perf: sharded OnceTable

[Huliiiiii]

Alternative of #205

Improvements

CPU: AMD Ryzen 9 9950X
OS: NixOS 25.11

Thread: 8

Benchmark (entry size)	Baseline	#205	This PR (Mutex)	This PR vs. #205	parking_lot::RwLock	RwLock vs. #205	RwLock vs. This PR
once_map::contended_compute_coalesced	7.027 µs	6.650 µs	6.906 µs	3.7% slower	6.666 µs	0.2% slower	3.5% faster
once_map::contended_compute_hit_same_key	964 ns	296 ns	1.113 µs	73.4% slower	312 ns	5.1% slower	3.6× faster
once_map::contended_compute_hit_disjoint (64)	1.114 µs	357 ns	110 ns	3.2× faster	103 ns	3.5× faster	6.3% faster
once_map::contended_compute_hit_disjoint (1024)	1.156 µs	296 ns	94 ns	3.1× faster	103 ns	2.9× faster	8.4% slower
once_map::contended_compute_miss_churn (64)	3.552 µs	4.979 µs	208 ns	23.9× faster	223 ns	22.4× faster	6.6% slower
once_map::contended_compute_miss_churn (1024)	3.542 µs	4.867 µs	223 ns	21.8× faster	231 ns	21.1× faster	3.4% slower
once_map::contended_compute_mixed (64)	2.589 µs	3.305 µs	163 ns	20.3× faster	166 ns	19.9× faster	1.9% slower
once_map::contended_compute_mixed (1024)	2.513 µs	3.200 µs	169 ns	18.9× faster	168 ns	19.0× faster	0.4% faster
once_map::contended_get_hit_same_key	852 ns	322 ns	960 ns	66.5% slower	638 ns	49.5% slower	1.5× faster
once_map::contended_get_hit_disjoint (64)	935 ns	464 ns	89 ns	5.2× faster	124 ns	3.8× faster	28.0% slower
once_map::contended_get_hit_disjoint (1024)	963 ns	461 ns	65 ns	7.1× faster	70 ns	6.6× faster	7.4% slower
singleflight::contended_work_coalesced	6.746 µs	6.937 µs	7.351 µs	5.6% slower	6.594 µs	4.9% faster	10.3% faster
singleflight::contended_work_same_key	1.881 µs	2.167 µs	1.719 µs	1.3× faster	1.790 µs	1.2× faster	4.0% slower
singleflight::contended_work_disjoint_churn	3.533 µs	4.780 µs	217 ns	22.0× faster	221 ns	21.7× faster	1.6% slower

Thread: 32

Benchmark (entry size)	Baseline	#205	This PR (Mutex)	This PR vs. #205	parking_lot::RwLock	RwLock vs. #205	RwLock vs. This PR
once_map::contended_compute_coalesced	28.900 µs	31.010 µs	29.900 µs	3.6% faster	46.520 µs	33.3% slower	35.7% slower
once_map::contended_compute_hit_same_key	3.915 µs	2.576 µs	6.607 µs	61.0% slower	2.097 µs	18.6% faster	3.2× faster
once_map::contended_compute_hit_disjoint (64)	5.513 µs	3.004 µs	882 ns	3.4× faster	278 ns	10.8× faster	3.2× faster
once_map::contended_compute_hit_disjoint (1024)	5.981 µs	2.869 µs	334 ns	8.6× faster	243 ns	11.8× faster	1.4× faster
once_map::contended_compute_miss_churn (64)	29.290 µs	42.010 µs	948 ns	44.3× faster	1.246 µs	33.7× faster	23.9% slower
once_map::contended_compute_miss_churn (1024)	26.960 µs	38.790 µs	988 ns	39.3× faster	1.257 µs	30.9× faster	21.4% slower
once_map::contended_compute_mixed (64)	19.780 µs	25.620 µs	822 ns	31.2× faster	872 ns	29.4× faster	5.7% slower
once_map::contended_compute_mixed (1024)	19.140 µs	24.010 µs	721 ns	33.3× faster	780 ns	30.8× faster	7.6% slower
once_map::contended_get_hit_same_key	3.428 µs	2.098 µs	5.190 µs	59.6% slower	2.954 µs	29.0% slower	1.8× faster
once_map::contended_get_hit_disjoint (64)	4.849 µs	2.755 µs	623 ns	4.4× faster	319 ns	8.6× faster	2.0× faster
once_map::contended_get_hit_disjoint (1024)	4.896 µs	2.810 µs	325 ns	8.6× faster	234 ns	12.0× faster	1.4× faster
singleflight::contended_work_coalesced	30.750 µs	28.620 µs	35.160 µs	18.6% slower	32.200 µs	11.1% slower	8.4% faster
singleflight::contended_work_same_key	12.920 µs	13.960 µs	13.120 µs	6.0% faster	17.480 µs	20.1% slower	24.9% slower
singleflight::contended_work_disjoint_churn	28.160 µs	39.170 µs	916 ns	42.8× faster	1.274 µs	30.7× faster	28.1% slower

