Phase 69: Refill tuning completion (Warm Pool Size=16 optimized)

- Promoted Warm Pool Size=16 as the new baseline (+3.26% gain). - Updated PERFORMANCE_TARGETS_SCORECARD.md with Phase 69 results. - Updated scripts/run_mixed_10_cleanenv.sh and core/bench_profile.h to use HAKMEM_WARM_POOL_SIZE=16 by default. - Clarified that TINY_REFILL_BATCH_SIZE is not currently connected.
2025-12-18 01:55:27 +09:00
parent 5c9b09148b
commit b6212bbe31
7 changed files with 122 additions and 68 deletions
--- a/CURRENT_TASK.md
+++ b/CURRENT_TASK.md
@ -2,11 +2,11 @@

 ## 0) 今の「正」

- **性能比較の正**: FAST PGO build（`make pgo-fast-full` → `bench_random_mixed_hakmem_minimal_pgo`）✓ **Phase 68 昇格済み** (seed/WS diversified)
+- **性能比較の正**: FAST PGO build（`make pgo-fast-full` → `bench_random_mixed_hakmem_minimal_pgo`）✓ **Phase 69 昇格済み** (Warm Pool Size=16)
 - **安全・互換の正**: Standard build（`make bench_random_mixed_hakmem`）
 - **観測の正**: OBSERVE build（`make perf_observe`）
- **スコアカード**: `docs/analysis/PERFORMANCE_TARGETS_SCORECARD.md`（M1 達成・超過: 50.93% vs 50% target）
- **計測の正（Mixed 10-run）**: `scripts/run_mixed_10_cleanenv.sh`（`ITERS=20000000 WS=400`）
+- **スコアカード**: `docs/analysis/PERFORMANCE_TARGETS_SCORECARD.md`（M1 達成・超過: 51.77% vs 50% target、M2 まで残り +3.23pp）
+- **計測の正（Mixed 10-run）**: `scripts/run_mixed_10_cleanenv.sh`（`ITERS=20000000 WS=400`、`HAKMEM_WARM_POOL_SIZE=16` デフォルト）

 ## 1) 現状（要点）

@ -18,7 +18,10 @@
  - Baseline: `bench_random_mixed_hakmem_minimal_pgo` = 60.89M ops/s = 50.32% (initial PGO)
 - Phase 68（PGO training set 最適化）: **GO & 昇格完了** ✓
  - 検証: 10-run で +1.19% vs Phase 66 (GO: +1.0% threshold超過)
-  - 新 baseline: `bench_random_mixed_hakmem_minimal_pgo` (upgraded) = 61.614M ops/s = **50.93%** (50% target 超過、+0.93pp)
+  - Baseline (upgraded): `bench_random_mixed_hakmem_minimal_pgo` = 61.614M ops/s = **50.93%** (50% target 超過、+0.93pp)
+- Phase 69（Refill tuning: Warm Pool Size 最適化）: **強GO & 昇格完了** ✓✓✓
+  - 検証: 10-run で +3.26% vs Phase 68 (強GO: +3.0% threshold超過)
+  - 新 baseline: `bench_random_mixed_hakmem_minimal_pgo` (upgraded) = 62.63M ops/s = **51.77%** (M1 超過、+1.77pp、M2 まで残り +3.23pp)

 ## 2) 次の指示書（Active）

@ -64,32 +67,63 @@

 - ✓ `docs/analysis/PHASE69_REFILL_TUNING_0_DESIGN.md` 作成
 - ✓ Tunable parameters 特定:
-  - `TINY_REFILL_BATCH_SIZE` (16 → 32/64/128)
+  - `HAKMEM_TINY_REFILL_COUNT_MID` / `HAKMEM_TINY_REFILL_COUNT_HOT`（refill 量の実体, ENV-only）
  - Unified Cache C5-C7 capacity (128 → 256/512)
  - Warm Pool size (12 → 16/24)
 - ✓ Sweep 計画立案（single-parameter → combined optimization）
 - ✓ Risk assessment & 判定基準定義

