# Multitude Performance Report
Generated by `scripts/perf_report.rs`:
- `cargo bench --bench criterion_alloc` and `criterion_drop` — criterion wall-clock timings.
- `cargo bench --bench gungraun_alloc` and `gungraun_drop` — Callgrind instruction-precise counts.
**Workload:** N = 1000 operations per measurement; slice element count = 8.
Criterion median is reported (default 30 samples, 1 s warm-up, 2 s measurement; override with `--samples` / `--measurement-time` / `--warm-up-time`).
Memory accesses = L1 Hits + LL Hits + RAM Hits (Callgrind D-cache references).
Bench names are aligned between criterion and gungraun via the `GROUPS` table in `scripts/perf_report.rs`.
## `arena_creation`
| `multitude` | 47 ns | — | — | — |
| `bumpalo` | 1 ns | — | — | — |
## `alloc_u64`
| `alloc` | 2.92 µs | 17,145 | 10 | 27,218 |
| `alloc_with` | 4.08 µs | 17,024 | 6 | 27,038 |
| `alloc_box` | 5.12 µs | 20,427 | 31 | 32,593 |
| `alloc_box_with` | 5.14 µs | 20,305 | 28 | 32,412 |
| `alloc_uninit_box` | 3.49 µs | 22,406 | 31 | 33,562 |
| `alloc_zeroed_box` | 4.81 µs | 24,404 | 35 | 36,558 |
| `alloc_rc` | 4.73 µs | 20,427 | 31 | 32,593 |
| `alloc_rc_with` | 4.70 µs | 20,305 | 28 | 32,412 |
| `alloc_uninit_rc` | 3.41 µs | 22,406 | 31 | 33,562 |
| `alloc_zeroed_rc` | 4.64 µs | 24,404 | 35 | 36,558 |
| `alloc_arc` | 4.96 µs | 22,412 | 35 | 34,571 |
| `alloc_arc_with` | 4.79 µs | 22,412 | 33 | 34,571 |
| `alloc_uninit_arc` | 3.54 µs | 22,410 | 34 | 33,567 |
| `alloc_zeroed_arc` | 4.87 µs | 23,408 | 34 | 35,563 |
| `bumpalo` | 5.41 µs | 21,022 | 5 | 31,037 |
| `bumpalo_with` | 5.26 µs | 21,020 | 5 | 31,034 |
## `alloc_str`
| `alloc_str` | 4.86 µs | 163,533 | 47 | 233,808 |
| `alloc_str_box` | 9.24 µs | 172,359 | 45 | 246,561 |
| `alloc_str_rc` | 9.52 µs | 172,359 | 45 | 246,561 |
| `alloc_str_arc` | 8.86 µs | 172,361 | 46 | 245,563 |
| `bumpalo` | 6.14 µs | 168,404 | 47 | 245,618 |
## `alloc_slice`
| `alloc_slice_copy` | 20.99 µs | 23,496 | 28 | 37,727 |
| `alloc_slice_clone` | 21.97 µs | 46,523 | 35 | 66,761 |
| `alloc_slice_fill_with` | 23.80 µs | 41,003 | 8 | 72,995 |
| `alloc_slice_fill_iter` | 23.80 µs | 43,007 | 7 | 77,001 |
| `alloc_slice_copy_box` | 23.31 µs | 38,747 | 39 | 60,074 |
| `alloc_slice_clone_box` | 23.48 µs | 59,770 | 43 | 88,100 |
| `alloc_slice_fill_with_box` | 25.44 µs | 52,250 | 22 | 90,336 |
| `alloc_slice_fill_iter_box` | 25.64 µs | 54,251 | 22 | 94,337 |
| `alloc_uninit_slice_box` | 4.99 µs | 28,375 | 25 | 42,531 |
| `alloc_zeroed_slice_box` | 21.69 µs | 31,376 | 24 | 47,532 |
| `alloc_slice_copy_rc` | 21.70 µs | 38,747 | 37 | 60,074 |
| `alloc_slice_clone_rc` | 21.91 µs | 59,770 | 42 | 88,100 |
| `alloc_slice_fill_with_rc` | 25.85 µs | 52,250 | 22 | 90,336 |
| `alloc_slice_fill_iter_rc` | 25.63 µs | 54,251 | 22 | 94,337 |
| `alloc_uninit_slice_rc` | 5.06 µs | 28,375 | 25 | 42,531 |
| `alloc_zeroed_slice_rc` | 21.91 µs | 31,376 | 24 | 47,532 |
| `alloc_slice_copy_arc` | 21.87 µs | 36,747 | 43 | 59,075 |
| `alloc_slice_clone_arc` | 22.39 µs | 66,733 | 43 | 101,054 |
