Stage H: DoE rebuilt on BIRTH/SEND/KILL, defs-only workloads; work heat floor on contact (FABRIC-3.8 §3.9, §3.13.7)

Workload capsules the DoE births are definitions only; each defines WL-RUN
(its old birth-time work plus its entry word twice) and WL-DONE-ARGS. The
driver births N workers, SENDs each 'WL-RUN WL-DONE-ARGS SEND', waits in an
ordinary loop for N MU-DONE messages, KILLs them (parked idle), and emits
the marker. The serialized VM-EXEC round-robin is gone.

Kernel: no switch checkpoint after a payload's last word; work heat is the
assessed per-beat charge, and a VM gaining work is raised to the busy set's
floor (Captain Bob's ruling) -- fixes a live stall where a busy-waiting Hera
never yielded, without serializing batch starts.

Evidence: cfg-4 rows spread evenly over Hera and all 8 worker slots (44k
switches); full cfg-4 trials on amd64/aarch64/riscv64 end fail=0 with 8/8
KILL dead; Stage G interleave byte-identical on all three. Acceptance:
dict_hash unchanged, relay 3, idle growth 0.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

FABRIC-3.8.md

@@ -715,6 +715,24 @@ a loop. So would the console under load (already recorded under C-b).
 conserved, and still appears in the DoE heat columns; it just no longer decides turns. The
 prediction held on all three ISAs: see the Stage G entry in §6.
 
+**Refined at Stage H, RULED 2026-09-23, Captain Bob: "floor on contact."** Two changes to what
+work heat is, both forced by a live stall (Stage H entry, §6):
+
+- **Assessed, not moved.** Work heat accrues each beat's assessed charge (the slope, in the
+  carry's sub-unit scale), whether or not the rest of the fleet held heat to move.
+  - As first built, it counted only heat actually moved.
+  - A busy-waiting Hera, with nobody else holding absolute heat, then stayed at 0 forever, tied
+    the workers, and never yielded.
+- **Floor on contact.** When a VM with no work gains some (its kernel-Hermes queue goes from
+  empty to one message, and it is neither running nor parked mid-task), its work heat is raised
+  to the lowest work heat among the VMs that have work right now: the running one, and every
+  eligible participant.
+  - Without the floor, a fresh worker at 0 would run alone until it had done as much work as
+    Hera has since boot. The batch would start serially.
+  - No tuning constant.
+  - Rejected alternatives: decay/dissipation (needs a rate constant), and assessed charge
+    without the floor.
+
 #### 3.13.6 Stage C, split
 