-**Phase 69-1（Active）: Sweep 実行**
+**Phase 69-1: Sweep 実行** ✅ **完了**

- **狙い**: +3〜6% (M2: 55% target への最短距離)
- **施策**（優先順）:
-  1. **Warm Pool Size** (ENV-only, no recompile): 12 → 16 → 24
-     - Expected: +0.5-1.0% (registry scan reduction)
-  2. **Unified Cache C5-C7** (ENV-only, no recompile): 128 → 256 → 512
-     - Expected: +1-2% (miss rate reduction for mid-size allocations)
-  3. **Refill Batch Size** (requires PGO rebuild): 16 → 32 → 64
-     - Expected: +1-3% (refill frequency reduction)
- **手順**:
-  - `scripts/run_mixed_10_cleanenv.sh` で 10-run (各パラメータ)
-  - 失敗時は `scripts/box/layout_tax_forensics_box.sh` を当てて原因分類
- **判定**:
-  - GO: +1.0%（まず1段目）
-  - "強GO": +3.0% 以上（M2射程の芯として昇格）
+- ✓ Baseline (Phase 68 PGO): 60.65M ops/s (10-run mean)
+- ✓ Warm Pool Size sweep:
+  - Size=16: **62.63M ops/s (+3.26%, 強GO)** ✓✓✓ **Winner**
+  - Size=24: 62.37M ops/s (+2.84%, GO)
+- ✓ Unified Cache C5-C7 sweep:
+  - Cache=256: 61.92M ops/s (+2.09%, GO)
+  - Cache=512: 61.80M ops/s (+1.89%, GO)
+- ✓ Combined optimization check:
+  - Warm=16 + Cache=256: 62.35M ops/s (+2.81%, non-additive)
+- ✓ “Refill Batch Size sweep” は無効（knob 未接続）:
+  - `TINY_REFILL_BATCH_SIZE` は現行 Tiny front に call site が無く、性能 knob として成立していない
+  - 参照: `docs/analysis/PHASE69_REFILL_TUNING_3C_REFILL_BATCH_KNOB_AUDIT.md`
+- **結果**: `docs/analysis/PHASE69_REFILL_TUNING_1_RESULTS.md`
+- **勝ち設定**: **Warm Pool Size=16 (ENV-only, +3.26%, 強GO)**

-**Phase 69-2（後続）: 勝ち設定を baseline に反映**
- 勝ちパラメータを `pgo_fast_profile_config.sh` / `core/hakmem_tiny_config.h` に反映
- `make pgo-fast-full` で再ビルド → baseline 昇格
+**Phase 69-2: 勝ち設定を baseline に反映** ✅ **完了**
+
+- ✓ `scripts/run_mixed_10_cleanenv.sh` に `HAKMEM_WARM_POOL_SIZE=16` デフォルト追加
+- ✓ `core/bench_profile.h` の `MIXED_TINYV3_C7_SAFE` preset に `bench_setenv_default("HAKMEM_WARM_POOL_SIZE","16")` 追加
+- ✓ `PERFORMANCE_TARGETS_SCORECARD.md` に新 baseline 追加:
+  - Phase 69 baseline: 62.63M ops/s = 51.77% of mimalloc
+  - M1 (50%) achievement: **EXCEEDED** (+1.77pp above target)
+  - M2 (55%) progress: Gap reduced to +3.23pp
+- ✓ Rollback: `HAKMEM_WARM_POOL_SIZE=12` or ENV 変数削除
+
+**新 baseline**: 62.63M ops/s = mimalloc の **51.77%** (Phase 68 から +3.26%、M2 まで残り +3.23pp)
+
+---
+
+**Phase 69-3（次候補）: refill 量（ENV-only）sweep OR 次の sweep**
+
+- **選択肢 A（推奨）**: Refill count の ENV sweep（コード変更なし）
+  - `HAKMEM_TINY_REFILL_COUNT_MID`（C4–C7）を 64/96/128/160… で sweep
+  - `HAKMEM_TINY_REFILL_COUNT_HOT`（C0–C3）も同様に sweep（ただし WarmPool/UnifiedCache と相互作用あり）
+  - 判定: 10-run mean で GO(+1.0%) / 強GO(+3.0%) / NO-GO(-1.0%)
+
+- **選択肢 B**: Unified Cache の fine sweep（ENV-only）
+  - C5/C6/C7 を 192/256/320… などで sweep（Phase 69-1 の 256/512 は coarse）
+  - WarmPool=16 との非加算性を “原因切り分け” する
+
+- **選択肢 C**: compile-time knob の新設（後回し）
+  - `TINY_REFILL_BATCH_SIZE` は未接続なので、そのまま追わない
+  - 必要なら別途 SSOT を作って実装する（Phase 70+）
+
+- **選択肢 D**: 別方向の最適化（M2: 55% への最短距離）
+  - 残り gap: +3.23pp (51.77% → 55%)
+  - Phase 67b（境界 inline/unroll チューニング）
+  - Top 50 hot functions の最適化
+  - PGO profile の再調整

