / 381.59 ms โ”‚ +1.27x faster โ”‚
   > โ”‚ QQuery 5  โ”‚  1056.66 / 1107.84 ยฑ29.29 / 1130.19 ms โ”‚ 1077.08 / 1109.96 
ยฑ22.86 / 1140.63 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 6  โ”‚      132.68 / 135.00 ยฑ3.95 / 142.86 ms โ”‚     132.56 / 134.73 
ยฑ3.34 / 141.30 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 7  โ”‚  1521.03 / 1540.06 ยฑ17.25 / 1561.64 ms โ”‚  1599.76 / 1606.38 
ยฑ6.21 / 1616.01 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 8  โ”‚ 1430.43 / 1865.35 ยฑ329.81 / 2147.76 ms โ”‚ 1263.17 / 1322.81 
ยฑ63.15 / 1408.66 ms โ”‚ +1.41x faster โ”‚
   > โ”‚ QQuery 9  โ”‚ 2057.48 / 2255.38 ยฑ105.31 / 2370.24 ms โ”‚ 1910.53 / 1942.67 
ยฑ21.25 / 1975.29 ms โ”‚ +1.16x faster โ”‚
   > โ”‚ QQuery 10 โ”‚     511.13 / 532.41 ยฑ23.71 / 577.56 ms โ”‚     523.11 / 528.92 
ยฑ4.32 / 534.88 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 11 โ”‚      454.61 / 463.82 ยฑ5.46 / 470.53 ms โ”‚    536.45 / 550.25 
ยฑ10.89 / 562.23 ms โ”‚  1.19x slower โ”‚
   > โ”‚ QQuery 12 โ”‚      284.23 / 294.33 ยฑ9.03 / 310.60 ms โ”‚     289.44 / 293.88 
ยฑ2.43 / 296.85 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 13 โ”‚      364.07 / 372.45 ยฑ6.22 / 381.34 ms โ”‚     363.05 / 369.08 
ยฑ3.94 / 374.97 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 14 โ”‚      194.44 / 198.29 ยฑ2.49 / 201.18 ms โ”‚     194.98 / 196.47 
ยฑ1.47 / 199.15 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 15 โ”‚      315.39 / 322.91 ยฑ4.06 / 327.75 ms โ”‚     320.90 / 325.80 
ยฑ4.09 / 333.36 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 16 โ”‚      122.20 / 124.70 ยฑ1.85 / 127.91 ms โ”‚     122.52 / 127.61 
ยฑ6.67 / 140.55 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 17 โ”‚  1570.09 / 1577.29 ยฑ10.02 / 1595.97 ms โ”‚  1437.17 / 1448.07 
ยฑ8.65 / 1458.41 ms โ”‚ +1.09x faster โ”‚
   > โ”‚ QQuery 18 โ”‚  1503.93 / 1547.81 ยฑ27.70 / 1588.49 ms โ”‚ 1511.50 / 1540.04 
ยฑ21.71 / 1576.84 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 19 โ”‚     273.93 / 280.90 ยฑ11.56 / 303.92 ms โ”‚    273.17 / 287.98 
ยฑ25.81 / 339.51 ms โ”‚     no change โ”‚
   > โ”‚ QQuery 20 โ”‚     439.39 / 457.63 ยฑ10.05 / 466.77 ms โ”‚     420.94 / 429.17 
ยฑ7.51 / 439.14 ms โ”‚ +1.07x faster โ”‚
   > โ”‚ QQuery 21 โ”‚ 2986.80 / 3293.47 ยฑ154.96 / 3403.32 ms โ”‚ 2902.96 / 2961.16 
ยฑ38.44 / 3019.92 ms โ”‚ +1.11x faster โ”‚
   > โ”‚ QQuery 22 โ”‚      185.61 / 190.31 ยฑ3.99 / 196.84 ms โ”‚     178.18 / 182.96 
ยฑ5.38 / 192.81 ms โ”‚     no change โ”‚
   > 
โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   > โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ณโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”“
   > โ”ƒ Benchmark Summary            โ”ƒ            โ”ƒ
   > โ”กโ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ•‡โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”ฉ
   > โ”‚ Total Time (HEAD)            โ”‚ 18581.80ms โ”‚
   > โ”‚ Total Time (sort_redesign)   โ”‚ 17186.81ms โ”‚
   > โ”‚ Average Time (HEAD)          โ”‚   844.63ms โ”‚
   > โ”‚ Average Time (sort_redesign) โ”‚   781.22ms โ”‚
   > โ”‚ Queries Faster               โ”‚          7 โ”‚
   > โ”‚ Queries Slower               โ”‚          1 โ”‚
   > โ”‚ Queries with No Change       โ”‚         14 โ”‚
   > โ”‚ Queries with Failure         โ”‚          0 โ”‚
   > โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
   > ```
   > 
   > Resource Usage
   > [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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to