 This splits the single C row in §4 in two:
@@ -1226,3 +1244,84 @@ explicit instruction.
       - Hera's work heat grows only when other VMs hold absolute heat for her charges to pull
         from. She runs cheaply while they hold ~0.
       - Both are properties of the ruled quantity, recorded here, not changed.
+      - *(Superseded at Stage H: both consequences above turned out to matter, and the "floor on
+        contact" refinement replaced them. See §3.13.7 and the Stage H entry.)*
+- [x] **Stage H — the DoE rebuilt on ordinary words; "floor on contact" (§3.9, §3.13.7).**
+      2026-09-23. Captain Bob: *"defs-only, no users factor yet, accept PRNG, go ahead with H"*,
+      and then ruled **floor on contact** when the first trial stalled.
+
+      **Changes, capsules:**
+      - The ten workload capsules the DoE births (`workload-0`, `-1-lite`, `-2`, `-3`, `-4`,
+        `-5-lite`, `-6`, `-7`, `-8`, `-9`) are now **definitions only**.
+        - Every top-level line that executed at birth is removed. That includes the heavy
+          self-exec tails and the mid-file `SEED`/wave/`RUN-ASYM`/`RUN-DUAL` runs in 2, 3 and 9.
+        - Each gains `: WL-RUN <exactly what it used to run at birth> <its entry word, twice> ;`.
+          That is the old trial's birth work plus its two `VM-EXEC` passes, now one job.
+        - Each also gains `: WL-DONE-ARGS S" MU-DONE" S" Hera" ;`.
+        - `workload-4` keeps its `4:dopi-ok` print inside `WL-RUN`.
+        - `workload-7` needed a second block for the two definitions. Block 2151 was unclaimed;
+          the conflict gate and `mkcapsule --lint` pass (37 files, 0 violations).
+        - `workload-1` and `-5` (full versions) are untouched, since the DoE doesn't birth them.
+      - `multiuser-doe.4th`:
+        - `WL-ENTRY`, `MU-TOUCH-WORKER/ALL` (the serialized `VM-EXEC` round-robin) and
+          `MU-CHECK-WORKER/ALL` are gone.
+        - New: `MU-DONE-COUNT`/`MU-DONE`, `MU-JOB` (the constant job text `WL-RUN WL-DONE-ARGS
+          SEND`), `MU-SEND-WORKER/ALL`, and `MU-WAIT-ALL` (an ordinary `BEGIN … WHILE REPEAT`;
+          Hera's own checkpoints drain the `MU-DONE`s).
+        - `MU-RUN-TRIAL` is now birth all, send all, wait, kill all, marker.
+        - A refused `SEND` counts as a failure **and** as done, so the wait can't hang on it.
+
+      **Changes, kernel:**
+      - `vm_core.c`: the between-top-level-words checkpoint (C-b) doesn't fire after the
+        **last** word (`vm_input_exhausted()`). The job's final word is the `SEND` to Hera, which
+        makes a decision due. Switching the worker out right there would park it "mid-task" with
+        nothing left, and `KILL` would refuse it. The trampoline decides a moment later anyway.
+      - **Floor on contact** (§3.13.7):
+        - `vm_physics_charge()` accrues the assessed slope into `work_heat_q48`.
+        - New `vm_physics_work_heat_floor()`.
+        - New `sk_vm_switch_signal_note_work_gained()`, called by `sk_hermes_send_one()` when the
+          recipient's queue goes from empty to one message.
+
+      **First trial, before the floor** (`logs/20260923-211253/amd64/`, kept): stalled.
+      - After 8,200+ beats the rows showed `switch_count_cumulative=0` and
+        `switch_current_slot=2` (Hera) throughout.
+      - Hera sat in `MU-WAIT-ALL`, with work heat 0 against workers at 0: a tie, so the
+        incumbent stays, forever.
+
+      **Evidence, as named in §4.**
+      - **Rows** (`logs/20260923-212353/amd64/`, cfg 4 with `HB-ON`, stopped deliberately once
+        captured because per-VM rows at every beat make a full trial take hours):
+        - `switch_count_cumulative` reached 44,297.
+        - Over the first ~7,600 rows, `switch_current_slot` was spread **evenly** across Hera
+          (slot 2: 851) and all eight workers (slots 5–12: 844–846 each).
+      - **Full trials** at cfg 4 (8 workers, uniform `workload-8`), rows off, one per ISA:
+        `logs/20260923-220439/amd64/`, `-222512/aarch64/`, `-231553/riscv64/`.
+        - Each ends with **`MU-TRIAL isa=<isa>run=999 cfg=4 rep=0 nw=8 mode=0 fail=0`**.
+        - **All 8 `KILL: muN dead`, 0 refused.** Every worker ended parked idle and killable.
+        - Times: amd64 ≈ 13 min, aarch64 ≈ 43 min, riscv64 ≈ 23 min.
+      - **Stage G's test, re-run on each boot under the refined rule:** byte-identical on all
+        three, `ABBAABBABAABBAABBABAABBA…` (A = 40, B = 40, **48** runs). It is finer than
+        Stage G's 16 runs because work heat now moves by sub-unit steps every beat.
+
+      **Acceptance,** in the same three boots after the trial:
+      - All reach `[zuse@Hera] ok>` with zero `UNKNOWN WORD`, failures and faults.
+      - The `dict_hash` set is unchanged from Stage G (word_count 533, M7.1a
+        `0x330b28ce0f4e1c75`).
+      - Mint → WIREBIND → `USE` → `1 2 + .` → `3`.
+      - Idle log growth: 0 bytes.
+      - Clean build on all three plus hosted, no new warnings.
+
+      **Known limits, recorded, not changed:**
+      - **A worker whose job errors never sends `MU-DONE`,** so `MU-WAIT-ALL` waits forever.
+        The error prints, but the trial produces no marker.
+      - **Worker execution errors are no longer counted.** `VM-ERROR?` can't see them, because
+        `sk_hermes_drain_checkpoint()` restores the VM's error flag after each payload (by
+        design: a bad message must not abort the enclosing context). `fail` now counts only
+        refused `SEND`s.
+      - **Hera's wait loop is a busy loop.** She takes a fair share of the CPU
+        (1/(N+1)), as the row tally shows.
+      - Per-VM, per-beat `HB-ON` rows make trials slow. One row in that capture shows a negative
+        (wrapped) `tick_interval_ns` for a VM a beat behind the others. That's a DoE-row
+        observation, not investigated.
+      - Stage F's remaining evidence (work proceeds with the console idle) is shown by Stage G's
+        interleave and by these trials.

