atlas-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Pierre Padovani <pierre.padov...@civitaslearning.com>
Subject Re: Performance of relationships
Date Mon, 04 Jun 2018 13:52:31 GMT
Here is the raw data:

child count v2 post v2 get container v1 post v1 get container v2 hybrid post v2
hybrid get
0 210 85 96 478 487 185
1 188 83 129 140 213 75
2 177 83 209 357 188 118
3 162 78 117 72 387 323
4 188 92 324 71 360 169
5 166 80 453 100 314 299
6 188 87 364 79 332 154
7 152 69 174 341 282 244
8 168 67 252 318 184 172
9 201 88 404 102 192 235
10 144 88 169 253 313 380
11 234 81 261 419 270 185
12 177 67 366 99 197 259
13 146 76 167 259 176 270
14 179 87 187 80 267 352
15 273 105 255 82 161 276
16 159 64 199 246 176 291
17 173 84 190 110 125 389
18 187 99 422 194 171 324
19 158 84 320 389 192 442
20 179 94 339 92 225 410
21 193 211 221 89 185 301
22 233 77 375 97 166 284
23 247 103 244 124 190 463
24 172 159 459 119 150 256
25 181 102 385 85 279 443
26 188 93 290 99 137 326
27 169 88 577 277 171 423
28 192 82 345 143 121 355
29 253 88 497 190 205 406
30 232 81 287 288 119 396
31 178 89 393 263 219 441
32 208 75 400 166 148 528
33 147 83 366 152 160 451
34 179 98 322 193 121 452
35 166 77 331 184 126 464
36 348 69 526 161 140 468
37 164 85 322 164 121 443
38 166 93 349 162 128 534
39 184 83 504 152 152 459
40 179 234 666 156 179 469
41 188 86 661 154 138 558
42 180 72 666 157 114 461
43 171 87 663 184 127 557
44 189 91 532 211 147 474
45 202 93 712 156 140 464
46 341 79 662 162 103 541
47 206 101 658 184 102 535
48 153 85 634 173 149 552
49 374 164 648 187 141 674
50 193 72 638 169 142 568
51 149 92 643 177 137 681
52 213 105 642 167 176 580
53 195 127 653 240 121 498
54 152 74 579 235 109 592
55 165 81 584 223 150 616
56 190 107 595 241 113 586
57 375 74 667 268 137 560
58 256 128 701 248 151 621
59 140 81 624 235 141 580
60 164 92 701 236 133 599
61 181 110 991 252 186 646
62 396 88 980 190 129 617
63 220 147 630 202 158 690
64 164 78 920 232 215 847
65 230 196 652 241 147 580
66 192 89 27 253 141 759
67 230 152 975 212 119 641
68 149 90 14 196 132 627
69 167 231 866 243 142 647
70 160 98 153 247 116 666
71 204 137 980 278 150 679
72 142 81 791 268 168 648
73 209 165 122 241 147 703
74 172 77 939 194 132 595
75 145 80 81 209 154 690
76 134 91 25 261 119 913
77 151 77 962 241 129 799
78 139 75 54 261 101 746
79 151 95 904 453 121 672
80 131 78 184 252 150 741
81 142 82 26 268 109 745
82 170 74 963 261 165 658
83 157 78 970 263 191 778
84 329 78 320 309 112 756
85 136 78 326 314 114 734
86 140 89 324 306 136 722
87 144 186 339 321 117 671
88 164 76 316 307 110 792
89 154 81 326 327 127 820
90 143 77 315 267 114 743
91 156 94 19 272 137 772
92 137 184 152 312 116 868
93 214 136 127 329 83 815
94 136 78 301 324 114 819
95 131 77 158 342 116 741
96 144 77 210 331 155 824
97 144 84 556 324 95 862
98 191 134 257 284 120 862
99 143 80 492 340 117 879

On Fri, Jun 1, 2018 at 3:48 PM, Apoorv Naik <anaik@hortonworks.com> wrote:

> Can you share the text of the benchmarks you performed ? My email is not
> rendering the attached image.
>
> From: Pierre Padovani <pierre.padovani@civitaslearning.com>
> Reply-To: "dev@atlas.apache.org" <dev@atlas.apache.org>
> Date: Friday, June 1, 2018 at 12:49 PM
> To: "dev@atlas.apache.org" <dev@atlas.apache.org>
> Subject: Re: Performance of relationships
>
> I was in a hurry writing the previous email and realized that I did not
> specify that the relationship type was COMPOSITION.
>
> On Fri, Jun 1, 2018 at 2:37 PM, Pierre Padovani <pierre.padovani@
> civitaslearning.com<mailto:pierre.padovani@civitaslearning.com>> wrote:
> All,
>
> I decided to spend a few cycles performing from basic performance tests to
> help us refine our type system. The goal was to determine how we would add
> relationships to our entities and if we would use the new relationship APIs
> or not. The test was simple. I created a type called container and a type
> called child both extending DataSet. I think created an instance of
> container, and added one child at a time to the container recording the
> cost of the add, and the cost of fetching the container after each add.
>
> There were three tests run:
>
> - Legacy - Assign the entities to the attributes in the attribute map
> - Hybrid - Leave the legacy attributes in the type system, and use the
> relationship APIs
> - V2 - Remove the legacy attributes and use the relationship APIs
>
> Here are the results for 100 children:
>
> [cid:ii_jhwdc79w0_163bcd934880f531]
>
> Is there a reason we are continuing to use legacy attributes in the type
> system(s) we are shipping with Atlas? From a performance perspective this
> does not seem like a reasonable thing to do.
>
> Thanks!
>
> Pierre
>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message