[Webkit-unassigned] [Bug 244754] New: String prototype functions are slower in JSC than V8

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=244754

            Bug ID: 244754
           Summary: String prototype functions are slower in JSC than V8
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: JavaScriptCore
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: jarred at jarredsumner.com

❯ jsc foo.mjs
cpu: unknown
runtime: unknown (unknown)

benchmark                                        time (avg)             (min … max)       p75       p99      p995
----------------------------------------------------------------------------------- -----------------------------
startsWith                                       11 ns/iter   (10.51 ns … 21.54 ns)  10.78 ns  15.12 ns  15.71 ns
prototype_startsWith                          34.36 ns/iter  (32.08 ns … 150.64 ns)  34.63 ns  39.24 ns  80.12 ns
_startsWith                                   38.76 ns/iter  (36.85 ns … 124.11 ns)  39.07 ns  47.63 ns  81.98 ns
substring(0,8)                                36.64 ns/iter  (34.45 ns … 148.93 ns)  37.06 ns   45.9 ns  74.68 ns
substring(0, length) === (substr strict)      38.45 ns/iter  (36.84 ns … 155.19 ns)  38.82 ns  47.86 ns  75.75 ns
substring(0, length) == (substr non-strict)   38.76 ns/iter   (36.8 ns … 130.45 ns)   39.1 ns  47.66 ns  69.83 ns
indexOf strict                                15.75 ns/iter   (15.49 ns … 24.07 ns)  15.55 ns  17.25 ns  17.47 ns
indexOf non-strict                            13.67 ns/iter      (13 ns … 18.34 ns)  14.52 ns  17.06 ns  17.14 ns


❯ node foo.mjs
cpu: Apple M1 Max
runtime: node v18.4.0 (arm64-darwin)

benchmark                                        time (avg)             (min … max)       p75       p99      p995
----------------------------------------------------------------------------------- -----------------------------
startsWith                                    19.11 ns/iter  (18.58 ns … 404.22 ns)  18.68 ns  23.09 ns  26.93 ns
prototype_startsWith                          16.19 ns/iter   (14.25 ns … 28.51 ns)  17.43 ns   23.3 ns  24.11 ns
_startsWith                                   16.14 ns/iter  (14.55 ns … 285.63 ns)   16.8 ns  21.92 ns  22.95 ns
substring(0,8)                                15.94 ns/iter   (14.25 ns … 48.39 ns)  16.06 ns  25.28 ns  28.27 ns
substring(0, length) === (substr strict)      15.34 ns/iter   (14.25 ns … 31.48 ns)  15.72 ns  22.83 ns  24.01 ns
substring(0, length) == (substr non-strict)   15.87 ns/iter   (14.25 ns … 31.93 ns)  15.82 ns  24.57 ns  25.63 ns
indexOf strict                                14.68 ns/iter    (13.88 ns … 21.2 ns)   14.8 ns  16.66 ns  16.98 ns
indexOf non-strict                            14.62 ns/iter   (14.29 ns … 18.91 ns)  14.46 ns   16.8 ns  17.02 ns



Code:

```
import { run, bench, baseline } from "./node_modules/mitata/src/cli.mjs";

function _startsWith(str, find) {
    return str.substring(0,find.length) === find
}

String.prototype._startsWith = function (find) {
    return this.substring(0,find.length) === find
}

baseline("startsWith", () => "/assets/omfg".startsWith('/assets/'));
baseline("prototype_startsWith", () => "/assets/omfg"._startsWith('/assets/'));
baseline("_startsWith", () => _startsWith("/assets/omfg", '/assets/'));
baseline("substring(0,8)", () => "/assets/omfg".substring(0,8) === '/assets/');
baseline("substring(0, length) === (substr strict)", () => "/assets/omfg".substring(0,'/assets/'.length) === '/assets/');
baseline("substring(0, length) == (substr non-strict)", () => "/assets/omfg".substring(0,'/assets/'.length) == '/assets/');
baseline("indexOf strict", () => "/assets/omfg".indexOf('/assets/') === 0);
baseline("indexOf non-strict", () => "/assets/omfg".indexOf('/assets/') == 0);

await run();
```

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20220903/9590ff3d/attachment-0001.htm>