capsules/BLOCK_MAP.md

@@ -1,5 +1,5 @@
 # Capsule Block Manifest — Auto-generated
-<!-- Generated by mkcapsule --manifest  2026-09-24T00:51:15Z -->
+<!-- Generated by mkcapsule --manifest  2026-09-24T03:15:17Z -->
 <!-- DO NOT EDIT — re-run mkcapsule --manifest to refresh.  -->
 <!-- Hand-written justifications and immutability notes live  -->
 <!-- in MANIFEST.md alongside this auto-generated index.      -->
@@ -25,23 +25,23 @@
 | `init-l8-volatile.4th` | 4810, 4811, 4812, 4813 | `0x98caabbbd92abac4` | yes |
 | `init.4th` | 2049, 2050, 2057 | `0xea038ba684c53443` | yes |
 | `lib.4th` | 4050 | `0x4b216635c359ef73` | yes |
-| `multiuser-doe.4th` | 5044, 5045, 5046, 5047, 5048, 5049, 5050, 5051, 5052, 5053, 5054, 5055 | `0xb4f6f0c2c582a4d4` | yes |
+| `multiuser-doe.4th` | 5044, 5045, 5046, 5047, 5048, 5049, 5050, 5051, 5052, 5053, 5054, 5055 | `0x2b268decf6065736` | yes |
 | `per-isa-doe.4th` | 5117, 5119 | `0xab13edecabfd78db` | yes |
 | `sdk.4th` | 5109, 5110, 5111, 5112, 5113, 5114, 5115 | `0x008fdbbb62c94a3a` | yes |
 | `turtle.4th` | 5100, 5101, 5102, 5103, 5104, 5105, 5106, 5107, 5108 | `0x4d470418ca543365` | yes |
 | `user-font-demo.4th` | 4200, 4201, 4202 | `0xce1fd7d1b581a56d` | yes |
-| `workload-0.4th` | 2200, 2201 | `0x93f86f60aeba8feb` | yes |
-| `workload-1-lite.4th` | 5058, 5059 | `0x44a7a7e3176dcc8d` | yes |
+| `workload-0.4th` | 2200, 2201 | `0x5597918289a363e3` | yes |
+| `workload-1-lite.4th` | 5058, 5059 | `0xbc13567fea25063a` | yes |
 | `workload-1.4th` | 4406, 4415, 4425, 4435 | `0x63e251adb0a03613` | yes |
-| `workload-2.4th` | 4506, 4515, 4525, 4535, 4545 | `0xf113b3d0bcccae47` | yes |
-| `workload-3.4th` | 4606, 4615, 4625, 4635, 4645, 4655, 4665 | `0x62b7a71576ad1041` | yes |
-| `workload-4.4th` | 2130, 2131, 2132 | `0x099619264e650b1c` | yes |
-| `workload-5-lite.4th` | 5056, 5057 | `0xea0303f38824f254` | yes |
+| `workload-2.4th` | 4506, 4515, 4525, 4535, 4545 | `0x9f87a07575e8f86a` | yes |
+| `workload-3.4th` | 4606, 4615, 4625, 4635, 4645, 4655, 4665 | `0xc4a88f6991036cc4` | yes |
+| `workload-4.4th` | 2130, 2131, 2132 | `0x65372d2d3787d2ba` | yes |
+| `workload-5-lite.4th` | 5056, 5057 | `0x05c651e140ef667b` | yes |
 | `workload-5.4th` | *(none — raw code capsule)* | `0x526964c7c9506a11` | yes |
-| `workload-6.4th` | 2080, 2081, 2082, 2083, 2084, 2085, 2086, 2087, 2088, 2089, 2090, 2091, 2092, 2093, 2094, 2095 | `0x06fc0ce1e369ef5a` | yes |
-| `workload-7.4th` | 2150 | `0x02862291387eb7ac` | yes |
-| `workload-8.4th` | 2160 | `0x56b7f2f0efa000df` | yes |
-| `workload-9.4th` | 4706, 4715, 4725, 4735, 4745 | `0x3f2bec73142aa424` | yes |
+| `workload-6.4th` | 2080, 2081, 2082, 2083, 2084, 2085, 2086, 2087, 2088, 2089, 2090, 2091, 2092, 2093, 2094, 2095 | `0x2526f98096581e63` | yes |
+| `workload-7.4th` | 2150, 2151 | `0x3441183353715e63` | yes |
+| `workload-8.4th` | 2160 | `0xe182180361013aa8` | yes |
+| `workload-9.4th` | 4706, 4715, 4725, 4735, 4745 | `0x6c08bf4ad3f286b2` | yes |
 | `workload-calib1.4th` | 5043 | `0x3b9f2d17b554fabc` | yes |
 | `zuse-eligibility.4th` | 4021, 4022 | `0x7b28f4776a32e0b7` | yes |
 | `zuse.4th` | 4016, 4017, 4018 | `0xa14d5a7a88791cf6` | yes |
@@ -69,22 +69,22 @@
 | 2077 | `init-l8-omni.4th` | `0x5979e314d6452045` | ok |
 | 2078 | `init-l8-omni.4th` | `0x5979e314d6452045` | ok |
 | 2079 | `init-l8-omni.4th` | `0x5979e314d6452045` | ok |
-| 2080 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2081 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2082 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2083 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2084 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2085 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2086 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2087 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2088 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2089 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2090 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2091 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2092 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2093 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2094 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
-| 2095 | `workload-6.4th` | `0x06fc0ce1e369ef5a` | ok |
+| 2080 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2081 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2082 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2083 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2084 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2085 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2086 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2087 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2088 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2089 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2090 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2091 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2092 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2093 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2094 | `workload-6.4th` | `0x2526f98096581e63` | ok |
+| 2095 | `workload-6.4th` | `0x2526f98096581e63` | ok |
 | 2100 | `doe.4th` | `0xf154616d248e861f` | ok |
 | 2101 | `doe.4th` | `0xf154616d248e861f` | ok |
 | 2102 | `doe.4th` | `0xf154616d248e861f` | ok |
@@ -93,13 +93,14 @@
 | 2105 | `doe.4th` | `0xf154616d248e861f` | ok |
 | 2106 | `doe.4th` | `0xf154616d248e861f` | ok |
 | 2107 | `doe.4th` | `0xf154616d248e861f` | ok |