[aisk]

Verified on bare-metal Linux: AMD Ryzen 7 5700X (8C/16T), NixOS 26.11, kernel 7.2.0, rustc 1.96.0, t=8, --sample-count 100 --sample-size 10000, interleaved rounds ×3 against base (c461305) and #205's head (6c55ca4). Values are medians of the 3 round-medians; per-round variance was <5% throughout, including the churn scenarios.

scenario	base	this PR	#205 (6c55ca4)
compute_hit_same_key	809.1 ns	810.3 ns	165.9 ns
get_hit_same_key	547.9 ns	599.0 ns	343.9 ns
compute_hit_disjoint (64)	908.3 ns	100.9 ns	168.5 ns
compute_hit_disjoint (1024)	913.0 ns	96.3 ns	174.7 ns
get_hit_disjoint (64)	715.6 ns	84.2 ns	518.9 ns
get_hit_disjoint (1024)	727.3 ns	78.8 ns	530.8 ns
compute_miss_churn (64)	3.06 µs	223.6 ns	4.05 µs
compute_miss_churn (1024)	2.91 µs	232.5 ns	3.92 µs
compute_mixed (64)	2.21 µs	175.1 ns	2.63 µs
compute_mixed (1024)	2.16 µs	179.0 ns	2.54 µs
compute_coalesced	18.41 µs	18.53 µs	18.47 µs
work_same_key	2.07 µs	1.80 µs	2.14 µs
work_disjoint_churn	3.02 µs	233.4 ns	3.74 µs
work_coalesced	18.58 µs	18.47 µs	18.67 µs

• The headline wins reproduce: 8.5–9.5× on disjoint hits, 12–13.7× on miss_churn/mixed/work_disjoint_churn (below the 16–24× on the 9950X, still decisive).
• The same-key penalty is milder here: get_hit_same_key −9% vs base, compute_hit_same_key on par, work_same_key +15%.
• Same-key hits are the one weak spot vs perf: take the keyed table read lock on hits and duplicate waiters #205: every hit takes the shard mutex exclusively, so same-key traffic serializes on one shard — 810 vs 166 ns on compute hits (4.9×), 599 vs 344 ns on get (1.7×).
• Question: the branch only contains the Mutex-shard variant; the parking_lot::RwLock column from the table doesn't seem to be in the code. Is that variant pushed anywhere?

The two approaches look complementary: sharding covers disjoint/churn, a shared reader path plus the hit shortcut from 6c55ca4 covers same-key. A per-shard reader path on top of this PR should cover every scenario — happy to help with that.

[Huliiiiii]

Refactor the storage of OnceMap, add a concurrent index for Ready values.
This makes the current implementation significantly superior to other solutions when using two or more threads.

Also fixed some issues in benchmarks, there is still too much variance in the results.

