diff --git a/Cargo.toml b/Cargo.toml
index 8da5e69f..1a5467a7 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -31,6 +31,11 @@ name = "reshape_observation"
 harness = false
 test = false
 
+[[bench]]
+name = "scale_bisect"
+harness = false
+test = false
+
 [dependencies]
 gp-tree = { path = "crates/gp-tree" }
 improvise-core = { path = "crates/improvise-core" }
diff --git a/bench/edge.py b/bench/edge.py
new file mode 100644
index 00000000..f5e199e5
--- /dev/null
+++ b/bench/edge.py
@@ -0,0 +1,8 @@
+import sys; sys.path.insert(0,'bench')
+from scale_bisect import run
+cfgs=[dict(k=k,prune=0) for k in (16,17,18,26,27)]+[dict(k=k,prune=1) for k in (7,11,12,13,14,21,22)]+[dict(k=1000,n=n,mode='records') for n in (1000,1200,1500,1800,2000,4600,14500)]
+for c in cfgs:
+    c.setdefault('n',100000); c['ops']='open,key,window'
+    v=[run('/tmp/sb',c,timeout=200) for _ in range(3)]
+    f=lambda k:'/'.join(f"{r[k][1]:.1f}" if k in r else '-' for r in v)
+    print({k:c[k] for k in c if k!='ops'},'rows',v[0].get('rows'),'open',f('open'),'key',f('keypress_down'),'win',f('browser_grid_of_top'),'tuiframe',f('tui_frame_grid_of_whole'),flush=True)
diff --git a/bench/full.py b/bench/full.py
new file mode 100644
index 00000000..27d79581
--- /dev/null
+++ b/bench/full.py
@@ -0,0 +1,10 @@
+import sys, json; sys.path.insert(0,'bench')
+from scale_bisect import run
+cfgs=[dict(k=9,prune=0),dict(k=10,prune=0),dict(k=17,prune=0,groups=1),dict(k=26,prune=0,groups=1),
+      dict(k=8,prune=1),dict(k=13,prune=1,groups=1),
+      dict(k=1000,n=1300,mode='records'),dict(k=1000,n=4400,mode='records'),dict(k=1000,n=14000,mode='records')]
+for c in cfgs:
+    c.setdefault('n',100000)
+    for rep in range(3):
+        r=run('/tmp/sb',c,timeout=300)
+        print(json.dumps({'cfg':c,'rep':rep,**r}),flush=True)
diff --git a/bench/scale_bisect.py b/bench/scale_bisect.py
new file mode 100644
index 00000000..9d576cfa
--- /dev/null
+++ b/bench/scale_bisect.py
@@ -0,0 +1,38 @@
+"""Drive benches/scale_bisect one configuration per process; parse its lines.
+usage: scale_bisect.py BIN k=.. n=.. [cats=..] [prune=..] [mode=..] [ops=..] [--timeout S]
+"""
+import subprocess, sys, re, json
+
+def run(binary, cfg, timeout=90):
+    args = [binary] + [f"{k}={v}" for k, v in cfg.items()]
+    try:
+        p = subprocess.run(args, capture_output=True, text=True, timeout=timeout)
+        out, status = p.stdout, "ok"
+    except subprocess.TimeoutExpired as e:
+        out, status = (e.stdout or b"").decode() if isinstance(e.stdout, bytes) else (e.stdout or ""), "timeout"
+    res = {"status": status}
+    for line in out.splitlines():
+        if line.startswith("REFUSED"):
+            res["status"] = "refused"
+        m = re.match(r"(\w+) min=([\d.]+) med=([\d.]+) max=([\d.]+)", line)
+        if m:
+            res[m[1]] = (float(m[2]), float(m[3]), float(m[4]))
+        for kv in re.findall(r"(\w+)=([\d.]+)", line):
+            if kv[0] in ("rows", "cols", "maxrss_mb", "improv_mb", "snapshot_kb", "rss_after_open_mb", "cells"):
+                res[kv[0]] = float(kv[1])
+        m = re.match(r"json_encode_setcell_updates ms=(.*)", line)
+        if m:
+            res["json_setcell"] = m[1]
+    return res
+
+if __name__ == "__main__":
+    binary = sys.argv[1]
+    cfg, timeout = {}, 90
+    rest = sys.argv[2:]
+    for a in rest:
+        if a.startswith("--timeout"):
+            timeout = int(a.split("=")[1])
+        else:
+            k, v = a.split("=", 1)
+            cfg[k] = v
+    print(json.dumps(run(binary, cfg, timeout)))
diff --git a/bench/sweep.py b/bench/sweep.py
new file mode 100644
index 00000000..f0c166d6
--- /dev/null
+++ b/bench/sweep.py
@@ -0,0 +1,10 @@
+import sys; sys.path.insert(0,'bench')
+from scale_bisect import run
+import itertools
+OPS=sys.argv[1]; 
+cfgs=eval(sys.argv[2])
+cols=sys.argv[3].split(',')
+for c in cfgs:
+    c=dict(c); c['ops']=OPS
+    r=run('/tmp/sb',c,timeout=int(sys.argv[4]) if len(sys.argv)>4 else 90)
+    print(c, r['status'], 'rows',r.get('rows'), *[f"{k}={r[k][1]:.1f}" if k in r else f"{k}=-" for k in cols], f"rss={r.get('maxrss_mb')}", flush=True)
diff --git a/benches/scale_bisect.rs b/benches/scale_bisect.rs
new file mode 100644
index 00000000..9b559514
--- /dev/null
+++ b/benches/scale_bisect.rs
@@ -0,0 +1,266 @@
+//! One configuration per process: build a sparse document, time the real paths.