-| 2130 | `workload-4.4th` | `0x099619264e650b1c` | ok |
-| 2131 | `workload-4.4th` | `0x099619264e650b1c` | ok |
-| 2132 | `workload-4.4th` | `0x099619264e650b1c` | ok |
-| 2150 | `workload-7.4th` | `0x02862291387eb7ac` | ok |
-| 2160 | `workload-8.4th` | `0x56b7f2f0efa000df` | ok |
-| 2200 | `workload-0.4th` | `0x93f86f60aeba8feb` | ok |
-| 2201 | `workload-0.4th` | `0x93f86f60aeba8feb` | ok |
+| 2130 | `workload-4.4th` | `0x65372d2d3787d2ba` | ok |
+| 2131 | `workload-4.4th` | `0x65372d2d3787d2ba` | ok |
+| 2132 | `workload-4.4th` | `0x65372d2d3787d2ba` | ok |
+| 2150 | `workload-7.4th` | `0x3441183353715e63` | ok |
+| 2151 | `workload-7.4th` | `0x3441183353715e63` | ok |
+| 2160 | `workload-8.4th` | `0xe182180361013aa8` | ok |
+| 2200 | `workload-0.4th` | `0x5597918289a363e3` | ok |
+| 2201 | `workload-0.4th` | `0x5597918289a363e3` | ok |
 | 4000 | `ACL.4th` | `0xf8890c05c0d8f921` | ok |
 | 4001 | `ACL.4th` | `0xf8890c05c0d8f921` | ok |
 | 4002 | `ACL.4th` | `0xf8890c05c0d8f921` | ok |
@@ -202,23 +203,23 @@
 | 4415 | `workload-1.4th` | `0x63e251adb0a03613` | ok |
 | 4425 | `workload-1.4th` | `0x63e251adb0a03613` | ok |
 | 4435 | `workload-1.4th` | `0x63e251adb0a03613` | ok |
-| 4506 | `workload-2.4th` | `0xf113b3d0bcccae47` | ok |
-| 4515 | `workload-2.4th` | `0xf113b3d0bcccae47` | ok |
-| 4525 | `workload-2.4th` | `0xf113b3d0bcccae47` | ok |
-| 4535 | `workload-2.4th` | `0xf113b3d0bcccae47` | ok |
-| 4545 | `workload-2.4th` | `0xf113b3d0bcccae47` | ok |
-| 4606 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4615 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4625 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4635 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4645 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4655 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4665 | `workload-3.4th` | `0x62b7a71576ad1041` | ok |
-| 4706 | `workload-9.4th` | `0x3f2bec73142aa424` | ok |
-| 4715 | `workload-9.4th` | `0x3f2bec73142aa424` | ok |
-| 4725 | `workload-9.4th` | `0x3f2bec73142aa424` | ok |
-| 4735 | `workload-9.4th` | `0x3f2bec73142aa424` | ok |
-| 4745 | `workload-9.4th` | `0x3f2bec73142aa424` | ok |
+| 4506 | `workload-2.4th` | `0x9f87a07575e8f86a` | ok |
+| 4515 | `workload-2.4th` | `0x9f87a07575e8f86a` | ok |
+| 4525 | `workload-2.4th` | `0x9f87a07575e8f86a` | ok |
+| 4535 | `workload-2.4th` | `0x9f87a07575e8f86a` | ok |
+| 4545 | `workload-2.4th` | `0x9f87a07575e8f86a` | ok |
+| 4606 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4615 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4625 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4635 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4645 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4655 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4665 | `workload-3.4th` | `0xc4a88f6991036cc4` | ok |
+| 4706 | `workload-9.4th` | `0x6c08bf4ad3f286b2` | ok |
+| 4715 | `workload-9.4th` | `0x6c08bf4ad3f286b2` | ok |
+| 4725 | `workload-9.4th` | `0x6c08bf4ad3f286b2` | ok |
+| 4735 | `workload-9.4th` | `0x6c08bf4ad3f286b2` | ok |
+| 4745 | `workload-9.4th` | `0x6c08bf4ad3f286b2` | ok |
 | 4806 | `init-l8-stable.4th` | `0xdc3830f189063a9a` | ok |
 | 4810 | `init-l8-volatile.4th` | `0x98caabbbd92abac4` | ok |
 | 4811 | `init-l8-volatile.4th` | `0x98caabbbd92abac4` | ok |
@@ -333,22 +334,22 @@
 | 5001 | `fabric.4th` | `0x9d9489cbeca4099b` | ok |
 | 5002 | `fabric.4th` | `0x9d9489cbeca4099b` | ok |
 | 5043 | `workload-calib1.4th` | `0x3b9f2d17b554fabc` | ok |
-| 5044 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5045 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5046 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5047 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5048 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5049 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5050 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5051 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5052 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5053 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5054 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5055 | `multiuser-doe.4th` | `0xb4f6f0c2c582a4d4` | ok |
-| 5056 | `workload-5-lite.4th` | `0xea0303f38824f254` | ok |
-| 5057 | `workload-5-lite.4th` | `0xea0303f38824f254` | ok |
-| 5058 | `workload-1-lite.4th` | `0x44a7a7e3176dcc8d` | ok |
-| 5059 | `workload-1-lite.4th` | `0x44a7a7e3176dcc8d` | ok |
+| 5044 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5045 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5046 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5047 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5048 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5049 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5050 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5051 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5052 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5053 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5054 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5055 | `multiuser-doe.4th` | `0x2b268decf6065736` | ok |
+| 5056 | `workload-5-lite.4th` | `0x05c651e140ef667b` | ok |
+| 5057 | `workload-5-lite.4th` | `0x05c651e140ef667b` | ok |
+| 5058 | `workload-1-lite.4th` | `0xbc13567fea25063a` | ok |
+| 5059 | `workload-1-lite.4th` | `0xbc13567fea25063a` | ok |
 | 5100 | `turtle.4th` | `0x4d470418ca543365` | ok |
 | 5101 | `turtle.4th` | `0x4d470418ca543365` | ok |
 | 5102 | `turtle.4th` | `0x4d470418ca543365` | ok |

