From 7839fb9dbdb9235bfff253a06c5a1389af0fb8ee Mon Sep 17 00:00:00 2001 From: Jinwoo-H Date: Fri, 3 Jul 2026 00:37:57 -0400 Subject: [PATCH] Findings + results: three stacked fixes unlock the pipeline (agent-tui 16x, DSR-load p50 161->18.8ms in dev) Co-authored-by: Orca --- notes/terminal-performance-initiative.md | 22 ++++ ...seline-jul02-2026-07-03T03-14-25-248Z.json | 116 ++++++++++++++++++ ...v-blockedfix-2026-07-03T04-36-30-875Z.json | 116 ++++++++++++++++++ 3 files changed, 254 insertions(+) create mode 100644 tools/benchmarks/results/terminal-pipeline-baseline-jul02-2026-07-03T03-14-25-248Z.json create mode 100644 tools/benchmarks/results/terminal-pipeline-dev-blockedfix-2026-07-03T04-36-30-875Z.json diff --git a/notes/terminal-performance-initiative.md b/notes/terminal-performance-initiative.md index f3331744f1b..3bcda9b1a23 100644 --- a/notes/terminal-performance-initiative.md +++ b/notes/terminal-performance-initiative.md @@ -329,6 +329,28 @@ plus the terminal-wait contract tests. Expected effect: removes ~85% of remaining onPtyData cost; combined with the two landed fixes should finally unlock the pipeline toward the renderer's measured 27–117 MB/s. +### 2026-07-03 — pipeline unlocked: three stacked fixes, 16× throughput, 9× latency + +Dev-build bench with all three fixes (parse-clocked drains 9e8bb2243, +windowed tail 4e08a28cd, throttled blocked-check 66f20258e), label +dev-blockedfix, same protocol/config as prior dev rows: + +| metric | pre-fix dev | +tail fix | +blocked fix | +|---|---|---|---| +| agent-tui MB/s | 0.7 | 1.0 | **11.5** | +| DSR load p50/p99 (ms) | 161 / 624 | 108 / 154 | **18.8 / 24.9** | +| DSR idle p50/p99 (ms) | 0.95 / 21 | 1.09 / 18 | **0.52 / 8.6** | +| ascii-log MB/s | 6.4 | 4.7 | **9.6** | + +The agent-TUI-specific penalty is gone (agent-tui ≈ cjk ≈ ascii now). The +throttled blocked-check delivered the predicted ~85% cut. Dev mode carries +~2× overhead vs prod, so the prod build should land near ~10ms DSR-under- +load — from the 134ms baseline (~13×) — pending a packaged-build rerun. +Remaining floor is structural cadence (8ms daemon batch + 4ms HP drain +ticks + xterm 12ms slices), which flow control (#6) does not target; +re-evaluate the "within 10× of Terminal.app" goal line after a prod +measurement. Next: term-speed-2 revival (#4), then flow control (#6). + ## Success criteria (baseline-relative; finalize after task 1) - DSR-under-load p90 in Orca within striking distance of iTerm2 on the same diff --git a/tools/benchmarks/results/terminal-pipeline-baseline-jul02-2026-07-03T03-14-25-248Z.json b/tools/benchmarks/results/terminal-pipeline-baseline-jul02-2026-07-03T03-14-25-248Z.json new file mode 100644 index 00000000000..102f7c07f9e --- /dev/null +++ b/tools/benchmarks/results/terminal-pipeline-baseline-jul02-2026-07-03T03-14-25-248Z.json @@ -0,0 +1,116 @@ +{ + "benchmark": "terminal-pipeline", + "label": "baseline-jul02", + "timestamp": "2026-07-03T03:14:25.248Z", + "terminal": { + "program": "ghostty", + "version": "1.3.1", + "term": "xterm-ghostty", + "cols": 195, + "rows": 50, + "tty": true + }, + "machine": { + "platform": "darwin", + "arch": "arm64", + "cpu": "Apple M4", + "release": "25.2.0", + "node": "v26.0.0" + }, + "config": { + "sizeMb": 10, + "iterations": 5, + "dsrCount": 200 + }, + "dsrIdle": { + "count": 200, + "p50": 0.15908399999989342, + "p90": 0.2396659999999997, + "p99": 0.4069170000000213, + "max": 0.5705420000000032, + "timeouts": 0 + }, + "throughput": [ + { + "fixture": "ascii-log", + "fixtureBytes": 10485767, + "wallMs": { + "count": 5, + "p50": 126.93312500000002, + "p90": 251.20050000000015, + "p99": 251.20050000000015, + "max": 251.20050000000015 + }, + "mbPerSec": { + "count": 5, + "p50": 78.78169450031436, + "p90": 84.26504436105196, + "p99": 84.26504436105196, + "max": 84.26504436105196 + } + }, + { + "fixture": "cjk-emoji", + "fixtureBytes": 