Enable sort optimization on float and half_float #126342
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mayya-sharipova
added
>enhancement
:Search Relevance/Search
elasticsearchmachine
added
the
Team:Search Relevance
|
Pinging @elastic/es-search-relevance (Team:Search Relevance) |
|
Hi @mayya-sharipova, I've created a changelog YAML for you. |
|
Tested on rally nyc_taxis track, where we index Results:
Full results for baseline| Metric | Task | Value | Unit |
|---------------------------------------------------------------:|--------------------------------:|-----------------:|-------:|
| Cumulative indexing time of primary shards | | 85.6696 | min |
| Min cumulative indexing time across primary shards | | 85.6696 | min |
| Median cumulative indexing time across primary shards | | 85.6696 | min |
| Max cumulative indexing time across primary shards | | 85.6696 | min |
| Cumulative indexing throttle time of primary shards | | 0 | min |
| Min cumulative indexing throttle time across primary shards | | 0 | min |
| Median cumulative indexing throttle time across primary shards | | 0 | min |
| Max cumulative indexing throttle time across primary shards | | 0 | min |
| Cumulative merge time of primary shards | | 61.2744 | min |
| Cumulative merge count of primary shards | | 73 | |
| Min cumulative merge time across primary shards | | 61.2744 | min |
| Median cumulative merge time across primary shards | | 61.2744 | min |
| Max cumulative merge time across primary shards | | 61.2744 | min |
| Cumulative merge throttle time of primary shards | | 32.8222 | min |
| Min cumulative merge throttle time across primary shards | | 32.8222 | min |
| Median cumulative merge throttle time across primary shards | | 32.8222 | min |
| Max cumulative merge throttle time across primary shards | | 32.8222 | min |
| Cumulative refresh time of primary shards | | 1.97777 | min |
| Cumulative refresh count of primary shards | | 69 | |
| Min cumulative refresh time across primary shards | | 1.97777 | min |
| Median cumulative refresh time across primary shards | | 1.97777 | min |
| Max cumulative refresh time across primary shards | | 1.97777 | min |
| Cumulative flush time of primary shards | | 2.84968 | min |
| Cumulative flush count of primary shards | | 32 | |
| Min cumulative flush time across primary shards | | 2.84968 | min |
| Median cumulative flush time across primary shards | | 2.84968 | min |
| Max cumulative flush time across primary shards | | 2.84968 | min |
| Total Young Gen GC time | | 43.38 | s |
| Total Young Gen GC count | | 423 | |
| Total Old Gen GC time | | 0 | s |
| Total Old Gen GC count | | 0 | |
| Dataset size | | 26.1902 | GB |
| Store size | | 26.1902 | GB |
| Translog size | | 5.12227e-08 | GB |
| Heap used for segments | | 0 | MB |
| Heap used for doc values | | 0 | MB |
| Heap used for terms | | 0 | MB |
| Heap used for norms | | 0 | MB |
| Heap used for points | | 0 | MB |
| Heap used for stored fields | | 0 | MB |
| Segment count | | 32 | |
| Total Ingest Pipeline count | | 0 | |
| Total Ingest Pipeline time | | 0 | s |
| Total Ingest Pipeline failed | | 0 | |
| Min Throughput | index | 196786 | docs/s |
| Mean Throughput | index | 200442 | docs/s |
| Median Throughput | index | 199493 | docs/s |
| Max Throughput | index | 205806 | docs/s |
| 50th percentile latency | index | 338.342 | ms |
| 90th percentile latency | index | 467.953 | ms |
| 99th percentile latency | index | 2003.41 | ms |
| 99.9th percentile latency | index | 5223.63 | ms |
| 99.99th percentile latency | index | 6114.89 | ms |
| 100th percentile latency | index | 6510.01 | ms |
| 50th percentile service time | index | 338.342 | ms |
| 90th percentile service time | index | 467.953 | ms |
| 99th percentile service time | index | 2003.41 | ms |
| 99.9th percentile service time | index | 5223.63 | ms |
| 99.99th percentile service time | index | 6114.89 | ms |
