recipes: GMU 0.88 in r27/r28 4x+2x (startup gate 107.0 vs 109.5 GiB; KV pinned, capacity unchanged)

recipes/glm53-flash-r27-4x-dcp4/recipe.toml

@@ -103,7 +103,11 @@ args = [
     # of the old "54-second death"), and post-crash reruns see ~106.6 GiB.
     # 0.82 = 99.7 GiB, below any realistic dip; does NOT change what gets
     # allocated (weights ~93 + pinned KV 10 + activations).
-    "--gpu-memory-utilization", "0.90",
+    # 0.88 (not 0.90): the startup admission gate is free-mem >= GMU*total
+    # (107.0 vs 109.5 GiB); ~12 GiB of starting vLLM processes + docker
+    # daemon can push the 0.90 check under the line (2026-09-08, #48140).
+    # KV is explicitly pinned (--kv-cache-memory-bytes) so capacity is unchanged.
+    "--gpu-memory-utilization", "0.88",
     # Pinned 10 GiB KV (≈1.16M tokens) — GMU sizing has no headroom at
     # 92.72 GiB weights/node; pinning also keeps warmup off the swap path
     "--kv-cache-memory-bytes", "17179869184",  # 16 GiB: 4x NVFP4 has the headroom (47.1 GiB weights/node vs 108 usable); 2x stays 10

recipes/glm53-flash-r27-spark/recipe.toml

@@ -99,7 +99,11 @@ args = [
     # of the old "54-second death"), and post-crash reruns see ~106.6 GiB.
     # 0.82 = 99.7 GiB, below any realistic dip; does NOT change what gets
     # allocated (weights ~93 + pinned KV 10 + activations).
-    "--gpu-memory-utilization", "0.90",
+    # 0.88 (not 0.90): the startup admission gate is free-mem >= GMU*total
+    # (107.0 vs 109.5 GiB); ~12 GiB of starting vLLM processes + docker
+    # daemon can push the 0.90 check under the line (2026-09-08, #48140).
+    # KV is explicitly pinned (--kv-cache-memory-bytes) so capacity is unchanged.
+    "--gpu-memory-utilization", "0.88",
     # Pinned 10 GiB KV (≈1.16M tokens) — GMU sizing has no headroom at
     # 92.72 GiB weights/node; pinning also keeps warmup off the swap path
     "--kv-cache-memory-bytes", "10737418240",

recipes/glm53-flash-r28-4x-dcp4-b12xload/recipe.toml

@@ -125,7 +125,11 @@ args = [
     # of the old "54-second death"), and post-crash reruns see ~106.6 GiB.
     # 0.82 = 99.7 GiB, below any realistic dip; does NOT change what gets
     # allocated (weights ~93 + pinned KV 10 + activations).
-    "--gpu-memory-utilization", "0.90",
+    # 0.88 (not 0.90): the startup admission gate is free-mem >= GMU*total
+    # (107.0 vs 109.5 GiB); ~12 GiB of starting vLLM processes + docker
+    # daemon can push the 0.90 check under the line (2026-09-08, #48140).
+    # KV is explicitly pinned (--kv-cache-memory-bytes) so capacity is unchanged.
+    "--gpu-memory-utilization", "0.88",
     # Pinned 10 GiB KV (≈1.16M tokens) — GMU sizing has no headroom at
     # 92.72 GiB weights/node; pinning also keeps warmup off the swap path
     "--kv-cache-memory-bytes", "17179869184",  # 16 GiB: 4x NVFP4 has the headroom (47.1 GiB weights/node vs 108 usable); 2x stays 10

recipes/glm53-flash-r28-4x-dcp4/recipe.toml

@@ -125,7 +125,11 @@ args = [
     # of the old "54-second death"), and post-crash reruns see ~106.6 GiB.
     # 0.82 = 99.7 GiB, below any realistic dip; does NOT change what gets
     # allocated (weights ~93 + pinned KV 10 + activations).
-    "--gpu-memory-utilization", "0.90",
+    # 0.88 (not 0.90): the startup admission gate is free-mem >= GMU*total
+    # (107.0 vs 109.5 GiB); ~12 GiB of starting vLLM processes + docker
+    # daemon can push the 0.90 check under the line (2026-09-08, #48140).
+    # KV is explicitly pinned (--kv-cache-memory-bytes) so capacity is unchanged.
+    "--gpu-memory-utilization", "0.88",
     # Pinned 10 GiB KV (≈1.16M tokens) — GMU sizing has no headroom at
     # 92.72 GiB weights/node; pinning also keeps warmup off the swap path
     "--kv-cache-memory-bytes", "17179869184",  # 16 GiB: 4x NVFP4 has the headroom (47.1 GiB weights/node vs 108 usable); 2x stays 10

recipes/glm53-flash-r28-4x-dcp4/reports/2026-09-08_211618_r28-4x-dcp4-full-rerun.log