+//! `scale_bisect k=1000 n=100000 cats=3 prune=1 mode=grid`; see context/test-evidence/scale-bisect.md.
+
+use std::collections::{HashMap, HashSet};
+use std::io::Write;
+use std::time::Instant;
+
+use improvise::model::cell::{CellKey, CellValue};
+use improvise::view::Axis;
+use improvise::workbook::Workbook;
+use improvise_command::ui::app::App;
+use improvise_command::ui::effect::{self, ViewTarget};
+use improvise_command::ui::host::Host;
+use improvise_command::ui::project::{self, Viewport};
+use improvise_command::ui::reduce::caught_up;
+use improvise_io::persistence::{format_md, parse_md, save_to};
+use improvise_io::storage::MemoryFs;
+use improvise_protocol::{
+    ClientFrame, KeyCode, KeyEvent, KeyModifiers, ModelCommand, RecordHash, render::RecordCell,
+};
+
+fn lcg(s: &mut u64, n: usize) -> usize {
+    *s = s
+        .wrapping_mul(6_364_136_223_846_793_005)
+        .wrapping_add(1_442_695_040_888_963_407);
+    (*s >> 33) as usize % n
+}
+
+struct Cfg {
+    k: usize,
+    n: usize,
+    cats: usize,
+    prune: bool,
+    mode: String,
+    groups: bool,
+    ops: String,
+}
+
+fn arg(args: &HashMap<String, String>, k: &str, d: &str) -> String {
+    args.get(k).cloned().unwrap_or_else(|| d.to_string())
+}
+
+fn document(c: &Cfg) -> (Workbook, Vec<CellKey>) {
+    let mut wb = Workbook::new("bench");
+    for ci in 0..c.cats {
+        let cat = format!("C{ci}");
+        wb.add_category(cat.clone()).expect("under the limit");
+        for i in 0..c.k {
+            if c.groups && ci == 0 {
+                wb.add_item_in_group(&cat, format!("i{i}"), format!("G{}", i / 10))
+                    .expect("added");
+            } else {
+                wb.add_item(&cat, format!("i{i}")).expect("added");
+            }
+        }
+    }
+    for m in ["Cost", "Rev"] {
+        wb.add_item("_Measure", m).expect("virtual");
+    }
+    let space = (c.k as f64).powi(c.cats as i32);
+    let wanted = c.n.min((space / 2.0).max(1.0) as usize);
+    let (mut s, mut seen, mut recs) = (0x9E37_79B9_7F4A_7C15u64, HashSet::new(), Vec::new());
+    while seen.len() < wanted {
+        let d: Vec<usize> = (0..c.cats).map(|_| lcg(&mut s, c.k)).collect();
+        if !seen.insert(d.clone()) {
+            continue;
+        }
+        let coords: Vec<(String, String)> = d
+            .iter()
+            .enumerate()
+            .map(|(ci, x)| (format!("C{ci}"), format!("i{x}")))
+            .collect();
+        recs.push(CellKey::new(coords.clone()));
+        for (m, v) in [("Cost", 3.0), ("Rev", 5.0)] {
+            let mut key = coords.clone();
+            key.push(("_Measure".to_string(), m.to_string()));
+            wb.model.set_cell(CellKey::new(key), CellValue::Number(v));
+        }
+    }
+    let view = wb.active_view_mut();
+    for ci in 0..c.cats {
+        let axis = if c.mode == "records" || c.mode == "none" { Axis::None } else { Axis::Row };
+        view.set_axis(&format!("C{ci}"), axis);
+    }
+    if c.mode == "records" {
+        view.set_axis("_Index", Axis::Row);
+        view.set_axis("_Dim", Axis::Column);
+    } else {
+        view.set_axis("_Measure", Axis::Column);
+    }
+    view.prune_empty = c.prune;
+    (wb, recs)
+}
+
+fn maxrss_mb() -> f64 {
+    let mut u: libc::rusage = unsafe { std::mem::zeroed() };
+    unsafe { libc::getrusage(libc::RUSAGE_SELF, &mut u) };
+    u.ru_maxrss as f64 / 1048576.0
+}
+
+fn rss_mb() -> f64 {
+    let out = std::process::Command::new("ps")
+        .args(["-o", "rss=", "-p", &std::process::id().to_string()])
+        .output()
+        .expect("ps");
+    String::from_utf8_lossy(&out.stdout).trim().parse::<f64>().unwrap_or(0.0) / 1024.0
+}
+
+/// Sample `f` up to `reps` times, stopping once `budget_ms` is spent (one sample minimum).