| 100th percentile service time | index | 6510.01 | ms |
| error rate | index | 0.01 | % |
| Min Throughput | default_no_target | 285.11 | ops/s |
| Mean Throughput | default_no_target | 285.11 | ops/s |
| Median Throughput | default_no_target | 285.11 | ops/s |
| Max Throughput | default_no_target | 285.11 | ops/s |
| 50th percentile latency | default_no_target | 1.95542 | ms |
| 90th percentile latency | default_no_target | 2.40826 | ms |
| 99th percentile latency | default_no_target | 2.67739 | ms |
| 100th percentile latency | default_no_target | 5.47121 | ms |
| 50th percentile service time | default_no_target | 1.95542 | ms |
| 90th percentile service time | default_no_target | 2.40826 | ms |
| 99th percentile service time | default_no_target | 2.67739 | ms |
| 100th percentile service time | default_no_target | 5.47121 | ms |
| error rate | default_no_target | 0 | % |
| Min Throughput | default | 3.02 | ops/s |
| Mean Throughput | default | 3.03 | ops/s |
| Median Throughput | default | 3.03 | ops/s |
| Max Throughput | default | 3.06 | ops/s |
| 50th percentile latency | default | 14.1325 | ms |
| 90th percentile latency | default | 17.3301 | ms |
| 99th percentile latency | default | 19.7128 | ms |
| 100th percentile latency | default | 20.8316 | ms |
| 50th percentile service time | default | 12.164 | ms |
| 90th percentile service time | default | 15.2736 | ms |
| 99th percentile service time | default | 18.3746 | ms |
| 100th percentile service time | default | 19.4802 | ms |
| error rate | default | 0 | % |
| Min Throughput | default_1k | 3.02 | ops/s |
| Mean Throughput | default_1k | 3.03 | ops/s |
| Median Throughput | default_1k | 3.03 | ops/s |
| Max Throughput | default_1k | 3.05 | ops/s |
| 50th percentile latency | default_1k | 22.7621 | ms |
| 90th percentile latency | default_1k | 24.814 | ms |
| 99th percentile latency | default_1k | 32.4233 | ms |
| 100th percentile latency | default_1k | 56.992 | ms |
| 50th percentile service time | default_1k | 20.5098 | ms |
| 90th percentile service time | default_1k | 22.5856 | ms |
| 99th percentile service time | default_1k | 26.0072 | ms |
| 100th percentile service time | default_1k | 29.9932 | ms |
| error rate | default_1k | 0 | % |
| Min Throughput | range | 49.86 | ops/s |
| Mean Throughput | range | 49.86 | ops/s |
| Median Throughput | range | 49.86 | ops/s |
| Max Throughput | range | 49.86 | ops/s |
| 50th percentile latency | range | 15.6114 | ms |
| 90th percentile latency | range | 17.125 | ms |
| 99th percentile latency | range | 21.7086 | ms |
| 100th percentile latency | range | 22.0313 | ms |
| 50th percentile service time | range | 15.6114 | ms |
| 90th percentile service time | range | 17.125 | ms |
| 99th percentile service time | range | 21.7086 | ms |
| 100th percentile service time | range | 22.0313 | ms |
| error rate | range | 0 | % |
| Min Throughput | desc_sort_tip_amount_float | 0.77 | ops/s |
| Mean Throughput | desc_sort_tip_amount_float | 0.81 | ops/s |
| Median Throughput | desc_sort_tip_amount_float | 0.82 | ops/s |
| Max Throughput | desc_sort_tip_amount_float | 0.83 | ops/s |
| 50th percentile latency | desc_sort_tip_amount_float | 1184.63 | ms |
| 90th percentile latency | desc_sort_tip_amount_float | 1195.78 | ms |
| 100th percentile latency | desc_sort_tip_amount_float | 1203.2 | ms |
| 50th percentile service time | desc_sort_tip_amount_float | 1184.63 | ms |
| 90th percentile service time | desc_sort_tip_amount_float | 1195.78 | ms |
| 100th percentile service time | desc_sort_tip_amount_float | 1203.2 | ms |
| error rate | desc_sort_tip_amount_float | 0 | % |
| Min Throughput | asc_sort_tip_amount_float | 0.85 | ops/s |
| Mean Throughput | asc_sort_tip_amount_float | 0.85 | ops/s |
| Median Throughput | asc_sort_tip_amount_float | 0.85 | ops/s |