capsules/multiuser-doe.4th

@@ -29,19 +29,14 @@ Block 5045
   ENDCASE ;
 
 Block 5046
-: WL-ENTRY ( idx -- c-addr u )
-  CASE
-    0 OF S" RUN-TRANSPOSE" ENDOF
-    1 OF S" MICRO-BURST" ENDOF
-    2 OF S" RUN-SAW" ENDOF
-    3 OF S" DUAL-SINE" ENDOF
-    4 OF S" RUN-PI" ENDOF
-    5 OF S" CHAOS-FIELD" ENDOF
-    6 OF S" RUN-OMNI" ENDOF
-    7 OF S" RUN-MMUL" ENDOF
-    8 OF S" RUN-FIB" ENDOF
-    9 OF S" RUN-SAW" ENDOF
-  ENDCASE ;
+( FABRIC-3.8 3.9 Stage H: workers are born defs-only and    )
+( each is SENT one job. WL-RUN, defined in every workload  )
+( capsule, is its old birth-time tail plus its entry word  )
+( twice, as the old two VM-EXEC passes did. The LAST word  )
+( of the job SENDs MU-DONE back to Hera, who counts them.  )
+VARIABLE MU-DONE-COUNT
+: MU-DONE ( -- ) MU-DONE-COUNT @ 1+ MU-DONE-COUNT ! ;
+: MU-JOB ( -- c-addr u ) S" WL-RUN WL-DONE-ARGS SEND" ;
 
 Block 5047
 : MU-WORKER-NAME ( n -- c-addr u )
@@ -66,17 +61,12 @@ Block 5048
   WL-CAPSULE
   MU-TMP-NC @ MU-TMP-NU @
   WORKER-BIRTH ;
-: MU-TOUCH-WORKER ( wl-idx worker-n -- )
-  MU-WORKER-NAME MU-TMP-NU ! MU-TMP-NC !
-  WL-ENTRY
-  MU-TMP-NC @ MU-TMP-NU @
-  VM-EXEC ;
 
 Block 5049
 VARIABLE MU-FAIL-COUNT
-: MU-CHECK-WORKER ( worker-n -- )
-  MU-WORKER-NAME VM-ERROR?
-  IF MU-FAIL-COUNT @ 1+ MU-FAIL-COUNT ! THEN ;
+: MU-SEND-WORKER ( worker-n -- )
+  MU-JOB ROT MU-WORKER-NAME SEND 0= IF
+  MU-FAIL-COUNT @ 1+ MU-FAIL-COUNT ! MU-DONE THEN ;
 : MU-KILL-WORKER ( worker-n -- )
   MU-WORKER-NAME KILL ;
 CREATE MU-RUN-MATRIX MU-MAX-RUNS CELLS ALLOT
@@ -125,14 +115,12 @@ Block 5052
   MU-NW @ 0 DO
     MU-MODE @ I MU-WL-FOR I MU-BIRTH-WORKER DROP
   LOOP ;
-: MU-TOUCH-ALL ( -- )
+: MU-SEND-ALL ( -- )
   MU-NW @ 0 DO
-    MU-MODE @ I MU-WL-FOR I MU-TOUCH-WORKER
-  LOOP ;
-: MU-CHECK-ALL ( -- )
-  MU-NW @ 0 DO
-    I MU-CHECK-WORKER
+    I MU-SEND-WORKER
   LOOP ;
+: MU-WAIT-ALL ( -- )
+  BEGIN MU-DONE-COUNT @ MU-NW @ < WHILE REPEAT ;
 
 Block 5053
 : MU-KILL-ALL ( -- )
@@ -141,13 +129,12 @@ Block 5053
   LOOP ;
 : MU-RUN-TRIAL ( cfg rep run-id -- )
   MU-RUN-ID ! MU-REP ! MU-CFG !
-  0 MU-FAIL-COUNT !
+  0 MU-FAIL-COUNT ! 0 MU-DONE-COUNT !
   MU-CFG @ MU-CONC-OF MU-NW !
   MU-CFG @ MU-MODE-OF MU-MODE !
   MU-BIRTH-ALL
-  MU-TOUCH-ALL
-  MU-TOUCH-ALL
-  MU-CHECK-ALL
+  MU-SEND-ALL
+  MU-WAIT-ALL
   MU-KILL-ALL
   MU-EMIT-TRIAL-MARKER ;
 

capsules/workload-0.4th

@@ -15,5 +15,6 @@ Block 2201
     TRANSPOSE-CHUNK 46 EMIT
   LOOP
 ;
-( RUN-TRANSPOSE )
-TRANSPOSE-CHUNK
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN TRANSPOSE-CHUNK RUN-TRANSPOSE RUN-TRANSPOSE ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-1-lite.4th

@@ -15,4 +15,6 @@ Block 5058
 Block 5059
 : RUN-MICRO-BURSTS ( -- )
   1000 0 DO MICRO-BURST LOOP ;
-RUN-MICRO-BURSTS
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN RUN-MICRO-BURSTS MICRO-BURST MICRO-BURST ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-2.4th

@@ -12,7 +12,6 @@ Block 4506
 
 Block 4515
 ( Rising edge with random perturbation )
-42 SEED
 : RISING-EDGE ( steps -- )
   0 DO
     I 1+
@@ -41,11 +40,8 @@ Block 4525
     TRIANGLE-CYCLE
   LOOP DROP ;
 
-3 10 TRIANGLE-WAVE
-
 Block 4535
 ( Asymmetric triangles - fast rise, slow fall )
-7777 SEED
 : ASYM-TRIANGLE ( -- )
   8 0 DO
     I 1+ 80 120 RANDOM 100 */ RAMP-STEP
@@ -56,14 +52,11 @@ Block 4535
     80 120 RANDOM WAIT
   LOOP ;
 : RUN-ASYM 4 0 DO ASYM-TRIANGLE LOOP ;
-RUN-ASYM
 
