This tries to benchmark a updated implementation of Prebid Util functions, specifically deepEqual and mergeDeep.
This is a micro-benchmark, and that comes with caveats (especially because different data will give different results).
The data used to test is in src/data.ts.
src/old-util.js contain the current versions of the functions, and src/new-util.js has the new one.
- Node
- NPM
I use proto to manage my Node/NPM versions.
To get the same version that I am using, look into the .prototools file.
npm inpm run startnpm run start:markdown- generates the benchmark results and a summary as markdown in the dist folder
| Task name | Latency Improvement (%) | Throughput Improvement (%) |
|---|---|---|
| deepEqual - array | -9.23% improvement | +8.50% improvement |
| deepEqual - deepNested | -9.08% improvement | +11.59% improvement |
| deepEqual - large | -5.92% improvement | +8.42% improvement |
| deepEqual - nested | -10.85% improvement | +13.85% improvement |
| deepEqual - simple | -18.99% improvement | +17.45% improvement |
| deepEqual - differentTypes | -9.44% improvement | +10.17% improvement |
| Task name | Latency Improvement (%) | Throughput Improvement (%) |
|---|---|---|
| mergeDeep - array | -40.10% improvement | +83.07% improvement |
| mergeDeep - deepNested | -7.84% improvement | +8.12% improvement |
| mergeDeep - large | -80.00% improvement | +395.95% improvement |
| mergeDeep - nested | -46.51% improvement | +85.18% improvement |
| mergeDeep - simple | -58.41% improvement | +132.15% improvement |
| mergeDeep - differentTypes | -56.36% improvement | +138.65% improvement |
| Task name | Latency average (ns) | Latency median (ns) | Throughput average (ops/s) | Throughput median (ops/s) | Samples |
|---|---|---|---|---|---|
| deepEqualOld - array | 389.94 ± 0.40% | 375.00 | 2724953 ± 0.01% | 2666667 | 2564480 |
| deepEqual - array | 353.95 ± 0.33% | 333.00 | 2956457 ± 0.01% | 3003003 | 2825294 |
| deepEqualOld - deepNested | 307.19 ± 1.05% | 292.00 | 3434427 ± 0.01% | 3424658 | 3255357 |
| deepEqual - deepNested | 279.29 ± 0.36% | 250.00 | 3832571 ± 0.01% | 4000000 | 3580542 |
| deepEqualOld - large | 59214.79 ± 0.60% | 55750.00 | 17371 ± 0.15% | 17937 | 16888 |
| deepEqual - large | 55710.91 ± 0.89% | 51541.00 | 18834 ± 0.16% | 19402 | 17950 |
| deepEqualOld - nested | 280.73 ± 0.46% | 250.00 | 3777229 ± 0.01% | 4000000 | 3562130 |
| deepEqual - nested | 250.28 ± 0.47% | 250.00 | 4300209 ± 0.01% | 4000000 | 3995558 |
| deepEqualOld - simple | 149.33 ± 2.40% | 125.00 | 7661044 ± 0.01% | 8000000 | 6696466 |
| deepEqual - simple | 120.98 ± 0.40% | 125.00 | 8998020 ± 0.01% | 8000000 | 8265833 |
| deepEqualOld - differentTypes | 337.18 ± 0.28% | 333.00 | 3063344 ± 0.01% | 3003003 | 2965802 |
| deepEqual - differentTypes | 305.35 ± 0.18% | 292.00 | 3374887 ± 0.01% | 3424658 | 3274944 |
| Task name | Latency average (ns) | Latency median (ns) | Throughput average (ops/s) | Throughput median (ops/s) | Samples |
|---|---|---|---|---|---|
| mergeDeepOld - array | 540.44 ± 0.59% | 500.00 | 1972887 ± 0.01% | 2000000 | 1850578 |
| mergeDeep - array | 323.70 ± 1.18% | 291.00 | 3611748 ± 0.01% | 3436426 | 3089913 |
| mergeDeepOld - deepNested | 1705.74 ± 0.34% | 1625.00 | 610094 ± 0.02% | 615385 | 586256 |
| mergeDeep - deepNested | 1572.03 ± 0.35% | 1500.00 | 659664 ± 0.02% | 666667 | 636119 |
| mergeDeepOld - large | 464709.40 ± 1.25% | 439958.00 | 2204 ± 0.41% | 2273 | 2152 |
| mergeDeep - large | 92962.00 ± 0.41% | 89000.00 | 10931 ± 0.16% | 11236 | 10758 |
| mergeDeepOld - nested | 1959.30 ± 0.47% | 1875.00 | 528007 ± 0.02% | 533333 | 510387 |
| mergeDeep - nested | 1048.11 ± 0.31% | 1000.00 | 977743 ± 0.01% | 1000000 | 954098 |
| mergeDeepOld - simple | 863.15 ± 3.61% | 791.00 | 1274121 ± 0.02% | 1264223 | 1158542 |
| mergeDeep - simple | 358.97 ± 0.60% | 333.00 | 2957820 ± 0.01% | 3003003 | 2785757 |
| mergeDeepOld - differentTypes | 913.25 ± 0.89% | 834.00 | 1167189 ± 0.01% | 1199041 | 1094986 |
| mergeDeep - differentTypes | 398.50 ± 1.01% | 375.00 | 2785454 ± 0.01% | 2666667 | 2509735 |