+fn sample(name: &str, reps: usize, budget_ms: f64, mut f: impl FnMut(usize)) {
+    let mut v = Vec::new();
+    let start = Instant::now();
+    for r in 0..reps {
+        let t = Instant::now();
+        f(r);
+        v.push(t.elapsed().as_secs_f64() * 1000.0);
+        if start.elapsed().as_secs_f64() * 1000.0 > budget_ms {
+            break;
+        }
+    }
+    v.sort_by(|a, b| a.partial_cmp(b).expect("finite"));
+    println!(
+        "{name} min={:.3} med={:.3} max={:.3} reps={}",
+        v[0],
+        v[v.len() / 2],
+        v[v.len() - 1],
+        v.len()
+    );
+    std::io::stdout().flush().ok();
+}
+
+fn once<T>(name: &str, f: impl FnOnce() -> T) -> T {
+    let t = Instant::now();
+    let v = f();
+    println!("{name} min={0:.3} med={0:.3} max={0:.3} reps=1", t.elapsed().as_secs_f64() * 1000.0);
+    std::io::stdout().flush().ok();
+    v
+}
+
+fn key(code: KeyCode) -> KeyEvent {
+    KeyEvent::new(code, KeyModifiers::NONE)
+}
+
+fn main() {
+    let args: HashMap<String, String> = std::env::args()
+        .skip(1)
+        .filter_map(|a| a.split_once('=').map(|(k, v)| (k.to_string(), v.to_string())))
+        .collect();
+    let c = Cfg {
+        k: arg(&args, "k", "1000").parse().expect("k"),
+        n: arg(&args, "n", "100000").parse().expect("n"),
+        cats: arg(&args, "cats", "3").parse().expect("cats"),
+        prune: arg(&args, "prune", "1") == "1",
+        mode: arg(&args, "mode", "grid"),
+        groups: arg(&args, "groups", "0") == "1",
+        ops: arg(&args, "ops", "open,key,frame,window,setcell,host,reshape,io"),
+    };
+    let rows_cap: f64 = arg(&args, "rows_cap", "30000000").parse().expect("cap");
+    let product = (c.k as f64).powi(c.cats as i32);
+    if c.mode == "grid" && product > rows_cap {
+        println!("REFUSED product_rows={product:.3e} (> {rows_cap:.0}: layout would materialize it)");
+        return;
+    }
+    let want = |o: &str| c.ops.split(',').any(|x| x == o);
+    let (wb, recs) = document(&c);
+    println!("cells={} records={} base_rss_mb={:.0}", wb.model.data.iter_cells().count(), recs.len(), rss_mb());
+    let mut app = once("open", || App::new(wb, None));
+    println!("rows={} cols={} rss_after_open_mb={:.0}", app.layout.row_count(), app.layout.col_count(), rss_mb());
+    app.term_width = 200;
+    app.term_height = 50;
+    let name = app.view_name().to_string();
+
+    if want("key") {
+        sample("keypress_down", 8, 6000.0, |_| {
+            app.dispatch_key(key(KeyCode::Down));
+        });
+    }
+    if want("frame") {
+        let whole = |app: &App| Viewport { row_offset: 0, col_offset: 0, rows: app.layout.row_count(), cols: app.layout.col_count() };
+        sample("tui_frame_grid_of_whole", 3, 8000.0, |_| {
+            std::hint::black_box(project::grid_of(&app, whole(&app)));
+        });
+        let cache = project::grid_of(&app, whole(&app));
+        let area = ratatui::layout::Rect::new(0, 0, 200, 50);
+        sample("tui_render_gridwidget_buffer", 5, 4000.0, |_| {
+            let mut buf = ratatui::buffer::Buffer::empty(area);
+            let w = improvise::ui::grid::GridWidget::new(
+                &cache, app.cursor(), &app.view_state.active_view, &app.layout,
+                &app.view_state.mode, &app.view_state.search_query,
+                &app.view_state.buffers, app.drill_state.as_ref(),
+            );
+            ratatui::widgets::Widget::render(w, area, &mut buf);
+            std::hint::black_box(buf);
+        });
+    }
+    if want("window") {
+        let mid = app.layout.row_count() / 2;
+        for (label, off) in [("top", 0), ("mid", mid)] {
+            let vp = Viewport { row_offset: off, col_offset: 0, rows: 45, cols: 12 };
+            sample(&format!("browser_grid_of_{label}"), 5, 6000.0, |_| {
+                std::hint::black_box(project::grid_of(&app, vp));
+            });
+        }
+    }