 Block 4545
 
 ( Sawtooth variant - all rise, sharp drop )
 
-54321 SEED
-
 : SAWTOOTH ( steps -- )
   0 DO
     I 1+ 85 115 RANDOM 100 */ RAMP-STEP
@@ -72,4 +65,7 @@ Block 4545
   50 WAIT ;
 
 : RUN-SAW 3 0 DO 12 SAWTOOTH LOOP ;
-RUN-SAW
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN 42 SEED 3 10 TRIANGLE-WAVE 7777 SEED RUN-ASYM
+  54321 SEED RUN-SAW RUN-SAW RUN-SAW ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-3.4th

@@ -36,7 +36,6 @@ Block 4625
 
 Block 4635
 ( Sinusoidal wave with random amplitude/frequency modulation )
-42 SEED
 : SINE-PULSE ( phase amplitude -- )
   SWAP FULL-SIN * 100 / 1+
   COMPUTE ;
@@ -61,14 +60,10 @@ Block 4645
     100 200 RANDOM WAIT    ( inter-cycle pause )
   LOOP ;
 
-3 SINE-WAVE
-
 Block 4655
 
 ( Phase-shifted dual sine - interference pattern )
 
-98765 SEED
-
 : DUAL-SINE ( -- )
   32 0 DO
     I FULL-SIN 80 120 RANDOM 100 */
@@ -78,11 +73,9 @@ Block 4655
   LOOP ;
 
 : RUN-DUAL 4 0 DO DUAL-SINE LOOP ;
-RUN-DUAL
 
 Block 4665
 ( Damped sine - decaying amplitude )
-11111 SEED
 : DAMPED-SINE ( cycles -- )
   100 SWAP               ( starting amplitude )
   0 DO
@@ -94,5 +87,7 @@ Block 4665
     85 * 100 /           ( decay by 15% each cycle )
     DUP 10 < IF DROP 10 THEN
   LOOP DROP ;
-
-3 DAMPED-SINE
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN 42 SEED 3 SINE-WAVE 98765 SEED RUN-DUAL
+  11111 SEED 3 DAMPED-SINE DUAL-SINE DUAL-SINE ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-4.4th

@@ -30,5 +30,6 @@ Block 2132
     DO-PI
   LOOP
 ;
-RUN-PI
-." 4:dopi-ok" CR
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN RUN-PI ." 4:dopi-ok" CR RUN-PI RUN-PI ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-5-lite.4th

@@ -21,4 +21,6 @@ Block 5057
     CHAOS-BLOCK
   LOOP
 ;
-CHAOS-FIELD
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN CHAOS-FIELD CHAOS-FIELD CHAOS-FIELD ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-6.4th

@@ -163,5 +163,6 @@ Block 2095
 : RUN-OMNI
   OMNI-WORK
 ;
-
-RUN-OMNI
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN RUN-OMNI RUN-OMNI RUN-OMNI ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-7.4th

@@ -14,4 +14,7 @@ Block 2150
     MM-CHUNK 46 EMIT
   LOOP
 ;
-RUN-MMUL
+Block 2151
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN RUN-MMUL RUN-MMUL RUN-MMUL ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-8.4th

@@ -12,5 +12,6 @@ Block 2160
     FIB-SEQ 46 EMIT
   LOOP
 ;
-FIB-SEQ
-( RUN-FIB )
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN FIB-SEQ RUN-FIB RUN-FIB ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

capsules/workload-9.4th

@@ -12,7 +12,6 @@ Block 4706
 
 Block 4715
 ( Rising edge with random perturbation )
-42 SEED
 : RISING-EDGE ( steps -- )
   0 DO
     I 1+
@@ -41,11 +40,8 @@ Block 4725
     TRIANGLE-CYCLE
   LOOP DROP ;
 
-3 10 TRIANGLE-WAVE
-
 Block 4735
 ( Asymmetric triangles - fast rise, slow fall )
-7777 SEED
 : ASYM-TRIANGLE ( -- )
   8 0 DO
     I 1+ 80 120 RANDOM 100 */ RAMP-STEP
@@ -56,14 +52,11 @@ Block 4735
     80 120 RANDOM WAIT
   LOOP ;
 : RUN-ASYM 4 0 DO ASYM-TRIANGLE LOOP ;
-RUN-ASYM
 
 Block 4745
 
 ( Sawtooth variant - all rise, sharp drop )
 
-54321 SEED
-
 : SAWTOOTH ( steps -- )
   0 DO
     I 1+ 85 115 RANDOM 100 */ RAMP-STEP
@@ -72,4 +65,7 @@ Block 4745
   50 WAIT ;
 
