use pre calculated utf16 table

[Uzlopak]

If #25 gets merged, we could compare it. It seems to double the encodeString performance for utf16 characters.

[anonrig]

Nice! I really liked the idea of pre-calculating utf-16 values. I'll merge this after the previous pull request is merged. GJ @Uzlopak

[anonrig]

@Uzlopak Can you update the branch and run a benchmark?

[Uzlopak]

I ran the Benchmarks. Basically a gain of +15 % for multi byte characters.

uzlopak@Battlestation:~/workspace/faster-querystring$ node benchmark/bench-cmp-branch.js 
Select the branch you want to compare (feature branch):
encodeString-fast
Select the branch you want to compare with (main branch):
main
Checking out "encodeString-fast"
Execute "npm run benchmark"

> fast-querystring@1.0.0 benchmark
> node benchmark/bench.js

encodeString: ""                                              x 301,800,480 ops/sec ±0.39% (193 runs sampled)
encodeString: "123"                                           x 124,991,538 ops/sec ±0.15% (194 runs sampled)
encodeString: "ä"                                             x 149,392,191 ops/sec ±0.16% (194 runs sampled)
encodeString: "𝌆"                                            x 33,263,591 ops/sec ±0.33% (191 runs sampled)
stringify: undefined                                       x 1,252,964,447 ops/sec ±0.12% (195 runs sampled)
stringify: null                                            x 1,256,352,113 ops/sec ±0.07% (194 runs sampled)
stringify: {}                                              x 123,117,277 ops/sec ±0.65% (182 runs sampled)
stringify: { id: true }                                    x 35,804,756 ops/sec ±0.62% (187 runs sampled)
stringify: { id: false }                                   x 37,444,187 ops/sec ±0.22% (194 runs sampled)
stringify: { id: 123 }                                     x 29,385,111 ops/sec ±0.62% (187 runs sampled)
stringify: { id: 1e+22 }                                   x 12,593,301 ops/sec ±0.52% (191 runs sampled)
stringify: { id: 123n }                                    x 19,516,558 ops/sec ±1.19% (186 runs sampled)
stringify: { id: Infinity }                                x 45,336,574 ops/sec ±0.60% (190 runs sampled)
stringify: { id: ["1", "3"] }                              x 16,322,006 ops/sec ±0.48% (187 runs sampled)
stringify: { id: "" }                                      x 39,664,504 ops/sec ±0.72% (187 runs sampled)
stringify: { id: "123" }                                   x 28,509,920 ops/sec ±0.47% (190 runs sampled)
stringify: { id: "ä" }                                     x 26,778,841 ops/sec ±0.67% (191 runs sampled)
stringify: { id: "𝌆" }                                    x 17,288,323 ops/sec ±0.63% (186 runs sampled)
stringify: { foo: ["1", "3"], bar: "2" }                   x 9,261,218 ops/sec ±0.24% (193 runs sampled)
parse:                                                 x 159,142,387 ops/sec ±0.16% (194 runs sampled)
parse: id=123                                          x 16,084,981 ops/sec ±0.66% (188 runs sampled)
parse: id=123&id=123                                   x 8,123,288 ops/sec ±0.32% (190 runs sampled)
parse: full%20name=Yagiz                               x 6,064,912 ops/sec ±0.26% (192 runs sampled)
parse: invalid%key=hello                               x 6,567,817 ops/sec ±0.30% (189 runs sampled)
parse: my+weird+field=q1%212%22%27w%245%267%2Fz8%29%3F x 1,384,351 ops/sec ±0.13% (193 runs sampled)

Checking out "main"
Execute "npm run benchmark"