| Max Throughput | asc_sort_tip_amount_float | 0.85 | ops/s |
| 50th percentile latency | asc_sort_tip_amount_float | 1166.21 | ms |
| 90th percentile latency | asc_sort_tip_amount_float | 1180.06 | ms |
| 100th percentile latency | asc_sort_tip_amount_float | 1194.18 | ms |
| 50th percentile service time | asc_sort_tip_amount_float | 1166.21 | ms |
| 90th percentile service time | asc_sort_tip_amount_float | 1180.06 | ms |
| 100th percentile service time | asc_sort_tip_amount_float | 1194.18 | ms |
| error rate | asc_sort_tip_amount_float | 0 | % |
| Min Throughput | desc_sort_tip_amount_half_float | 0.37 | ops/s |
| Mean Throughput | desc_sort_tip_amount_half_float | 0.38 | ops/s |
| Median Throughput | desc_sort_tip_amount_half_float | 0.38 | ops/s |
| Max Throughput | desc_sort_tip_amount_half_float | 0.39 | ops/s |
| 50th percentile latency | desc_sort_tip_amount_half_float | 2565.13 | ms |
| 90th percentile latency | desc_sort_tip_amount_half_float | 2575.25 | ms |
| 100th percentile latency | desc_sort_tip_amount_half_float | 2613.72 | ms |
| 50th percentile service time | desc_sort_tip_amount_half_float | 2565.13 | ms |
| 90th percentile service time | desc_sort_tip_amount_half_float | 2575.25 | ms |
| 100th percentile service time | desc_sort_tip_amount_half_float | 2613.72 | ms |
| error rate | desc_sort_tip_amount_half_float | 0 | % |
| Min Throughput | asc_sort_tip_amount_half_float | 0.39 | ops/s |
| Mean Throughput | asc_sort_tip_amount_half_float | 0.39 | ops/s |
| Median Throughput | asc_sort_tip_amount_half_float | 0.39 | ops/s |
| Max Throughput | asc_sort_tip_amount_half_float | 0.39 | ops/s |
| 50th percentile latency | asc_sort_tip_amount_half_float | 2569.29 | ms |
| 90th percentile latency | asc_sort_tip_amount_half_float | 2589.36 | ms |
| 100th percentile latency | asc_sort_tip_amount_half_float | 2622.1 | ms |
| 50th percentile service time | asc_sort_tip_amount_half_float | 2569.29 | ms |
| 90th percentile service time | asc_sort_tip_amount_half_float | 2589.36 | ms |
| 100th percentile service time | asc_sort_tip_amount_half_float | 2622.1 | ms |
| error rate | asc_sort_tip_amount_half_float | 0 | % |
| Min Throughput | distance_amount_agg | 0.88 | ops/s |
| Mean Throughput | distance_amount_agg | 0.89 | ops/s |
| Median Throughput | distance_amount_agg | 0.89 | ops/s |
| Max Throughput | distance_amount_agg | 0.89 | ops/s |
| 50th percentile latency | distance_amount_agg | 1105.46 | ms |
| 90th percentile latency | distance_amount_agg | 1176.85 | ms |
| 100th percentile latency | distance_amount_agg | 1287.86 | ms |
| 50th percentile service time | distance_amount_agg | 1105.46 | ms |
| 90th percentile service time | distance_amount_agg | 1176.85 | ms |
| 100th percentile service time | distance_amount_agg | 1287.86 | ms |
| error rate | distance_amount_agg | 0 | % |
| Min Throughput | autohisto_agg | 12.54 | ops/s |
| Mean Throughput | autohisto_agg | 13.23 | ops/s |
| Median Throughput | autohisto_agg | 13.31 | ops/s |
| Max Throughput | autohisto_agg | 13.72 | ops/s |
| 50th percentile latency | autohisto_agg | 66.8652 | ms |
| 90th percentile latency | autohisto_agg | 67.3299 | ms |
| 99th percentile latency | autohisto_agg | 69.9935 | ms |
| 100th percentile latency | autohisto_agg | 70.5453 | ms |
| 50th percentile service time | autohisto_agg | 66.8652 | ms |
| 90th percentile service time | autohisto_agg | 67.3299 | ms |
| 99th percentile service time | autohisto_agg | 69.9935 | ms |
| 100th percentile service time | autohisto_agg | 70.5453 | ms |
| error rate | autohisto_agg | 0 | % |
| Min Throughput | date_histogram_agg | 239.68 | ops/s |
| Mean Throughput | date_histogram_agg | 240.83 | ops/s |
| Median Throughput | date_histogram_agg | 240.83 | ops/s |
| Max Throughput | date_histogram_agg | 241.99 | ops/s |
| 50th percentile latency | date_histogram_agg | 3.4941 | ms |