+    if want("setcell") && !recs.is_empty() {
+        let existing = recs[recs.len() / 2].clone().with("_Measure", "Rev");
+        sample("setcell_existing", 5, 6000.0, |r| {
+            app.apply_effects(vec![Box::new(effect::SetCell(existing.clone(), CellValue::Number(10.0 + r as f64)))]);
+        });
+        let mut s = 99u64;
+        sample("setcell_new_record", 5, 6000.0, |_| {
+            let mut coords: Vec<(String, String)> = (0..c.cats).map(|ci| (format!("C{ci}"), format!("i{}", lcg(&mut s, c.k)))).collect();
+            coords.push(("_Measure".into(), "Rev".into()));
+            app.apply_effects(vec![Box::new(effect::SetCell(CellKey::new(coords), CellValue::Number(1.0)))]);
+        });
+    }
+    if want("reshape") && c.mode == "grid" {
+        let v = || ViewTarget::Active;
+        sample("reshape_hide_item", 3, 6000.0, |r| {
+            app.apply_effects(vec![Box::new(effect::HideItem { view: v(), category: "C1".into(), item: format!("i{r}") })]);
+        });
+        if c.groups {
+            sample("reshape_toggle_group", 3, 6000.0, |_| {
+                app.apply_effects(vec![Box::new(effect::ToggleGroup { view: v(), category: "C0".into(), group: "G0".into() })]);
+            });
+        }
+        let last = format!("C{}", c.cats - 1);
+        sample("reshape_set_axis", 4, 6000.0, |r| {
+            let axis = if r % 2 == 0 { Axis::Page } else { Axis::Row };
+            app.apply_effects(vec![Box::new(effect::SetAxis { view: v(), category: last.clone(), axis })]);
+        });
+        sample("reshape_toggle_prune", 2, 6000.0, |_| {
+            app.apply_effects(vec![Box::new(effect::TogglePruneEmpty(v()))]);
+        });
+    }
+    if want("io") {
+        let wbr = &app.model_state.workbook;
+        let text = once("format_md", || format_md(wbr));
+        println!("improv_mb={:.2}", text.len() as f64 / 1048576.0);
+        once("parse_md", || parse_md(&text).expect("parses"));
+        let fs = MemoryFs::new();
+        sample("persist_per_command_save_to", 3, 6000.0, |_| {
+            save_to(&fs, wbr, std::path::Path::new("d.improv")).expect("write");
+        });
+    }
+    if want("host") && !recs.is_empty() {
+        let mut host = Host::new(app);
+        let (id, snap) = once("host_connect_snapshot", || host.connect(45, 12));
+        sample("json_encode_snapshot", 3, 4000.0, |_| {
+            std::hint::black_box(serde_json::to_vec(&snap.msg).expect("json"));
+        });
+        println!("snapshot_kb={:.1}", serde_json::to_vec(&snap.msg).expect("json").len() as f64 / 1024.0);
+        let rec = recs[recs.len() / 2].clone();
+        let cell = RecordCell { record: RecordHash::of(&rec), measure: Some("Rev".into()) };
+        let mut enc: Vec<(f64, usize)> = Vec::new();
+        sample("host_handle_setcell_and_project", 4, 8000.0, |r| {
+            let wire = caught_up(host.app(), ModelCommand::SetCell { view: name.clone(), cell: cell.clone(), value: CellValue::Number(50.0 + r as f64) });
+            let out = host.handle(id, ClientFrame::command(wire));
+            let t = Instant::now();
+            let bytes: usize = out.iter().map(|a| serde_json::to_vec(&a.msg).expect("json").len()).sum();
+            enc.push((t.elapsed().as_secs_f64() * 1000.0, bytes));
+        });
+        println!("json_encode_setcell_updates ms={:?}", enc);
+    }
+    println!("maxrss_mb={:.0}", maxrss_mb());
+}
diff --git a/context/test-evidence/scale-bisect.md b/context/test-evidence/scale-bisect.md
new file mode 100644
index 00000000..7bf797d6
--- /dev/null
+++ b/context/test-evidence/scale-bisect.md
@@ -0,0 +1,218 @@
+# Scale bisect: how big a document does improvise handle today (improvise-felo.10)
+
+Measured on `gh/main` @ 8220a065, **Apple M4 Max, 64 GB** (`sysctl`: `hw.memsize 68719476736`), macOS,
+release build (`lto`, `opt-level 3`), native only (no wasm, no real OPFS, no terminal).