Threads	Benchmark	Range (Span)
1	compute_hit_disjoint (64)	Current: 30.76–33.92 (10.2%)
1	compute_hit_disjoint (1024)	Current: 31.03–34.44 (10.8%)
2	compute_hit_disjoint (64)	Mutex: 40.23–50.19 (24.6%); RwLock: 36.16–67.90 (85.0%)
2	compute_hit_disjoint (1024)	Mutex: 48.72–53.93 (10.4%)
2	get_hit_disjoint (64)	Mutex: 24.17–48.20 (98.0%); RwLock: 24.40–43.33 (77.0%)
2	get_hit_same_key	Mutex: 115.2–137.2 (16.5%)
8	compute_hit_disjoint (64)	Mutex: 56.25–70.49 (23.8%); RwLock: 52.71–69.58 (28.0%)
8	compute_hit_disjoint (1024)	Mutex: 70.51–78.34 (11.1%)
8	compute_hit_same_key	Mutex: 972.3–1126 (14.1%); RwLock: 304.0–374.5 (22.1%)
8	compute_miss_churn (1024)	Mutex: 222.5–251.6 (12.8%); Current: 172.0–208.1 (20.9%)
8	compute_mixed (64)	Current: 98.62–110.0 (11.4%)
8	compute_mixed (1024)	Mutex: 160.7–208.3 (29.5%)
8	get_hit_disjoint (64)	Mutex: 37.26–54.95 (41.5%); RwLock: 36.37–46.83 (24.3%)
8	get_hit_same_key	RwLock: 625.9–1000 (52.7%)
32	compute_hit_disjoint (64)	Mutex: 149.6–212.6 (36.6%); RwLock: 109.9–150.8 (30.5%)
32	compute_hit_disjoint (1024)	RwLock: 163.3–181.9 (10.9%)
32	compute_hit_same_key	RwLock: 2087–3257 (52.7%)
32	get_hit_disjoint (64)	Mutex: 99.28–144.3 (35.1%); RwLock: 93.72–127.0 (33.1%)
32	get_hit_same_key	RwLock: 2575–3316 (28.0%)

New Result:

1 thread

Benchmark	Mutex vs. baseline	RwLock vs. baseline	RwLock (non-sharded) vs. baseline	Current vs. baseline	Current vs. Mutex	Current vs. RwLock	Current vs. RwLock (non-sharded)
once_map::contended_compute_hit_disjoint (64)	23.8% slower	16.3% slower	On par	4.0% slower	16.0% faster	10.6% faster	4.9% slower
once_map::contended_compute_hit_disjoint (1024)	26.2% slower	15.9% slower	On par	2.9% slower	23.0% faster	11.3% faster	3.1% slower
once_map::contended_compute_hit_same_key	28.5% slower	22.7% slower	2.4% slower	2.6% faster	32.0% faster	26.0% faster	4.8% faster
once_map::contended_compute_miss_churn (64)	3.9% slower	On par	On par	15.0% slower	10.7% slower	13.2% slower	15.9% slower
once_map::contended_compute_miss_churn (1024)	5.9% slower	5.5% slower	3.4% faster	14.3% slower	7.9% slower	8.4% slower	18.3% slower
once_map::contended_compute_mixed (64)	9.0% slower	6.7% slower	On par	12.5% slower	3.2% slower	5.4% slower	11.3% slower
once_map::contended_compute_mixed (1024)	14.2% slower	10.6% slower	On par	13.1% slower	On par	2.3% slower	13.0% slower
once_map::contended_get_hit_disjoint (64)	30.8% slower	33.0% slower	5.7% faster	6.9% faster	41.0% faster	43.0% faster	On par
once_map::contended_get_hit_disjoint (1024)	31.3% slower	38.2% slower	5.7% faster	6.9% faster	41.0% faster	48.0% faster	On par
once_map::contended_get_hit_same_key	39.1% slower	55.2% slower	8.7% faster	7.9% faster	51.0% faster	68.0% faster	On par