| 90th percentile latency | date_histogram_agg | 5.62715 | ms |
| 99th percentile latency | date_histogram_agg | 11.4177 | ms |
| 100th percentile latency | date_histogram_agg | 119.41 | ms |
| 50th percentile service time | date_histogram_agg | 3.4941 | ms |
| 90th percentile service time | date_histogram_agg | 5.62715 | ms |
| 99th percentile service time | date_histogram_agg | 11.4177 | ms |
| 100th percentile service time | date_histogram_agg | 119.41 | ms |
| error rate | date_histogram_agg | 0 | % |
Full results for candidate| Metric | Task | Value | Unit |
|---------------------------------------------------------------:|--------------------------------:|-----------------:|-------:|
| Cumulative indexing time of primary shards | | 83.5571 | min |
| Min cumulative indexing time across primary shards | | 83.5571 | min |
| Median cumulative indexing time across primary shards | | 83.5571 | min |
| Max cumulative indexing time across primary shards | | 83.5571 | min |
| Cumulative indexing throttle time of primary shards | | 0 | min |
| Min cumulative indexing throttle time across primary shards | | 0 | min |
| Median cumulative indexing throttle time across primary shards | | 0 | min |
| Max cumulative indexing throttle time across primary shards | | 0 | min |
| Cumulative merge time of primary shards | | 55.4796 | min |
| Cumulative merge count of primary shards | | 76 | |
| Min cumulative merge time across primary shards | | 55.4796 | min |
| Median cumulative merge time across primary shards | | 55.4796 | min |
| Max cumulative merge time across primary shards | | 55.4796 | min |
| Cumulative merge throttle time of primary shards | | 29.3164 | min |
| Min cumulative merge throttle time across primary shards | | 29.3164 | min |
| Median cumulative merge throttle time across primary shards | | 29.3164 | min |
| Max cumulative merge throttle time across primary shards | | 29.3164 | min |
| Cumulative refresh time of primary shards | | 1.6677 | min |
| Cumulative refresh count of primary shards | | 70 | |
| Min cumulative refresh time across primary shards | | 1.6677 | min |
| Median cumulative refresh time across primary shards | | 1.6677 | min |
| Max cumulative refresh time across primary shards | | 1.6677 | min |
| Cumulative flush time of primary shards | | 2.98622 | min |
| Cumulative flush count of primary shards | | 34 | |
| Min cumulative flush time across primary shards | | 2.98622 | min |
| Median cumulative flush time across primary shards | | 2.98622 | min |
| Max cumulative flush time across primary shards | | 2.98622 | min |
| Total Young Gen GC time | | 38.445 | s |
| Total Young Gen GC count | | 389 | |
| Total Old Gen GC time | | 0 | s |
| Total Old Gen GC count | | 0 | |
| Dataset size | | 25.9803 | GB |
| Store size | | 25.9803 | GB |
| Translog size | | 5.12227e-08 | GB |
| Heap used for segments | | 0 | MB |
| Heap used for doc values | | 0 | MB |
| Heap used for terms | | 0 | MB |
| Heap used for norms | | 0 | MB |
| Heap used for points | | 0 | MB |
| Heap used for stored fields | | 0 | MB |
| Segment count | | 35 | |
| Total Ingest Pipeline count | | 0 | |
| Total Ingest Pipeline time | | 0 | s |
| Total Ingest Pipeline failed | | 0 | |
| Min Throughput | index | 201096 | docs/s |
| Mean Throughput | index | 209905 | docs/s |
| Median Throughput | index | 207515 | docs/s |
| Max Throughput | index | 228610 | docs/s |
| 50th percentile latency | index | 340.992 | ms |
| 90th percentile latency | index | 478.772 | ms |
| 99th percentile latency | index | 1929.45 | ms |
| 99.9th percentile latency | index | 5484.74 | ms |
| 99.99th percentile latency | index | 7569.14 | ms |
| 100th percentile latency | index | 7844.92 | ms |
| 50th percentile service time | index | 340.992 | ms |
| 90th percentile service time | index | 478.772 | ms |
| 99th percentile service time | index | 1929.45 | ms |