| `alloc_slice_fill_with_arc` | 25.36 µs | 51,379 | 27 | 89,536 |
| `alloc_slice_fill_iter_arc` | 25.43 µs | 54,379 | 26 | 94,536 |
| `alloc_uninit_slice_arc` | 4.83 µs | 27,379 | 28 | 41,536 |
| `alloc_zeroed_slice_arc` | 21.00 µs | 30,380 | 28 | 46,537 |
| `bumpalo_copy` | 17.80 µs | 43,374 | 26 | 65,556 |
| `bumpalo_clone` | 18.38 µs | 61,376 | 27 | 78,557 |
| `bumpalo_fill_with` | 19.50 µs | 42,020 | 5 | 74,033 |
| `bumpalo_fill_iter` | 20.05 µs | 42,020 | 5 | 74,033 |
## `string_builder`
| `alloc_string` | 4.54 µs | 149,846 | 52 | 210,254 |
| `alloc_string_with_capacity` | 4.46 µs | 150,348 | 43 | 210,556 |
| `bumpalo_grow` | 7.16 µs | 150,866 | 99 | 211,948 |
| `bumpalo_with_cap` | 6.91 µs | 149,733 | 50 | 210,243 |
## `vec_builder`
| `alloc_vec` | 1.52 µs | 12,062 | 42 | 18,518 |
| `alloc_vec_with_capacity` | 1.46 µs | 12,464 | 28 | 19,680 |
| `bumpalo_grow` | 2.67 µs | 12,429 | 70 | 19,101 |
| `bumpalo_with_cap` | 2.61 µs | 11,219 | 15 | 17,332 |
## `drop`
| `box_u64` | 3.62 µs | — | — | — |
| `rc_u64` | 3.59 µs | — | — | — |
| `arc_u64` | 10.75 µs | — | — | — |
| `box_droppy` | 17.22 µs | 177,497 | 95 | 246,500 |
| `rc_droppy` | 18.25 µs | 213,875 | 77 | 304,904 |
| `arc_droppy` | 23.98 µs | 215,037 | 180 | 307,094 |
| `str_box` | 3.16 µs | 10,771 | 45 | 14,164 |
| `str_rc` | 3.25 µs | 10,771 | 45 | 14,164 |
| `str_arc` | 10.15 µs | 11,350 | 91 | 15,507 |
| `slice_box_u64` | 5.59 µs | — | — | — |
| `slice_rc_u64` | 5.61 µs | — | — | — |
| `slice_arc_u64` | 14.08 µs | — | — | — |
| `slice_box_droppy` | 101.98 µs | 1,184,336 | 1,122 | 1,604,745 |
| `slice_rc_droppy` | 129.42 µs | 1,520,784 | 3,122 | 2,147,736 |
| `slice_arc_droppy` | 134.02 µs | 1,522,860 | 3,246 | 2,150,872 |
| `alloc` | 1.03 µs | 535 | 25 | 812 |
## Multitude vs Bumpalo Head-to-Head
Direct comparisons of multitude versus bumpalo on identical workloads (the multitude variant chosen is the closest semantic equivalent to bumpalo's plain bump-allocation).
| `alloc_u64/alloc` vs `bumpalo` | 2.92 µs | 5.41 µs | -45.9% | 17,145 | 21,022 | -18.4% |
| `alloc_str/alloc_str` vs `bumpalo` | 4.86 µs | 6.14 µs | -20.9% | 163,533 | 168,404 | -2.9% |
| `alloc_slice/alloc_slice_copy` vs `bumpalo_copy` | 20.99 µs | 17.80 µs | +17.9% | 23,496 | 43,374 | -45.8% |
| `alloc_slice/alloc_slice_clone` vs `bumpalo_clone` | 21.97 µs | 18.38 µs | +19.5% | 46,523 | 61,376 | -24.2% |
| `alloc_slice/alloc_slice_fill_with` vs `bumpalo_fill_with` | 23.80 µs | 19.50 µs | +22.1% | 41,003 | 42,020 | -2.4% |
| `alloc_slice/alloc_slice_fill_iter` vs `bumpalo_fill_iter` | 23.80 µs | 20.05 µs | +18.7% | 43,007 | 42,020 | +2.3% |
| `string_builder/alloc_string` vs `bumpalo_grow` | 4.54 µs | 7.16 µs | -36.6% | 149,846 | 150,866 | -0.7% |
| `string_builder/alloc_string_with_capacity` vs `bumpalo_with_cap` | 4.46 µs | 6.91 µs | -35.5% | 150,348 | 149,733 | +0.4% |
| `vec_builder/alloc_vec` vs `bumpalo_grow` | 1.52 µs | 2.67 µs | -42.9% | 12,062 | 12,429 | -3.0% |
| `vec_builder/alloc_vec_with_capacity` vs `bumpalo_with_cap` | 1.46 µs | 2.61 µs | -44.1% | 12,464 | 11,219 | +11.1% |