2 threads

Benchmark	Mutex vs. baseline	RwLock vs. baseline	RwLock (non-sharded) vs. baseline	Current vs. baseline	Current vs. Mutex	Current vs. RwLock
once_map::contended_compute_hit_disjoint (64)	3.06× faster	3.23× faster	3.2% slower	3.74× faster	22.0% faster	13.6% faster
once_map::contended_compute_hit_disjoint (1024)	2.54× faster	2.72× faster	2.2% slower	3.94× faster	56.0% faster	45.0% faster
once_map::contended_compute_hit_same_key	22.2% slower	5.1% slower	3.3% faster	4.29× faster	5.24× faster	4.51× faster
once_map::contended_compute_miss_churn (64)	3.41× faster	3.41× faster	4.8% slower	3.31× faster	3.1% slower	2.9% slower
once_map::contended_compute_miss_churn (1024)	3.08× faster	3.09× faster	3.3% slower	3.06× faster	On par	On par
once_map::contended_compute_mixed (64)	3.38× faster	3.45× faster	6.6% slower	3.58× faster	5.5% faster	3.6% faster
once_map::contended_compute_mixed (1024)	2.97× faster	3.02× faster	On par	3.47× faster	14.5% faster	13.0% faster
once_map::contended_get_hit_disjoint (64)	4.87× faster	4.62× faster	64.0% faster	6.81× faster	40.0% faster	47.0% faster
once_map::contended_get_hit_disjoint (1024)	3.67× faster	3.71× faster	54.0% faster	6.52× faster	77.0% faster	76.0% faster
once_map::contended_get_hit_same_key	32.0% slower	4.9% slower	53.0% faster	6.28× faster	8.29× faster	6.59× faster

8 threads

Benchmark	Mutex vs. baseline	RwLock vs. baseline	Current vs. baseline	Current vs. Mutex	Current vs. RwLock
once_map::contended_compute_hit_disjoint (64)	18.66× faster	17.56× faster	31.49× faster	69.0% faster	79.0% faster
once_map::contended_compute_hit_disjoint (1024)	14.42× faster	14.90× faster	31.74× faster	2.20× faster	2.13× faster
once_map::contended_compute_hit_same_key	20.9% slower	12.8% slower	25.82× faster	31.21× faster	29.13× faster
once_map::contended_compute_miss_churn (64)	16.01× faster	16.46× faster	17.05× faster	6.1% faster	3.5% faster
once_map::contended_compute_miss_churn (1024)	15.61× faster	13.91× faster	16.33× faster	4.4% faster	14.8% faster
once_map::contended_compute_mixed (64)	14.23× faster	18.12× faster	19.88× faster	40.0% faster	8.8% faster
once_map::contended_compute_mixed (1024)	13.63× faster	13.08× faster	20.81× faster	53.0% faster	59.0% faster
once_map::contended_get_hit_disjoint (64)	18.56× faster	22.29× faster	51.48× faster	2.77× faster	2.31× faster
once_map::contended_get_hit_disjoint (1024)	17.19× faster	17.26× faster	52.50× faster	3.05× faster	3.04× faster
once_map::contended_get_hit_same_key	24.2% slower	38.0% faster	48.16× faster	59.82× faster	34.90× faster

32 threads

Benchmark	Mutex vs. baseline	RwLock vs. baseline	Current vs. baseline	Current vs. Mutex	Current vs. RwLock
once_map::contended_compute_hit_disjoint (64)	35.88× faster	36.90× faster	119.78× faster	3.34× faster	3.25× faster
once_map::contended_compute_hit_disjoint (1024)	23.48× faster	20.96× faster	136.17× faster	5.80× faster	6.50× faster
once_map::contended_compute_hit_same_key	46.0% slower	36.6% slower	114.70× faster	167.46× faster	156.64× faster
once_map::contended_compute_miss_churn (64)	28.91× faster	29.47× faster	35.27× faster	22.0% faster	16.5% faster
once_map::contended_compute_miss_churn (1024)	29.01× faster	28.94× faster	36.95× faster	27.0% faster	28.0% faster
once_map::contended_compute_mixed (64)	29.71× faster	29.23× faster	51.15× faster	72.0% faster	75.0% faster
once_map::contended_compute_mixed (1024)	26.58× faster	27.89× faster	50.45× faster	90.0% faster	81.0% faster
once_map::contended_get_hit_disjoint (64)	34.82× faster	39.42× faster	174.31× faster	5.01× faster	4.42× faster
once_map::contended_get_hit_disjoint (1024)	18.09× faster	28.76× faster	179.66× faster	9.93× faster	6.25× faster
once_map::contended_get_hit_same_key	37.8% slower	43.0% faster	153.96× faster	212.13× faster	107.46× faster