 : RUN-SAW 3 0 DO 12 SAWTOOTH LOOP ;
-RUN-SAW
+( FABRIC-3.8 3.9: defs only; the DoE SENDs WL-RUN )
+: WL-RUN 42 SEED 3 10 TRIANGLE-WAVE 7777 SEED RUN-ASYM
+  54321 SEED RUN-SAW RUN-SAW RUN-SAW ;
+: WL-DONE-ARGS S" MU-DONE" S" Hera" ;

include/starkernel/capsule_vm_physics.h

@@ -114,15 +114,29 @@ void vm_physics_charge(VMUuid vm_id);
 /**
  * vm_physics_work_heat_of - Cumulative heat a VM has pulled in by working
  *
- * FABRIC-3.8.md §3.13.7, ruling C-4 (β), Captain Bob 2026-09-23: the sum of
- * every transfer vm_physics_charge() actually moved INTO this VM -- heat
- * earned by its own work, independent of any structural endowment (Hera's
- * initial Q48_ONE). This is the quantity the switch decision compares
- * ("coolest VM with work" = least work heat). Only grows; 0 for an unknown
- * or dead VM.
+ * FABRIC-3.8.md §3.13.7, ruling C-4 (β), Captain Bob 2026-09-23: heat
+ * earned by the VM's own work -- the charge vm_physics_charge() assessed at
+ * each of its beats (beats x slope, in the charge carry's sub-unit scale),
+ * whether or not the rest of the fleet held heat to move -- independent of
+ * any structural endowment (Hera's initial Q48_ONE). This is the quantity
+ * the switch decision compares ("coolest VM with work" = least work heat).
+ * Never decreases, except that it may be raised by
+ * vm_physics_work_heat_floor(). 0 for an unknown or dead VM.
  */
 uint64_t vm_physics_work_heat_of(VMUuid vm_id);
 
+/**
+ * vm_physics_work_heat_floor - "Floor on contact" (Captain Bob, Stage H)
+ *
+ * Raise vm_id's work heat to floor_q48 if it is below it; never lowers it.
+ * Called when a VM that had no work gains some (sk_vm_switch_signal_note_
+ * work_gained()), with the lowest work heat among the VMs that currently
+ * have work -- so a newly-busy VM joins the running set level with its
+ * floor instead of at whatever it last had (0 for a new VM), which would
+ * otherwise let it run alone until it "caught up."
+ */
+void vm_physics_work_heat_floor(VMUuid vm_id, uint64_t floor_q48);
+
 /**
  * vm_physics_tick - Heartbeat-gated inference pass
  *

include/starkernel/capsule_vm_switch_signal.h

@@ -150,6 +150,15 @@ int  sk_vm_switch_signal_take_decision_due(void);
  */
 struct VM *sk_vm_switch_decide(struct VM *self, int self_competes);
 
+/* Stage H, "floor on contact" (Captain Bob, 2026-09-23): vm_id just gained
+ * work (kernel_hermes.c: its queue went from empty to one message). If it
+ * had none before -- not running, not parked mid-task -- raise its work
+ * heat to the lowest work heat among the VMs that have work now (the one
+ * running and every eligible participant). A new or long-idle VM thus
+ * joins the busy set level with its floor, rather than at 0 and running
+ * alone until it "caught up" with VMs that have worked since boot. */
+void sk_vm_switch_signal_note_work_gained(VMUuid vm_id);
+
 /* Reap one participant marked pending_reap (a WIREBIND identity whose
  * device detached while it was parked), if any. Call only at a safe point
  * (checkpoint, trampoline, idle point) -- same contract the old Stage 3

src/starkernel/capsule/capsule_vm_physics.c

@@ -68,15 +68,16 @@ typedef struct {
                                * away, so a one-beat charge (slope < 1 unit)
                                * still moves heat once enough beats accrue */
     uint64_t work_heat_q48;   /* FABRIC-3.8.md §3.13.7, ruling C-4 (β), Captain
-                               * Bob 2026-09-23: cumulative heat this VM has
-                               * pulled in through its OWN per-beat charges --
-                               * heat earned by working. The switch decision
+                               * Bob 2026-09-23: heat earned by working -- the
+                               * charge assessed at each of this VM's own
+                               * beats (sub-unit scale), whether or not it
+                               * could be moved; raised to the busy set's
+                               * floor when the VM gains work (Stage H,
+                               * "floor on contact"). The switch decision
                                * compares this, not execution_heat_q48, so
-                               * Hera's structural endowment (all of Q48_ONE
-                               * at init) never counts against her. Only ever
-                               * grows; a record of transfers that already
-                               * happened, not heat of its own (fleet K is
-                               * unaffected). */
+                               * Hera's structural endowment never counts
+                               * against her. Bookkeeping, not heat of its
+                               * own: fleet K is unaffected. */
 } VMPhysics;
 
 typedef struct vm_physics_node {
@@ -388,6 +389,15 @@ void vm_physics_charge(VMUuid vm_id)
 
     if (!target || !target->physics.is_live) return;
 
+    /* §3.13.7 (β), refined at Stage H (Captain Bob, "floor on contact"):
+     * work heat accrues the charge each beat was ASSESSED, in the carry's
+     * own sub-unit scale -- not what could actually be moved. Counting
+     * only moved heat left a busy VM at 0 whenever the rest of the fleet
+     * held nothing to pull (seen live: Hera busy-waiting in the DoE never
+     * yielded). This is work done, beats x slope, independent of who holds
+     * heat. */
+    target->physics.work_heat_q48 += fleet_transfer_slope_q48;
+
     target->physics.charge_carry += fleet_transfer_slope_q48;
     amount = target->physics.charge_carry >> 16;
     if (amount == 0) return;
@@ -419,9 +429,6 @@ void vm_physics_charge(VMUuid vm_id)
      * vanishing. Conservation is unaffected either way -- only what
      * vm_physics_transfer() actually moved ever changed any heat. */
     target->physics.charge_carry += (moved_total - moved) << 16;
-
-    /* §3.13.7 (β): what this VM's own work actually pulled in. */
-    target->physics.work_heat_q48 += moved;
 }
 
 uint64_t vm_physics_work_heat_of(VMUuid vm_id)
@@ -431,6 +438,13 @@ uint64_t vm_physics_work_heat_of(VMUuid vm_id)
     return n->physics.work_heat_q48;
 }
 