| 99.9th percentile service time | index | 5484.74 | ms |
| 99.99th percentile service time | index | 7569.14 | ms |
| 100th percentile service time | index | 7844.92 | ms |
| error rate | index | 0.01 | % |
| Min Throughput | default_no_target | 319.98 | ops/s |
| Mean Throughput | default_no_target | 319.98 | ops/s |
| Median Throughput | default_no_target | 319.98 | ops/s |
| Max Throughput | default_no_target | 319.98 | ops/s |
| 50th percentile latency | default_no_target | 1.79496 | ms |
| 90th percentile latency | default_no_target | 2.2277 | ms |
| 99th percentile latency | default_no_target | 2.73558 | ms |
| 100th percentile latency | default_no_target | 6.80229 | ms |
| 50th percentile service time | default_no_target | 1.79496 | ms |
| 90th percentile service time | default_no_target | 2.2277 | ms |
| 99th percentile service time | default_no_target | 2.73558 | ms |
| 100th percentile service time | default_no_target | 6.80229 | ms |
| error rate | default_no_target | 0 | % |
| Min Throughput | default | 3.02 | ops/s |
| Mean Throughput | default | 3.03 | ops/s |
| Median Throughput | default | 3.03 | ops/s |
| Max Throughput | default | 3.06 | ops/s |
| 50th percentile latency | default | 11.9339 | ms |
| 90th percentile latency | default | 14.1389 | ms |
| 99th percentile latency | default | 16.1274 | ms |
| 100th percentile latency | default | 18.8175 | ms |
| 50th percentile service time | default | 10.0321 | ms |
| 90th percentile service time | default | 11.9062 | ms |
| 99th percentile service time | default | 14.1333 | ms |
| 100th percentile service time | default | 16.4925 | ms |
| error rate | default | 0 | % |
| Min Throughput | default_1k | 3.02 | ops/s |
| Mean Throughput | default_1k | 3.03 | ops/s |
| Median Throughput | default_1k | 3.03 | ops/s |
| Max Throughput | default_1k | 3.05 | ops/s |
| 50th percentile latency | default_1k | 19.3333 | ms |
| 90th percentile latency | default_1k | 21.428 | ms |
| 99th percentile latency | default_1k | 22.6266 | ms |
| 100th percentile latency | default_1k | 42.182 | ms |
| 50th percentile service time | default_1k | 17.1181 | ms |
| 90th percentile service time | default_1k | 19.2999 | ms |
| 99th percentile service time | default_1k | 19.93 | ms |
| 100th percentile service time | default_1k | 20.4645 | ms |
| error rate | default_1k | 0 | % |
| Min Throughput | range | 54.13 | ops/s |
| Mean Throughput | range | 54.13 | ops/s |
| Median Throughput | range | 54.13 | ops/s |
| Max Throughput | range | 54.13 | ops/s |
| 50th percentile latency | range | 15.0456 | ms |
| 90th percentile latency | range | 17.1059 | ms |
| 99th percentile latency | range | 18.237 | ms |
| 100th percentile latency | range | 18.2661 | ms |
| 50th percentile service time | range | 15.0456 | ms |
| 90th percentile service time | range | 17.1059 | ms |
| 99th percentile service time | range | 18.237 | ms |
| 100th percentile service time | range | 18.2661 | ms |
| error rate | range | 0 | % |
| Min Throughput | desc_sort_tip_amount_float | 46.78 | ops/s |
| Mean Throughput | desc_sort_tip_amount_float | 46.78 | ops/s |
| Median Throughput | desc_sort_tip_amount_float | 46.78 | ops/s |
| Max Throughput | desc_sort_tip_amount_float | 46.78 | ops/s |
| 50th percentile latency | desc_sort_tip_amount_float | 11.31 | ms |
| 90th percentile latency | desc_sort_tip_amount_float | 11.9369 | ms |
| 100th percentile latency | desc_sort_tip_amount_float | 12.7691 | ms |
| 50th percentile service time | desc_sort_tip_amount_float | 11.31 | ms |
| 90th percentile service time | desc_sort_tip_amount_float | 11.9369 | ms |
| 100th percentile service time | desc_sort_tip_amount_float | 12.7691 | ms |
| error rate | desc_sort_tip_amount_float | 0 | % |
| Min Throughput | asc_sort_tip_amount_float | 248.45 | ops/s |
| Mean Throughput | asc_sort_tip_amount_float | 248.45 | ops/s |