10485782, + "wallMs": { + "count": 5, + "p50": 104.4577919999997, + "p90": 109.85120899999993, + "p99": 109.85120899999993, + "max": 109.85120899999993 + }, + "mbPerSec": { + "count": 5, + "p50": 95.73264750642048, + "p90": 106.42174244458813, + "p99": 106.42174244458813, + "max": 106.42174244458813 + } + }, + { + "fixture": "agent-tui", + "fixtureBytes": 10486087, + "wallMs": { + "count": 5, + "p50": 127.35175000000027, + "p90": 135.539084, + "p99": 135.539084, + "max": 135.539084 + }, + "mbPerSec": { + "count": 5, + "p50": 78.52512314515855, + "p90": 81.831482788438, + "p99": 81.831482788438, + "max": 81.831482788438 + } + }, + { + "fixture": "styles-stress", + "fixtureBytes": 10485775, + "wallMs": { + "count": 5, + "p50": 95.08958299999995, + "p90": 104.82441699999981, + "p99": 104.82441699999981, + "max": 104.82441699999981 + }, + "mbPerSec": { + "count": 5, + "p50": 105.16414090400156, + "p90": 105.89837321603056, + "p99": 105.89837321603056, + "max": 105.89837321603056 + } + } + ], + "dsrUnderLoad": { + "paceBytesPerSec": 1048576, + "count": 50, + "p50": 0.17150000000037835, + "p90": 0.35129200000119454, + "p99": 2.7332079999996495, + "max": 2.7332079999996495, + "timeouts": 0 + } +} diff --git a/tools/benchmarks/results/terminal-pipeline-dev-blockedfix-2026-07-03T04-36-30-875Z.json b/tools/benchmarks/results/terminal-pipeline-dev-blockedfix-2026-07-03T04-36-30-875Z.json new file mode 100644 index 00000000000..bfccf292f16 --- /dev/null +++ b/tools/benchmarks/results/terminal-pipeline-dev-blockedfix-2026-07-03T04-36-30-875Z.json @@ -0,0 +1,116 @@ +{ + "benchmark": "terminal-pipeline", + "label": "dev-blockedfix", + "timestamp": "2026-07-03T04:36:30.875Z", + "terminal": { + "program": "Orca", + "version": "1.4.120-rc.2", + "term": "xterm-256color", + "cols": 282, + "rows": 80, + "tty": true + }, + "machine": { + "platform": "darwin", + "arch": "arm64", + "cpu": "Apple M4", + "release": "25.2.0", + "node": "v26.0.0" + }, + "config": { + "sizeMb": 3, + "iterations": 5, + "dsrCount": 200 + }, + "dsrIdle": { + "count": 200, + "p50": 0.5164159999999924, + "p90": 1.431041999999934, + "p99": 8.555624999999964, + "max": 20.349042000000004, + "timeouts": 0 + }, + "throughput": [ + { + "fixture": "ascii-log", + "fixtureBytes": 3145851, + "wallMs": { + "count": 5, + "p50": 313.5015000000003, + "p90": 323.46779100000003, + "p99": 323.46779100000003, + "max": 323.46779100000003 + }, + "mbPerSec": { + "count": 5, + "p50": 9.569706371232401, + "p90": 10.165677880605026, + "p99": 10.165677880605026, + "max": 10.165677880605026 + } + }, + { + "fixture": "cjk-emoji", + "fixtureBytes": 3145785, + "wallMs": { + "count": 5, + "p50": 251.9332089999998, + "p90": 254.40795800000024, + "p99": 254.40795800000024, + "max": 254.40795800000024 + }, + "mbPerSec": { + "count": 5, + "p50": 11.90813379206406, + "p90": 12.268678601770057, + "p99": 12.268678601770057, + "max": 12.268678601770057 + } + }, + { + "fixture": "agent-tui", + "fixtureBytes": 3146560, + "wallMs": { + "count": 5, + "p50": 261.10541699999976, + "p90": 321.0371669999995, + "p99": 321.0371669999995, + "max": 321.0371669999995 + }, + "mbPerSec": { + "count": 5, + "p50": 11.492651096668945, + "p90": 11.992095098957183, + "p99": 11.992095098957183, + "max": 11.992095098957183 + } + }, + { + "fixture": "styles-stress", + "fixtureBytes": 3145734, + "wallMs": { + "count": 5, + "p50": 589.9538329999996, + "p90": 600.6057920000003, + "p99": 600.6057920000003, + "max": 600.6057920000003 + }, + "mbPerSec": { + "count": 5, + "p50": 5.0851533700365, + "p90": 5.580751006417512, + "p99": 5.580751006417512, + "max": 5.580751006417512 + } + } + ], + "dsrUnderLoad": { + "paceBytesPerSec": 1048576, + "count": 50, + "p50": 18.74962499999856, + "p90": 23.03216699999939, + "p99": 24.845541000000594, + "max": 24.845541000000594, + "timeouts": 0 + } +}