+void vm_physics_work_heat_floor(VMUuid vm_id, uint64_t floor_q48)
+{
+    vm_physics_node_t *n = vm_physics_find(vm_id);
+    if (!n || !n->physics.is_live) return;
+    if (n->physics.work_heat_q48 < floor_q48) n->physics.work_heat_q48 = floor_q48;
+}
+
 /**
  * l8_regime_modulation_q16 - Read-only Q16 multiplier from Hera's own L8
  *

src/starkernel/capsule/capsule_vm_switch_signal.c

@@ -269,6 +269,46 @@ VM *sk_vm_switch_decide(VM *self, int self_competes) {
     return best;
 }
 
+/* Stage H, "floor on contact" (Captain Bob, 2026-09-23). See the header. */
+void sk_vm_switch_signal_note_work_gained(VMUuid vm_id) {
+    VM      *current = sk_vm_switch_current_vm();
+    VM      *mama    = (VM *)sk_get_mama_vm();
+    VMRegistryEntry e;
+    VM      *v;
+    uint64_t floor_q48 = 0;
+    int      have = 0, i;
+
+    if (capsule_vm_registry_get(vm_id, &e) != 0 || !e.vm_ptr) return;
+    v = (VM *)e.vm_ptr;
+
+    /* Only a VM that had NO work gains some here: not the one running, and
+     * not one parked mid-task (either already had work, and keeps its own
+     * work heat). */
+    if (v == current) return;
+    if (v->native_stack_saved_sp != 0 && v->frames.count > 0 &&
+        v->frames.entries[v->frames.count - 1].kind != VM_FRAME_IDLE) return;
+
+    /* The floor: the lowest work heat among the VMs that have work right
+     * now -- the one running, and every other eligible participant. */
+    if (current) {
+        floor_q48 = vm_physics_work_heat_of(current->stadium_vm_id);
+        have = 1;
+    }
+    for (i = 0; i < g_slot_count; i++) {
+        VMRegistryEntry o;
+        VM      *ov;
+        uint64_t heat, stamp;
+        if (vm_uuid_equal(g_slots[i].vm_id, vm_id)) continue;
+        if (capsule_vm_registry_get(g_slots[i].vm_id, &o) != 0) continue;
+        ov = (VM *)o.vm_ptr;
+        if (!ov || ov == current || o.pending_reap) continue;
+        if (o.state == VM_STATE_DEAD || o.state == VM_STATE_STILLBORN) continue;
+        if (!candidate_standing(ov, mama, g_slots[i].parked_stamp, &heat, &stamp)) continue;
+        if (!have || heat < floor_q48) { floor_q48 = heat; have = 1; }
+    }
+    if (have) vm_physics_work_heat_floor(vm_id, floor_q48);
+}
+
 void sk_vm_switch_reap_pending(void) {
     VM *current = sk_vm_switch_current_vm();
     int i;

src/starkernel/vm/kernel_hermes.c

@@ -536,6 +536,7 @@ int sk_hermes_send_one(VMUuid from, VMUuid to, uint32_t type, uint32_t channel_i
      * the switch decision told something changed -- the recipient may now
      * be the coolest VM with work. */
     msg->wait_stamp = sk_vm_switch_signal_next_stamp();
+    if (q->count == 1) sk_vm_switch_signal_note_work_gained(to); /* Stage H floor */
     sk_vm_switch_signal_note_mail();
 
     return 0;

src/starkernel/vm/vm_core.c

@@ -1213,6 +1213,19 @@ int sk_vm_owns_outermost_frame(VM *vm) {
            vm->mode != MODE_COMPILE;
 }
 
+/* Stage H (FABRIC-3.8.md §3.9): 1 if the rest of the input holds no more
+ * words -- the same whitespace set vm_parse_word() skips, so "exhausted"
+ * means exactly "the next vm_parse_word() would return 0". */
+static int vm_input_exhausted(const VM *vm) {
+    size_t p = vm->input_pos;
+    while (p < vm->input_length) {
+        char c = vm->input_buffer[p];
+        if (c != ' ' && c != '\t' && c != '\n' && c != '\r') return 0;
+        p++;
+    }
+    return 1;
+}
+
 void vm_interpret(VM* vm, const char* input)
 {
     vm_interpret_kind(vm, input, VM_FRAME_INTERPRET);
@@ -1276,7 +1289,15 @@ void vm_interpret_kind(VM* vm, const char* input, uint32_t kind)
          * is a clean switch point too -- without it, a payload made only
          * of primitives (no colon word, so no execute_colon_word()
          * checkpoint) could never be displaced. Same ownership gate. */
-        if (!vm->error && !vm->abort_requested && sk_vm_owns_outermost_frame(vm)) {
+        /* Stage H (2026-09-23): not after the LAST word. A VM with nothing
+         * left to interpret isn't mid-task -- its trampoline decides the
+         * moment this returns -- and switching it out here would park it
+         * "mid-task" with no work left, which KILL (rightly) refuses. Seen
+         * in design for the DoE's completion notice: a payload whose final
+         * word is a SEND to Hera sets decision-due, and this checkpoint
+         * would otherwise act on it one word too late to matter. */
+        if (!vm->error && !vm->abort_requested && sk_vm_owns_outermost_frame(vm) &&
+            !vm_input_exhausted(vm)) {
             sk_vm_switch_checkpoint(vm);
         }
 #endif