@@ -6,3 +6,79 @@
 │ latency=generation                                                           │
 ╰──────────────────────────────────────────────────────────────────────────────╯
 
+  ✓ Complete ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 18/18 0:02:11
+
+  llama-benchy 0.4.0
+  Estimated latency: 73.8 ms
+
+                               llama-benchy Results                             
+┏━━━━━━━━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━
+┃                      ┃    ┃         ┃         ┃     TTFT ┃    Total ┃         
+┃ Test                 ┃ c  ┃  pp t/s ┃  tg t/s ┃     (ms) ┃     (ms) ┃   Tokens
+┡━━━━━━━━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━
+│ pp2048 tg128 @ d0    │ c1 │   1,900 │    47.7 │    1,026 │    3,634 │ 2048+128
+│ pp2048 tg128 @ d0    │ c2 │   1,716 │    58.9 │    1,567 │    5,181 │ 2048+128
+│ pp2048 tg128 @ d4096 │ c1 │   1,910 │    44.6 │    2,899 │    5,692 │ 2048+128
+│ pp2048 tg128 @ d4096 │ c2 │   1,950 │    56.3 │    5,100 │    8,949 │ 2048+128
+│ pp2048 tg128 @ d8192 │ c1 │   2,000 │    47.7 │    4,602 │    7,211 │ 2048+128
+│ pp2048 tg128 @ d8192 │ c2 │   1,954 │    52.6 │    8,536 │   12,208 │ 2048+128
+└──────────────────────┴────┴─────────┴─────────┴──────────┴──────────┴─────────
+
+  ℹ Metrics sourced from llama-benchy — see https://github.com/eugr/llama-benchy
+for methodology.
+
+{"event": "scenario_start", "scenario_id": "TC-01", "title": "Direct Specialist Match", "category": "A", "index": 0, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-01", "status": "pass", "points": 2, "index": 0, "total": 88, "duration_seconds": 4.72}
+{"event": "scenario_start", "scenario_id": "TC-02", "title": "Distractor Resistance", "category": "A", "index": 1, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-02", "status": "pass", "points": 2, "index": 1, "total": 88, "duration_seconds": 3.89}
+{"event": "scenario_start", "scenario_id": "TC-03", "title": "Implicit Tool Need", "category": "A", "index": 2, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-03", "status": "pass", "points": 2, "index": 2, "total": 88, "duration_seconds": 5.8}
+{"event": "scenario_start", "scenario_id": "TC-04", "title": "Unit Handling", "category": "B", "index": 3, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-04", "status": "pass", "points": 2, "index": 3, "total": 88, "duration_seconds": 2.25}
+{"event": "scenario_start", "scenario_id": "TC-05", "title": "Date and Time Parsing", "category": "B", "index": 4, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-05", "status": "pass", "points": 2, "index": 4, "total": 88, "duration_seconds": 11.55}
+{"event": "scenario_start", "scenario_id": "TC-06", "title": "Multi-Value Extraction", "category": "B", "index": 5, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-06", "status": "pass", "points": 2, "index": 5, "total": 88, "duration_seconds": 4.01}
+{"event": "scenario_start", "scenario_id": "TC-07", "title": "Search \u2192 Read \u2192 Act", "category": "C", "index": 6, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-07", "status": "pass", "points": 2, "index": 6, "total": 88, "duration_seconds": 11.13}
+{"event": "scenario_start", "scenario_id": "TC-08", "title": "Conditional Branching", "category": "C", "index": 7, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-08", "status": "pass", "points": 2, "index": 7, "total": 88, "duration_seconds": 8.43}
+{"event": "scenario_start", "scenario_id": "TC-09", "title": "Parallel Independence", "category": "C", "index": 8, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-09", "status": "pass", "points": 2, "index": 8, "total": 88, "duration_seconds": 5.19}
+{"event": "scenario_start", "scenario_id": "TC-10", "title": "Trivial Knowledge", "category": "D", "index": 9, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-10", "status": "pass", "points": 2, "index": 9, "total": 88, "duration_seconds": 1.93}
+{"event": "scenario_start", "scenario_id": "TC-11", "title": "Simple Math", "category": "D", "index": 10, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-11", "status": "pass", "points": 2, "index": 10, "total": 88, "duration_seconds": 1.91}