+Harness: `benches/scale_bisect.rs` (one configuration per process) driven by `bench/scale_bisect.py`,
+`bench/sweep.py`, `bench/edge.py`, `bench/full.py`. Reproduce:
+`nix develop --command cargo build --release --bench scale_bisect`, copy the binary from
+`target/release/deps/scale_bisect-*` to `/tmp/sb`, then `python3 bench/edge.py`.
+
+Budgets (user ruling 2026-10-03): **8 ms is the headline** (120 FPS, hot path or defer to background);
+16 ms, 100 ms (noticeable) and 1 s (broken) are reference crossings.
+
+## Method, and what is real
+
+Every number goes through the production path: `App::new` (open: formula recompute +
+`GridLayout::with_frozen_records`), `App::dispatch_key(Down)` (**keypress/scroll**: it calls
+`rebuild_layout`, then `cmd_context`, which runs `compute_col_widths` over the whole grid, then applies effects),
+`project::grid_of` (TUI frame: whole grid; browser: a 45x12 window), `App::apply_effects` with `effect::SetCell`,
+`SetAxis`/`HideItem`/`ToggleGroup`/`TogglePruneEmpty` (reshape), `Host::connect` and `Host::handle` with a
+`caught_up` `ModelCommand::SetCell` (browser path incl. projection), `format_md`/`parse_md`/`save_to(MemoryFs)`.
+Document: C0..C{cats-1} with K items, N distinct random tuples ("records", capped at K^cats/2), two measures
+each (2N cells), `_Measure` on Column. Timings are the **median** of up to 3-8 in-process repeats; boundary
+rows are the median of **3 separate processes** run sequentially (spread shown). Earlier exploratory sweeps
+ran 2-3 processes in parallel and agree with the sequential runs within ~5%.
+
+**`View::prune_empty` defaults to false** (`view/types.rs`), so "prune off" is what a user gets until they
+toggle it; "prune on" is reported next to it. Rows below are the layout's `row_count`.
+
+**Dominant cost (reasoned, not profiled):** the release profile strips symbols, so `sample` could not
+symbolicate. The reason is in the code and the timings agree with it: `GridLayout::cell_key`/`row_item` do
+`data_items(&self.row_items).nth(row)` (`view/layout.rs` ~253/292/807/870), O(row) per lookup, and
+`compute_col_widths` (every keypress, via `cmd_context`) and `grid_of` visit every cell, so both are
+O(rows^2). Evidence: with prune off, rows 729 -> 5.4 ms, 4096 -> 74.6 ms, 19683 -> 1248 ms (x6.7 rows ->
+x14 time -> x17 time). A 45x12 browser window costs **the same** as the whole-grid TUI frame (K=30:
+2274 vs 2298 ms) because the widths pass is whole-grid whatever the viewport says, so viewport narrowing buys
+nothing today. `GridLayout` build itself is cheap (~90 ns/row, ADR 008).
+
+## Attribution: state+projection vs render/serialize (K=10, 1000 rows, prune off; 3 runs)
+
+| step | ms |
+|---|---|
+| keypress (state: rebuild + widths + effect) | 9.5-9.7 |
+| `grid_of` whole grid (TUI projection) | 11.7-12.0 |
+| `GridWidget` render into a 200x50 ratatui `Buffer` | 0.12 |
+| `grid_of` 45x12 window (browser projection) | 9.1-9.3 |
+| JSON encode of one SetCell's projected updates | 0.001-0.005 (0.6-1.1 KB) |
+| JSON encode of the connect Snapshot | 0.02 (21 KB) |
+
+Render and serialize are **<=0.5 ms in every row measured** (GridWidget 0.08-0.53 ms up to 14,500 records;
+snapshot encode <=0.07 ms; records-mode SetCell updates 21 KB in 0.02 ms). The 8 ms is entirely state and
+projection. Not measured: the rest of `draw()` (panels, status; `draw` is private to the binary), terminal I/O,
+browser JSON *decode* and DOM paint.
