Kwker benchmark results: each language's built-in sort against Kwker's binding ================================================================================ Published on https://kwker.io/benchmarks/#langs. Raw output of the bindings' own benchmark (part of their test run). Machine Intel Xeon (Cascade Lake, family 6 model 85, AVX-512), cloud VM; Kwker 0.26.0, AVX-512 engine Date JavaScript 2026-10-08; Java 2026-10-03. Re-run on 2026-10-08, Java 25's own Arrays.sort (its x86-simd-sort AVX-512 intrinsics) read 24 ms on 1M ints in three runs against 9 ms before, while Java 21's scalar sort was unchanged - an AVX-512 anomaly of that machine, so the Java rows stay those of 2026-10-03 until a clean re-run. Data 2^20 (1,048,576) random keys of the named type; the built-in sort and Kwker's sort each run on a fresh copy of the same array in the same process. Java: 9 rounds, the best of the last 7 (2 warm-up rounds); JavaScript: best of 5. The same test run first compares Kwker's results with the built-in sort's ("all N checks passed"). Runtimes Java 21.0.10 (Kwker through JNI) and Java 25.0.4.1 (through the FFM API); Node.js 22.22.2, Bun 1.3.11, Deno 2.9.6 (Node-API addon) Note: Java 22 and later use x86-simd-sort intrinsics inside Arrays.sort for primitive arrays, which is why Java 25's built-in sort is 4-9x faster than Java 21's; the published table uses Java 25. Java [java21 avx512] kwker 0.26.0 (avx512, JNI), Java 21.0.10: all 743 checks passed [java21 avx512] 1M int Arrays.sort 78.57 ms Kwker.sort 2.87 ms 27.4x [java21 avx512] 1M long Arrays.sort 80.42 ms Kwker.sort 5.62 ms 14.3x [java21 avx512] 1M float Arrays.sort 82.73 ms Kwker.sort 3.04 ms 27.2x [java21 avx512] 1M double Arrays.sort 85.46 ms Kwker.sort 5.97 ms 14.3x [java25 avx512] kwker 0.26.0 (avx512, FFM), Java 25.0.4.1: all 743 checks passed [java25 avx512] 1M int Arrays.sort 9.03 ms Kwker.sort 2.97 ms 3.0x [java25 avx512] 1M long Arrays.sort 16.78 ms Kwker.sort 5.58 ms 3.0x [java25 avx512] 1M float Arrays.sort 11.74 ms Kwker.sort 3.20 ms 3.7x [java25 avx512] 1M double Arrays.sort 19.19 ms Kwker.sort 6.13 ms 3.1x JavaScript (TypedArray.prototype.sort vs kwker.sort) [node avx512] kwker 0.1.0 (avx512), node v22.22.2: all 6711 checks passed [node avx512] 1M Int32Array TypedArray.sort 59.85 ms kwker.sort 3.11 ms 19.2x [node avx512] 1M Float32Array TypedArray.sort 99.04 ms kwker.sort 2.99 ms 33.2x [node avx512] 1M Float64Array TypedArray.sort 112.76 ms kwker.sort 7.51 ms 15.0x [node avx512] 1M BigInt64Array TypedArray.sort 67.34 ms kwker.sort 8.62 ms 7.8x [node avx512] 1M Uint16Array TypedArray.sort 43.05 ms kwker.sort 3.57 ms 12.1x [bun avx512] kwker 0.1.0 (avx512), bun 1.3.11: all 6711 checks passed [bun avx512] 1M Int32Array TypedArray.sort 55.36 ms kwker.sort 4.88 ms 11.4x [bun avx512] 1M Float32Array TypedArray.sort 63.49 ms kwker.sort 3.00 ms 21.2x [bun avx512] 1M Float64Array TypedArray.sort 64.72 ms kwker.sort 7.17 ms 9.0x [bun avx512] 1M BigInt64Array TypedArray.sort 66.49 ms kwker.sort 8.40 ms 7.9x [bun avx512] 1M Uint16Array TypedArray.sort 44.51 ms kwker.sort 3.49 ms 12.7x [deno avx512] kwker 0.1.0 (avx512), deno 2.9.6: all 6711 checks passed [deno avx512] 1M Int32Array TypedArray.sort 24.23 ms kwker.sort 3.66 ms 6.6x [deno avx512] 1M Float32Array TypedArray.sort 81.50 ms kwker.sort 3.19 ms 25.6x [deno avx512] 1M Float64Array TypedArray.sort 88.13 ms kwker.sort 7.10 ms 12.4x [deno avx512] 1M BigInt64Array TypedArray.sort 24.82 ms kwker.sort 8.94 ms 2.8x [deno avx512] 1M Uint16Array TypedArray.sort 21.83 ms kwker.sort 3.55 ms 6.1x