关于「实战工具与检查器规格:并发协同台账、四件套条款与预注册检验」,《并发拓扑耦合与黑板复用》给出的定量结论是什么?

结论:第八章给出并发耦合度与黑板复用增益试算器、检查器规格表、台账节奏与预注册检验,检验每臂需 156 次记录。书中以标注框区分已发表的证据、可运行的计算(21 段程序)、原典与作者的假说。

数据口径:书中参数为示意,数据为合成,全部计算由书中程序生成;本书是非个人化经营建议,不提供任何加速比承诺。 第一章说明峰值从何而来,示意参数下峰值在 6.78 个窗口附近。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 156

A. repeat-failure share 0.30 vs 0.18156 encounters per arm
if coupled arm is 0.1595 per arm
if coupled arm is 0.21289 per arm
if coupled arm is 0.24676 per arm
B. reuse rate 0.30 vs 0.4562 entries
C. simulated power at n156: 0.792 (2,000 studies)
false-alarm rate when arms are equal0.043
查看完整验算程序
# Chapter 8. p82: pre-registered sample sizes and a power check.
# Test 1: share of pitfall encounters that repeat a failure another
# node already posted; isolated arm p0 = 0.30, coupled arm p1 = 0.18
# (illustrative). One-sided two-proportion z test, alpha 0.05, power
# 0.80. Test 2: reuse rate (entries cited by another node / entries
# written); null 0.30, target 0.45, one-sample, same alpha and power.
# A Monte Carlo of 2000 simulated studies checks the power of test 1.
import math
import random
from statistics import NormalDist

Z = NormalDist()
ALPHA, POWER = 0.05, 0.80


def n_two(p0, p1):
    za, zb = Z.inv_cdf(1 - ALPHA), Z.inv_cdf(POWER)
    pb = (p0 + p1) / 2
    num = (za * math.sqrt(2 * pb * (1 - pb))
           + zb * math.sqrt(p0 * (1 - p0) + p1 * (1 - p1))) ** 2
    return math.ceil(num / (p0 - p1) ** 2)


def n_one(p0, p1):
    za, zb = Z.inv_cdf(1 - ALPHA), Z.inv_cdf(POWER)
    num = (za * math.sqrt(p0 * (1 - p0))
           + zb * math.sqrt(p1 * (1 - p1))) ** 2
    return math.ceil(num / (p1 - p0) ** 2)


def power_mc(p0, p1, n, runs, rng):
    za = Z.inv_cdf(1 - ALPHA)
    hits = 0
    for _ in range(runs):
        x0 = x1 = 0
        for _ in range(n):
            if rng.random() < p0:
                x0 += 1
            if rng.random() < p1:
                x1 += 1
        pb = (x0 + x1) / (2 * n)
        se = math.sqrt(2 * pb * (1 - pb) / n)
        if se > 0 and (x0 - x1) / n / se > za:
            hits += 1
    return hits / runs


if __name__ == "__main__":
    n1 = n_two(0.30, 0.18)
    print("A. repeat-failure share 0.30 vs 0.18: {:,} encounters per"
          " arm".format(n1))
    for p1 in (0.15, 0.21, 0.24):
        print("   if coupled arm is {:.2f}: {:,} per arm".format(
            p1, n_two(0.30, p1)))
    n2 = n_one(0.30, 0.45)
    print("B. reuse rate 0.30 vs 0.45: {:,} entries".format(n2))
    rng = random.Random(41182)
    pw = power_mc(0.30, 0.18, n1, 2000, rng)
    print("C. simulated power at n = {:,}: {:.3f} ({:,} studies)".format(
        n1, pw, 2000))
    alpha = power_mc(0.30, 0.30, n1, 2000, rng)
    print("   false-alarm rate when arms are equal: {:.3f}".format(alpha))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《多核并发操作拓扑耦合、共享黑板无锁通信与跨节点涌现增益母典》第 8 章「实战工具与检查器规格:并发协同台账、四件套条款与预注册检验」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-411 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026411,
  author    = {Fan, Yafu},
  title     = {Topological Coupling of Concurrent Multi-Core Operations: Lock-Free Blackboard Signalling and Cross-Node Emergent Gains},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论