+
+## Axis B: items per category K, 3 categories on Row, N = 100k (capped at K^3/2), prune OFF
+
+| K | rows | open | keypress | browser window | SetCell | maxrss MB |
+|---|---|---|---|---|---|---|
+| 5 | 125 | 0.1 | 0.5 | 0.5 | 0.2 | 8 |
+| 9 | 729 | 0.2 | **5.4** (5.3-5.5) | 5.0 | 0.9 | 13 |
+| 10 | 1,000 | 0.2 | **9.6** (9.5-9.7) | 9.2 | 1.2 | 14 |
+| 11 | 1,331 | 0.3 | 12.7 | 12.7 | 1.6 | 10 |
+| 12 | 1,728 | 0.3 | 20.0 | 19.3 | 2.1 | 11 |
+| 15 | 3,375 | 0.5 | 65 | 64 | 4.0 | 14 |
+| 16 | 4,096 | 0.6 | 74.6 (74.3-75.2) | 73.4 | - | - |
+| 17 | 4,913 | 0.7 | **105** (105.0-105.5) | 103 | - | - |
+| 20 | 8,000 | 1.1 | 302 | 301 | 9.4 | 21 |
+| 25 | 15,625 | 1.9 | 773 | 771 | 18.8 | 33 |
+| 26 | 17,576 | 2.2 | 967-977 | 962 | - | - |
+| 27 | 19,683 | 2.4 | **1,248** (1,244-1,251) | 1,244 | - | - |
+| 30 | 27,000 | 3.1 | 2,262 | 2,274 | 31 | 49 |
+| 40 | 64,000 | 6.7 | 10,858 | 10,833 | 73 | 235 |
+| 60 | 216,000 | 21.6 | timeout (>60 s) | - | - | - |
+| 1000 | 10^9 | **refused** (see below) | | | | |
+
+Hot-path keypress/scroll and browser window: **largest inside 8 ms: K=9 (729 rows)**; inside 16 ms: K=11;
+first over 100 ms: **K=17 (4,913 rows)**; first over 1 s: **K=27 (19,683 rows)**. SetCell alone
+(no widths pass; layout rebuild only, ~linear in rows): inside 8 ms to K~19 (K=15 4.0, K=20 9.4), 16 ms ~K=24,
+100 ms ~K=45 (K=40 = 73 ms), 1 s extrapolated ~K=95 (**extrapolated, not measured**). Open stays <25 ms to K=60.
+
+**3 x 1,000 on Row cannot be opened at any N**: the layout materializes the cross product (10^9 rows, ~150 B
+each = ~150 GB, ADR 008). The harness refuses above 30M rows (`rows_cap`) rather than swap the machine, so
+the 10^9 number is the ADR's estimate, not a measurement here.
+
+### Prune ON (same documents)
+
+| K | rows | open | keypress | browser window | SetCell |
+|---|---|---|---|---|---|
+| 7 | 171 | 2.0 | 4.5 (4.4-4.6) | 1.1 | 5.4 |
+| 8 | 256 | 3.0 | **7.0** (7.0-7.1) | 1.6 | 8.5 |
+| 9 | 364 | 5.3 | 12.8 | 2.7 | 15.3 |
+| 11 | 665 | 12.2 | 29.6 | 6.2 | ~28 |
+| 13 | 1,098 | 29.0 | 71 | 14.5 | ~70 |
+| 14 | 1,372 | 48 | **119** | 24 | - |
+| 20 | 4,000 | 300 | 747 | 145 | 903 |
+| 21 | 4,630 | 401 | 968-992 | 190 | - |
+| 25 | 7,812 | 767 | 1,899 | 366 | 2,328 |
+| 30 | 13,500 | 2,255 | 5,555 | 1,047 | 6,804 |
+
+Prune on: keypress inside 8 ms to **K=8**, first over 100 ms **K=14**, ~1 s at **K~21-22** (K=22 measured
+0.93-0.95 s, a different N cap; first clearly over: K=25). Pruning shrinks the grid, so the browser window is
+cheaper than prune off at the same K (K=20: 145 ms against 302 ms), but **every keypress re-runs
+the prune** (`dispatch_key` calls `rebuild_layout`), so keypress and SetCell are *worse* with prune on
+(K=10: 23 vs 9.9 ms). Prune is also quadratic (`nth`). Conclusion: prune on/off moves the boundary by about
+one K at 8 ms and makes 100 ms/1 s *earlier* (K=14/21 vs 17/27); it never rescues the 3x1000 case.
+
+## Axis A: records N at fixed K (K=30, 27,000 rows)
+
+| N | prune off keypress | prune on rows | prune on keypress | prune on window |
+|---|---|---|---|---|
+| 100 | 451 | 100 | 861 | 0.3 |
+| 1,000 | 527 | 1,000 | 1,059 | 6.4 |
+| 5,000 | 1,013 | 5,000 | 2,178 | 140 |
+| 13,500 (cap) | 2,267 | 13,500 | 5,555 | 1,047 |
+
+Prune off: ~450 ms of widths over 27k empty rows + ~1.3 us per stored cell; **N never gets the keypress
+under 100 ms at K=30**, and at K=1000 N is irrelevant because the layout cannot be built (rows = K^3).