| Median Throughput | asc_sort_tip_amount_float | 248.45 | ops/s |
| Max Throughput | asc_sort_tip_amount_float | 248.45 | ops/s |
| 50th percentile latency | asc_sort_tip_amount_float | 2.05563 | ms |
| 90th percentile latency | asc_sort_tip_amount_float | 2.42422 | ms |
| 100th percentile latency | asc_sort_tip_amount_float | 3.50667 | ms |
| 50th percentile service time | asc_sort_tip_amount_float | 2.05563 | ms |
| 90th percentile service time | asc_sort_tip_amount_float | 2.42422 | ms |
| 100th percentile service time | asc_sort_tip_amount_float | 3.50667 | ms |
| error rate | asc_sort_tip_amount_float | 0 | % |
| Min Throughput | desc_sort_tip_amount_half_float | 59.87 | ops/s |
| Mean Throughput | desc_sort_tip_amount_half_float | 59.87 | ops/s |
| Median Throughput | desc_sort_tip_amount_half_float | 59.87 | ops/s |
| Max Throughput | desc_sort_tip_amount_half_float | 59.87 | ops/s |
| 50th percentile latency | desc_sort_tip_amount_half_float | 10.8056 | ms |
| 90th percentile latency | desc_sort_tip_amount_half_float | 12.7851 | ms |
| 100th percentile latency | desc_sort_tip_amount_half_float | 17.7602 | ms |
| 50th percentile service time | desc_sort_tip_amount_half_float | 10.8056 | ms |
| 90th percentile service time | desc_sort_tip_amount_half_float | 12.7851 | ms |
| 100th percentile service time | desc_sort_tip_amount_half_float | 17.7602 | ms |
| error rate | desc_sort_tip_amount_half_float | 0 | % |
| Min Throughput | asc_sort_tip_amount_half_float | 290.69 | ops/s |
| Mean Throughput | asc_sort_tip_amount_half_float | 290.69 | ops/s |
| Median Throughput | asc_sort_tip_amount_half_float | 290.69 | ops/s |
| Max Throughput | asc_sort_tip_amount_half_float | 290.69 | ops/s |
| 50th percentile latency | asc_sort_tip_amount_half_float | 1.97798 | ms |
| 90th percentile latency | asc_sort_tip_amount_half_float | 2.20325 | ms |
| 100th percentile latency | asc_sort_tip_amount_half_float | 2.3505 | ms |
| 50th percentile service time | asc_sort_tip_amount_half_float | 1.97798 | ms |
| 90th percentile service time | asc_sort_tip_amount_half_float | 2.20325 | ms |
| 100th percentile service time | asc_sort_tip_amount_half_float | 2.3505 | ms |
| error rate | asc_sort_tip_amount_half_float | 0 | % |
| Min Throughput | distance_amount_agg | 0.95 | ops/s |
| Mean Throughput | distance_amount_agg | 0.95 | ops/s |
| Median Throughput | distance_amount_agg | 0.95 | ops/s |
| Max Throughput | distance_amount_agg | 0.96 | ops/s |
| 50th percentile latency | distance_amount_agg | 1032.4 | ms |
| 90th percentile latency | distance_amount_agg | 1056.36 | ms |
| 100th percentile latency | distance_amount_agg | 1077.32 | ms |
| 50th percentile service time | distance_amount_agg | 1032.4 | ms |
| 90th percentile service time | distance_amount_agg | 1056.36 | ms |
| 100th percentile service time | distance_amount_agg | 1077.32 | ms |
| error rate | distance_amount_agg | 0 | % |
| Min Throughput | autohisto_agg | 12.13 | ops/s |
| Mean Throughput | autohisto_agg | 13.04 | ops/s |
| Median Throughput | autohisto_agg | 13.16 | ops/s |
| Max Throughput | autohisto_agg | 13.64 | ops/s |
| 50th percentile latency | autohisto_agg | 66.1679 | ms |
| 90th percentile latency | autohisto_agg | 70.9618 | ms |
| 99th percentile latency | autohisto_agg | 78.2489 | ms |
| 100th percentile latency | autohisto_agg | 81.4795 | ms |
| 50th percentile service time | autohisto_agg | 66.1679 | ms |
| 90th percentile service time | autohisto_agg | 70.9618 | ms |
| 99th percentile service time | autohisto_agg | 78.2489 | ms |
| 100th percentile service time | autohisto_agg | 81.4795 | ms |
| error rate | autohisto_agg | 0 | % |
| Min Throughput | date_histogram_agg | 165.42 | ops/s |
| Mean Throughput | date_histogram_agg | 168.77 | ops/s |
| Median Throughput | date_histogram_agg | 168.77 | ops/s |
| Max Throughput | date_histogram_agg | 172.13 | ops/s |
