Files
MobileGL/docs/Disaggregated/MEASUREMENTS.md
T
BZLZHH 7c72dfde2b [Docs] (P12): the ~5 ms per round trip is the DEVICE's deep-idle exit, not the link and not the client - which is what the handoff measured and could not explain
- Three one-variable experiments on the same device, same workload (the 43k-record frame): a 40x
  larger client spin moves the per-wait mean by ~10% (so it is not the client park/wake); the
  same frames over adb forward - a link whose raw RTT is several times lower - by ~25-30% (so it
  is not the link, which is what the handoff saw in 3.6 and could not account for); and pinning
  the device (governor performance on all eight cores, cpuoff_l / clusteroff_l / mcusysoff /
  s2idle disabled) by 26% per wait and 25% of the whole frame: 221.5 s -> 166.5 s for 43,218
  records and 100.5 MB staged, against 43,232 and 101.4 MB in the baseline.
- The mechanism: the server's apply thread parks between records, the SoC drops into a deep idle
  state, and every record pays the exit latency - milliseconds, once per record, which is the
  entire shape of the frame. It also matches the server profile the handoff took (65% socket
  syscalls, ~15% doorbell spin: waiting and waking, not working).
- Stated where it matters: the handoff's 249 s was taken on an UNPINNED device, so its absolute
  numbers carry this idle tax - the project has a pinning discipline for exactly this reason.
- The lever list is now ordered: batch the waits (K <= 8 by the reply-slot pool, client-only),
  keep the server hot between records (its own spin budget, which the APK does not wire yet), and
  measure split numbers on a pinned device. The device was restored to stock afterwards.
2026-09-25 21:28:13 +08:00

46 lines
5.4 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# 实测(索引)
> 这一页只有结论和去向;完整的表、命令、协议与读法在各阶段 `notes/<阶段>/README.md` 的"实测"节。代码和旧笔记里的 `MEASUREMENTS.md §N` 按下表"§"列找到对应阶段(小节编号在那里原样保留)。
**口径**:性能只记录、不设门;看**逐线程 CPU 时间**,不看整机帧率。性能对比只用一台手机(Redmi `2f7cbe2e`,Adreno 830),定频、同热窗口、两臂交替跑([`guide/pin-verification-2026-09-07.md`](guide/pin-verification-2026-09-07.md))。
## 关键结论
- **接口本身的代价**:单进程里改成"前端推、后端收",CPU 多约 6–12%(P2),已接受;顶点数组切换密集的场景更高(P3a,+27–30%)。
- **拆成两线程**:起初游戏里只有 7–13 fps;优化后 103–106 fps(P5d);client 发完就走以后,重负载(渲染距离 32)下与单线程持平(P5e)。
- **拆成两进程**:真机上与两线程成本持平;两者的 GL 线程 CPU 都比单线程少约 30%(后端工作挪到了另一线程 / 进程)(P6)。
- **跨机**:电脑经 TCP 连手机,102 个集成用例全过;server 被杀后 133 ms、Wi-Fi 断开后约 5 s 干净报"设备丢失"(P6.5)。
- **纹理上传的等待粒度(P12,handoff §4)**:wire 臂上纹理那半 `resource_subdata` 不再逐条买回包,等待点从「每条记录」落到「SEG_STAGE 窗口用尽」。