+{"event": "scenario_start", "scenario_id": "TC-12", "title": "Impossible Request", "category": "D", "index": 11, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-12", "status": "pass", "points": 2, "index": 11, "total": 88, "duration_seconds": 9.74}
+{"event": "scenario_start", "scenario_id": "TC-13", "title": "Empty Results", "category": "E", "index": 12, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-13", "status": "pass", "points": 2, "index": 12, "total": 88, "duration_seconds": 7.27}
+{"event": "scenario_start", "scenario_id": "TC-14", "title": "Malformed Response", "category": "E", "index": 13, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-14", "status": "pass", "points": 2, "index": 13, "total": 88, "duration_seconds": 6.02}
+{"event": "scenario_start", "scenario_id": "TC-15", "title": "Conflicting Information", "category": "E", "index": 14, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-15", "status": "pass", "points": 2, "index": 14, "total": 88, "duration_seconds": 4.82}
+{"event": "scenario_start", "scenario_id": "TC-16", "title": "German Language Tool Call", "category": "F", "index": 15, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-16", "status": "pass", "points": 2, "index": 15, "total": 88, "duration_seconds": 5.25}
+{"event": "scenario_start", "scenario_id": "TC-17", "title": "Timezone-Aware Scheduling", "category": "F", "index": 16, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-17", "status": "pass", "points": 2, "index": 16, "total": 88, "duration_seconds": 5.8}
+{"event": "scenario_start", "scenario_id": "TC-18", "title": "Translate & Forward", "category": "F", "index": 17, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-18", "status": "pass", "points": 2, "index": 17, "total": 88, "duration_seconds": 8.51}
+{"event": "scenario_start", "scenario_id": "TC-19", "title": "Message Routing", "category": "G", "index": 18, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-19", "status": "pass", "points": 2, "index": 18, "total": 88, "duration_seconds": 4.78}
+{"event": "scenario_start", "scenario_id": "TC-20", "title": "Data Extraction & Calculation", "category": "G", "index": 19, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-20", "status": "pass", "points": 2, "index": 19, "total": 88, "duration_seconds": 10.57}
+{"event": "scenario_start", "scenario_id": "TC-21", "title": "Constraint Validation", "category": "G", "index": 20, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-21", "status": "pass", "points": 2, "index": 20, "total": 88, "duration_seconds": 20.96}
+{"event": "scenario_start", "scenario_id": "TC-22", "title": "Output Format Compliance", "category": "H", "index": 21, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-22", "status": "pass", "points": 2, "index": 21, "total": 88, "duration_seconds": 1.64}
+{"event": "scenario_start", "scenario_id": "TC-23", "title": "Explicit Tool Prohibition", "category": "H", "index": 22, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-23", "status": "pass", "points": 2, "index": 22, "total": 88, "duration_seconds": 12.04}
+{"event": "scenario_start", "scenario_id": "TC-24", "title": "Multi-Constraint Instruction", "category": "H", "index": 23, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-24", "status": "pass", "points": 2, "index": 23, "total": 88, "duration_seconds": 2.03}
+{"event": "scenario_start", "scenario_id": "TC-25", "title": "Cross-Reference Prior Results", "category": "I", "index": 24, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-25", "status": "pass", "points": 2, "index": 24, "total": 88, "duration_seconds": 11.39}
+{"event": "scenario_start", "scenario_id": "TC-26", "title": "State Consistency (Multi-Turn)", "category": "I", "index": 25, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-26", "status": "pass", "points": 2, "index": 25, "total": 88, "duration_seconds": 7.4}
+{"event": "scenario_start", "scenario_id": "TC-27", "title": "Deduplication Awareness", "category": "I", "index": 26, "total": 88}
+{"event": "scenario_result", "scenario_id": "TC-27", "status": "pass", "points": 2, "index": 26, "total": 88, "duration_seconds": 5.7}
+{"event": "scenario_start", "scenario_id": "TC-28", "title": "Read-Before-Write", "category": "J", "index": 27, "total": 88}