+The N-axis has no bisect point in the 3x1000 regime; N only matters once K is small enough that rows fit, and
+there it is a second-order term. Open is flat (3.1-3.6 ms) in N with prune off.
+
+## Axis C: categories on Row (prune off)
+
+| K | cats | rows | open | keypress | note |
+|---|---|---|---|---|---|
+| 10 | 2 | 100 | 0.1 | 0.2 | |
+| 10 | 3 | 1,000 | 0.2 | 9.9 | over 8 ms |
+| 10 | 4 | 10,000 | 1.4 | **880** | ~1 s |
+| 10 | 5 | 100,000 | 14 | timeout >100 s | |
+| 3 | 6 | 729 | 0.3 | 27.9 | 5x worse than K=9 x 3 cats at the same rows |
+| 3 | 8 | 6,561 | 2.0 | 2,798 | |
+| 3 | 10-12 | 59k-531k | 18-210 | timeout | open alone passes 100 ms at 11 cats |
+| 2 | 10 | 1,024 | 0.6 | 172 | |
+| 2 | 12 | 4,096 | 2.2 | 3,190 | |
+| 2 | 14 | 16,384 | 9 | 63,100 | |
+
+The cap of 24 is unreachable: rows = product of items, so a view with c Row categories of K items dies by
+K^c ~ 10^4 rows (1 s) regardless of N, and with two items per category by 12-14 categories. Per-row cost
+also grows with category count (each row_item is a Vec of c items). Not measured: Row categories with prune
+on, other axes absorbing categories (the realistic mitigation: only 1-2 on Row).
+
+## Axis D: records mode over N records (`with_frozen_records`, the improvise-sk32 quadratic)
+
+| N | open | keypress | browser window | SetCell (apply) | host SetCell+project | maxrss MB |
+|---|---|---|---|---|---|---|
+| 300 | 1.5 | 0.7 | 0.1 | 1.6 | - | 10 |
+| 1,000 | 2.9 | 5.6 (5.5-5.7) | 0.3 | 10 | - | 12 |
+| 1,200 | 3.9 | **7.9** (7.7-8.0) | 0.3 | - | - | - |
+| 1,300 | 4.5 | 9.3 | 0.3 | 15.5 | 17 | 26 |
+| 1,800 | 8.4 | 16.5 | 0.4 | - | - | - |
+| 3,000 | 22.9 | 47 | 0.7 | 74 | - | 20 |
+| 4,400-4,600 | 47-51 | **95-102** | 0.9 | 149 | 158 | 55 |
+| 10,000 | 248 | 492 | 1.8 | 734 | - | 46 |
+| 14,000-14,500 | 472-500 | **940-1,002** | 2.6-2.8 | 1,435 | 1,459 | 135-148 |
+| 30,000 | 2,198 | 4,314 | 5.6 | 6,467 | - | 125 |
+| 100,000 | **25,420** | **51,234** | 19.3 | not reached (killed at 100 s) | - | not reached |
+
+Records mode, keypress/scroll: inside 8 ms **to N~1,200**; 16 ms N~1,800; 100 ms N~4,500 (first over: 4,600 =
+102 ms); 1 s N~14,500. Open crosses 8 ms ~1,750, 16 ms ~2,300, 100 ms ~6,500, 1 s ~20,000. Time ~ N^1.95.
+SetCell is ~1.5x keypress (it rebuilds and re-groups). The *browser window* is cheap (<20 ms even at 100k)
+because records mode has two columns and the width walk is short, so the 51 s is `rebuild_layout`, not
+projection. Prune on/off does not change records mode (prune applies to pivoted axes; the sweep ran off).
+Records per KB: ~4 KB/record process-wide at 14k (maxrss includes the document generator, so read as an
+upper bound); the 100k memory was not obtained.
+
+## Axis E: save / load and per-command persistence
+
+Mode `none` (all categories off-axis, so layout is trivial and only the store scales), 3 x 1,000, N records
+= 2N cells:
+
+| cells | format_md | text MB | parse_md | persist per accepted command (`save_to` MemoryFs) | maxrss MB |
+|---|---|---|---|---|---|
+| 2,000 | 13.8 | 0.1 | 4.5 | 12.4 | 18 |
+| 20,000 | 156 | 0.86 | 37 | 161 | 80 |
+| 60,000 | 527 | 2.5 | 119 | 539 | 236 |
+| **200,000 (the target)** | **1,974** | 8.4 | **401** | **1,984** | 663 |
+| 600,000 | 6,470 | 25 | 1,320 | 6,460 | 1,894 |
+| 2,000,000 | 23,120 | 84 | 4,668 | 23,088 | 5,611 |
+
+Linear: ~10 us/cell to format, ~2 us/cell to parse. Crossings, format_md: 8 ms ~700 cells, 16 ms ~1.5k,
+100 ms ~11k, 1 s ~100k cells. parse_md: 8 ms ~4k cells, 100 ms ~50k, 1 s ~450k. **At the target (200k cells)
+every accepted command in the browser worker costs ~2 s of format alone** before OPFS write time (not
+measured: MemoryFs only). Persist per command also appears in smaller documents inside the grid axes
+(K=10 grid: 5.5 ms; records N=4,400: 63 ms).