| 50th percentile latency | date_histogram_agg | 4.43994 | ms |
| 90th percentile latency | date_histogram_agg | 7.24175 | ms |
| 99th percentile latency | date_histogram_agg | 11.2178 | ms |
| 100th percentile latency | date_histogram_agg | 19.3952 | ms |
| 50th percentile service time | date_histogram_agg | 4.43994 | ms |
| 90th percentile service time | date_histogram_agg | 7.24175 | ms |
| 99th percentile service time | date_histogram_agg | 11.2178 | ms |
| 100th percentile service time | date_histogram_agg | 19.3952 | ms |
| error rate | date_histogram_agg | 0 | % | |
john-wagster
reviewed
Apr 7, 2025
| @@ -199,7 +200,8 @@ private XFieldComparatorSource comparatorSource( | |||
| Nested nested | |||
| ) { | |||
| return switch (targetNumericType) { | |||
| case HALF_FLOAT, FLOAT -> new FloatValuesComparatorSource(this, missingValue, sortMode, nested); | |||
There was a problem hiding this comment.
more a question than feedback. Since HALF_FLOAT is no longer handled by FloatValuesComparatorSource are there missing unit tests that are needed. I noticed FloatValuesComparatorSource is referenced by FloatNestedSortingTests. Does it make sense to have a HalfFloatNestedSortingTests?
There was a problem hiding this comment.
@john-wagster Thanks for the feedback, addressed in ee021d7
john-wagster
approved these changes
Apr 7, 2025
benwtrent
approved these changes
Apr 8, 2025
| @@ -524,6 +524,9 @@ public void testSimpleSorts() throws Exception { | |||
| .startObject("float_value") | |||
| .field("type", "float") | |||
| .endObject() | |||
| .startObject("half_float_value") | |||
There was a problem hiding this comment.
my initial concern was if this integration test would ever be multi-node. This way we have multiple data nodes and thus might sort again on the coordinator. But it looks like this is randomly multiple data nodes. So I think we are good on coverage there.
jimczi
approved these changes
Apr 8, 2025
mayya-sharipova
added
the
auto-backport
💔 Backport failed
You can use sqren/backport to manually backport by running |
elasticsearchmachine
pushed a commit
that referenced
this pull request
Apr 9, 2025
* Enable sort optimization on float and half_float (#126342) Before we enabled sort optimization on long, double and date types, but left other types for follow-up. This enables sort optimization on float and half_float types. Optimizations on INT, BYTE, SHORT are left for follow-up, because they need more work: we currently use SORT.LONG type for all integer types and this doesn't allow to use optimization. Backport for #126342 * Modification
mayya-sharipova
added a commit
to mayya-sharipova/elasticsearch
that referenced
this pull request
Apr 9, 2025
Before we enabled sort optimization on long, double and date types, but left other types for follow-up. This enables sort optimization on float and half_float types. Optimizations on INT, BYTE, SHORT are left for follow-up, because they need more work: we currently use SORT.LONG type for all integer types and this doesn't allow to use optimization.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Apr 9, 2025
Before we enabled sort optimization on long, double and date types, but left other types for follow-up. This enables sort optimization on float and half_float types. Optimizations on INT, BYTE, SHORT are left for follow-up, because they need more work: we currently use SORT.LONG type for all integer types and this doesn't allow to use optimization.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Before we enabled sort optimization on long, double and date types,
but left other types for follow-up.
This enables sort optimization on float and half_float types.
Optimizations on INT, BYTE, SHORT are left for follow-up,
because they need more work: we currently use SORT.LONG
type for all integer types and this doesn't allow to use
optimization.