单条纯上传占用的回包等待 **1 → 0**(`wait_replies`,正是 §3.2 里"一帧 55,428"的那个计数器)。
按形状实测(`RemoteClientControls.ReplyWaitsByOpForAnAtlasShapedLoad`,主机 inproc,按 op 计数):**图集拼接**(一张 level 上 64 次 `glTexSubImage2D`)**64 → 0** 次回包等待;
**每 sprite 一张纹理**(64 张纹理 + 256 次 `glTexParameteri`)**384 → 320**,剩下的全是 per-resource / per-call 的三条 row(create 76 / respecify 130 / params 130)——
它们的拒绝是「服务端没有这个 handle」的唯一客户端可见信号,删掉等待就是删掉信号,裁定见 `CONTRACT-P5E.md` §2.5。
真机(TrebleDroid GSI,MC 26.2 经 TCP 入世界)那一帧的构成为 `ResourceRespecify` 20,633 + `SetTextureParams` 18,056 + `ResourceCreate` 4,536 条**各等一次**,
而 `ResourceSubData` 12,671 条**零等待**(B 的实机证据);这些记录都是真实状态变化,客户端没有可去重的(同一值的 `glTexParameteri` 本来就不产生记录)。
- **单次往返那 ~5 ms 是设备侧的深度空闲退出,不是链路也不是客户端**(P12,本轮实测):同一帧、同一批记录,
客户端自旋 40× 只快 ~10%;换 `adb forward`(USB)快 ~25–30%;而把设备 8 核钉 `performance` 并关掉 `cpuoff_l`/`clusteroff_l`/`mcusysoff`/`s2idle` 后,
单次等待 **5,040 → 3,715 µs**、那一帧墙钟 **221.5 s → 166.5 s(−25%)**——记录数与字节数都不变。
这解释了 handoff §3.6「换快 3.4 倍的链路墙钟不变」:钱花在设备每次唤醒上;handoff §3.2 那 249 s 是在**未定频**的机器上取的,含这份空闲税。
见 [`notes/p12/DEVICELOST-AND-UPLOAD-WAITS.md`](notes/p12/DEVICELOST-AND-UPLOAD-WAITS.md) §2.6–2.10。
- **画面正确**:Vulkan 后端真机画面检查 36/36 通过,CTS 五块相对基线没有超过 0.5 个百分点的退步、没有新崩溃(P7);server 自有窗口上屏两后端 SSIM 1.0(P12)。
## 按阶段
| § | 阶段 | 一句结论 | 全文 |
|---|---|---|---|
| 1 | P0 | 能从应用进程拉起第二个进程;跨进程共享 GPU 内存只有一条路在所有设备上可用 | [`notes/p0`](notes/p0/README.md) |
| 2 | P1 | 逐 draw 核对推送状态,79 条 trace 零分歧 | [`notes/p1`](notes/p1/README.md) |
| 3 | P2 | 桌面基准:新 tracker 比它替掉的旧填充更便宜 | [`notes/p2`](notes/p2/README.md) |
| 4 | P3a | 两机真机 Release:接口代价 +6–12%,VAO 切换密集的场景 +27–30% | [`notes/p3a`](notes/p3a/README.md) |
| 5 | P4a | 换到 Redmi;P4a 自己只加 3–6 个百分点 | [`notes/p4a`](notes/p4a/README.md) |
| 6 | P5 | 两线程第一帧画对;真机两线程臂还跑不完 | [`notes/p5`](notes/p5/README.md) |
| 7 | P5b | 真机两线程跑通四条游戏 trace;两线程额外代价 +6–18% | [`notes/p5b`](notes/p5b/README.md) |
| 8 | P5c | 两线程之间零共享内存后,全部测试仍绿 | [`notes/p5c`](notes/p5c/README.md) |
| 9 | P5d | 游戏内 7–13 → 103–106 fps,client CPU 15.0 → 9.2 ms/帧 | [`notes/p5d`](notes/p5d/README.md) |
| 10–11 | P5e | 发完就走比每条都等快 69%;重负载下与单线程持平 | [`notes/p5e`](notes/p5e/README.md) |
| 12 | P6 | 两进程与两线程持平;每帧数据量与记录数首次实测 | [`notes/p6`](notes/p6/README.md) |
| 13 | P3b/P4b | 纹理上传形状在四种拓扑下完全一致 | [`notes/p34b`](notes/p34b/README.md) |
| 14 | P6.5 | 跨机 TCP 102/102;断线约 5 s 检测到 | [`notes/p65`](notes/p65/README.md) |
| 15 | P7 | 真机画面 36/36;server 内存无界增长收住(主机 825 → 546 MiB) | [`notes/p7`](notes/p7/README.md) |
| 16 | P12 | 审查后真机复测:Espryt SSIM 1.0;Magma 1.0 / 0.999999511;P12 定向 CTest 46/46;纹理上传回包等待 1→0(主机机制门,真机墙钟待测);FCL 与跨机 TCP 收官门仍待完成 | [`notes/p12`](notes/p12/README.md) |