+
+## Other operations (3 runs each; "grouped" fixtures start with C0 groups collapsed, so rows are not K^3)
+
+Reshape (prune off, K=10, 1000 rows): hide item 0.25 ms, set axis 0.19 ms, **toggle prune 12.7 ms**.
+K=26 grouped, 2,028 rows: hide 0.5, toggle group 1.8, set axis 1.8, **toggle prune 727-750 ms**.
+Prune on K=13 grouped, 321 rows: hide 9.4, toggle group 37, set axis 37, toggle prune 37-39 ms.
+`Host::connect` snapshot: K=10 6.5 ms, K=26 375 ms, records N=14,000 2.9 ms. Host SetCell+project: K=10 7.7 ms,
+K=26 ~400 ms, records N=4,400 158 ms (equal to apply + a whole-grid width walk).
+
+## Classification and feasibility of yielding
+
+| operation | class | today's first size over 8 ms | dominated by | chunked/yielding form |
+|---|---|---|---|---|
+| keypress / scroll | **hot (<=8 ms)** | 1,000 rows (K=10) / N=1,300 records | `rebuild_layout` + whole-grid `compute_col_widths` (O(rows^2) `nth`), recomputed though nothing changed | not a slicing problem: cache widths/layout across keys, O(1) row lookup (ADR 008/009). Slicing would only hide it |
+| browser window projection | **hot** | 1,000 rows | same width walk, whole grid regardless of window | needs the same fix; then window cost is O(window) |
+| SetCell apply + project | **hot** | K~19 (prune off), 1,300 records | layout rebuild O(rows) + widths + projection | incremental update of one row/total, not a rebuild |
+| column totals | hot | inside `grid_of` (not separable here) | whole-grid sum in the projection | maintained total (ADR 009) |
+| open / layout build | deferrable | 40k rows (prune off); N~1,750 (records) | build is ~90 ns/row (sliceable by outer index); the prune/records grouping is the quadratic part | build is feasible in slices once the quadratic is removed; the quadratic itself is not sliceable |
+| reshape (hide/collapse/axis) | deferrable | cheap prune-off (<2 ms to ~2k rows); 37 ms prune-on at 321 rows | layout rebuild + prune | one sort/remap at 100k occupied tuples was 2.3 ms in ADR 008's prototype, so a reshape can fit inline; prune toggle needs the quadratic fixed first |
+| save (format_md) | deferrable | ~700 cells | per-cell formatting, linear | the Data section is a flat per-cell line stream, so chunking by cell range with a yield every ~400 cells looks feasible, but needs a consistent snapshot (clone or journal); not profiled by function |
+| load (parse_md) | deferrable | ~4k cells | pest parse + tree walk, linear | tree walk is chunkable by line; the pest parse of one text is not, unless the Data section is parsed line by line (grammar is line oriented); not profiled by function |
+| persistence per command | deferrable (debounce) | ~700 cells | the whole-document `format_md` | coalesce/debounce or append-journal: formatting 2 s per command at target is unusable inline at any slicing |
+
+## What I could not measure
+
+* wasm: all numbers are native; the browser worker is typically slower. Real OPFS I/O, JSON *decode* and
+  DOM paint, the full `draw()` frame, and Safari are not covered.
+* Records mode at 100k: SetCell, host and RSS (killed after open and keypress, ~76 s).
+* Per-function profile: symbols are stripped in `[profile.release]`; attribution is reasoned from the
+  code and the timing exponents, not a flame graph.
+* SetCell 100 ms/1 s for prune off past K=40 are extrapolations.
+* Memory (`maxrss`) includes the generator's hash set and `recs` vector; it is an upper bound on the
+  document. Layout cost ~150 B/row per ADR 008, consistent with K=40 (64k rows, 235 MB peak with prune off
+  including the generator and a second layout during rebuild).
+* Noise: boundary rows are 3 sequential processes, spread <=3% (e.g. K=17: 105.0-105.5 ms). Non-boundary
+  rows are single runs, some taken with 2-3 concurrent processes on a 16+ core machine; treat them as +-5-10%.