> fast-querystring@1.0.0 benchmark
> node benchmark/bench.js

encodeString: ""                                              x 292,799,787 ops/sec ±0.43% (184 runs sampled)
encodeString: "123"                                           x 117,099,112 ops/sec ±0.41% (190 runs sampled)
encodeString: "ä"                                             x 78,184,297 ops/sec ±0.23% (192 runs sampled)
encodeString: "𝌆"                                            x 33,798,774 ops/sec ±0.20% (192 runs sampled)
stringify: undefined                                       x 1,267,372,485 ops/sec ±0.10% (193 runs sampled)
stringify: null                                            x 1,266,193,122 ops/sec ±0.12% (195 runs sampled)
stringify: {}                                              x 135,395,111 ops/sec ±0.90% (189 runs sampled)
stringify: { id: true }                                    x 35,467,821 ops/sec ±0.87% (191 runs sampled)
stringify: { id: false }                                   x 36,171,446 ops/sec ±0.34% (192 runs sampled)
stringify: { id: 123 }                                     x 30,249,729 ops/sec ±0.53% (191 runs sampled)
stringify: { id: 1e+22 }                                   x 13,052,192 ops/sec ±0.41% (192 runs sampled)
stringify: { id: 123n }                                    x 20,219,939 ops/sec ±1.17% (194 runs sampled)
stringify: { id: Infinity }                                x 44,973,263 ops/sec ±0.30% (191 runs sampled)
stringify: { id: ["1", "3"] }                              x 16,815,705 ops/sec ±0.55% (189 runs sampled)
stringify: { id: "" }                                      x 41,464,595 ops/sec ±0.22% (193 runs sampled)
stringify: { id: "123" }                                   x 29,420,972 ops/sec ±0.47% (190 runs sampled)
stringify: { id: "ä" }                                     x 23,466,180 ops/sec ±0.57% (188 runs sampled)
stringify: { id: "𝌆" }                                    x 16,935,083 ops/sec ±0.56% (186 runs sampled)
stringify: { foo: ["1", "3"], bar: "2" }                   x 9,004,421 ops/sec ±0.28% (191 runs sampled)
parse:                                                 x 159,501,239 ops/sec ±0.13% (191 runs sampled)
parse: id=123                                          x 14,838,103 ops/sec ±0.43% (187 runs sampled)
parse: id=123&id=123                                   x 7,829,497 ops/sec ±0.59% (187 runs sampled)
parse: full%20name=Yagiz                               x 5,691,966 ops/sec ±0.20% (192 runs sampled)
parse: invalid%key=hello                               x 6,565,851 ops/sec ±0.27% (190 runs sampled)
parse: my+weird+field=q1%212%22%27w%245%267%2Fz8%29%3F x 1,370,609 ops/sec ±0.13% (194 runs sampled)

encodeString: ""                                             .+3.07%
encodeString: "123"                                          .+6.74%
encodeString: "ä"                                            +91.08%
encodeString: "𝌆"                                           .-1.58%
stringify: undefined                                      .-1.14%
stringify: null                                           .-0.78%
stringify: {}                                             .-9.07%
stringify: { id: true }                                   .+0.95%
stringify: { id: false }                                  .+3.52%
stringify: { id: 123 }                                    .-2.86%
stringify: { id: 1e+22 }                                  .-3.52%
stringify: { id: 123n }                                   .-3.48%
stringify: { id: Infinity }                               .+0.81%
stringify: { id: ["1", "3"] }                             .-2.94%
stringify: { id: "" }                                     .-4.34%
stringify: { id: "123" }                                  ..-3.1%
stringify: { id: "ä" }                                    +14.12%
stringify: { id: "𝌆" }                                   .+2.09%
stringify: { foo: ["1", "3"], bar: "2" }                  .+2.85%
parse:                                                .-0.22%
parse: id=123                                         ..+8.4%
parse: id=123&id=123                                  .+3.75%
parse: full%20name=Yagiz                              .+6.55%
parse: invalid%key=hello                              .+0.03%
parse: my+weird+field=q1%212%22%27w%245%267%2Fz8%29%3F....+1%
Back to encodeString-fast df5744c

[Fdawgs]

Morning @anonrig, could you cut a new release of this package please? Would love to take advantage of these performance improvements!

[anonrig]

I just released a new version. Thanks for the patience @Fdawgs https://github.com/anonrig/fast-querystring/releases/tag/1.1.0