 ---

--- a/6
+++ b/6
@ -1060,6 +1060,12 @@ clean:
 	rm -f $(OBJS) $(TARGET) $(BENCH_HAKMEM_OBJS) $(BENCH_SYSTEM_OBJS) $(BENCH_HAKMEM) $(BENCH_SYSTEM) $(SHARED_OBJS) $(SHARED_LIB) *.csv libhako_ffi_stub.a hako_ffi_stub.o
 	rm -f bench_comprehensive.o bench_comprehensive_hakmem bench_comprehensive_system
 	rm -f bench_tiny bench_tiny.o bench_tiny_mt bench_tiny_mt.o test_mf2 test_mf2.o bench_tiny_hakmem
+	rm -f bench_random_mixed_hakmem.o bench_random_mixed_system.o bench_random_mixed_mi.o
+	rm -f bench_tiny_hot_hakmem.o bench_tiny_hot_system.o bench_tiny_hot_mi.o bench_mi_force.o
+	rm -f bench_random_mixed_hakmem bench_random_mixed_system bench_random_mixed_mi bench_random_mixed_hakx
+	rm -f bench_random_mixed_hakmem_minimal bench_random_mixed_hakmem_minimal_pgo
+	rm -f bench_random_mixed_hakmem_fast_fixed bench_random_mixed_hakmem_fast_pruned bench_random_mixed_hakmem_fast_pgo
+	rm -f bench_tiny_hot_hakmem bench_tiny_hot_system bench_tiny_hot_mi bench_tiny_hot_hakmi bench_tiny_hot_hakx bench_tiny_hot_hakx_p0 bench_tiny_hot_direct

 # Help
 help:
--- a/core/bench_profile.h
+++ b/core/bench_profile.h
@ -103,6 +103,8 @@ static inline void bench_apply_mixed_tinyv3_c7_common(void) {
  bench_setenv_default("HAKMEM_TINY_STATIC_ROUTE", "1");
  // Phase 3 D1: Free route cache (TLS cache for free path routing, +2.19% proven)
  bench_setenv_default("HAKMEM_FREE_STATIC_ROUTE", "1");
+  // Phase 69-1: Warm Pool Size=16 (+3.26% Strong GO, ENV-only)
+  bench_setenv_default("HAKMEM_WARM_POOL_SIZE", "16");
 }

 static inline void bench_apply_profile(void) {
--- a/core/hakmem_tiny_config.h
+++ b/core/hakmem_tiny_config.h
@ -83,7 +83,10 @@ int tiny_cap_max_for_class(int class_idx);
 // Refill/Drain Configuration
 // ============================================================================

-// Number of blocks to refill from SuperSlab to magazine
+// Number of blocks to refill from SuperSlab to magazine.
+// NOTE: As of 2025-12, this macro is not wired into the active Tiny front path
+// (no call sites). Keep it for documentation/future work, but do not treat it as
+// a performance knob until it is connected to the refill logic.
 #define TINY_REFILL_BATCH_SIZE 16

 // Number of blocks to drain from magazine to SuperSlab
--- a/docs/analysis/PERFORMANCE_TARGETS_SCORECARD.md
+++ b/docs/analysis/PERFORMANCE_TARGETS_SCORECARD.md
@ -27,7 +27,8 @@ mimalloc との比較は **FAST build** で行う（Standard は fixed tax を
 | FAST v3 | 58.478 | 58.876 | 48.34% | 旧 baseline（Phase 59b rebase）。性能評価の正から昇格 → Phase 66 PGO へ |
 | FAST v3 + PGO | 59.80 | 60.25 | 49.41% | Phase 47: NEUTRAL (+0.27% mean, +1.02% median, research box) |
 | **FAST v3 + PGO (Phase 66)** | **60.89** | **61.35** | **50.32%** | **GO: +3.0% mean (3回検証済み、安定 <±1%)**。Phase 66 PGO initial baseline |
-| **FAST v3 + PGO (Phase 68)** | **61.614** | **61.924** | **50.93%** | **GO: +1.19% vs Phase 66** ✓ (seed/WS diversification) → **昇格済み 新 FAST baseline** ✓ |
+| **FAST v3 + PGO (Phase 68)** | **61.614** | **61.924** | **50.93%** | **GO: +1.19% vs Phase 66** ✓ (seed/WS diversification) |
+| **FAST v3 + PGO (Phase 69)** | **62.63** | **63.38** | **51.77%** | **強GO: +3.26% vs Phase 68** ✓✓✓ (Warm Pool Size=16, ENV-only) → **昇格済み 新 FAST baseline** ✓ |
 | Standard | 53.50 | - | 44.21% | 安全・互換基準（Phase 48 前計測、要 rebase） |
 | OBSERVE | TBD | - | - | 診断カウンタ ON |

@ -65,14 +66,14 @@ Notes:

 推奨マイルストーン（Mixed 16–1024B, FAST build）：

-| Milestone | Target | Current (FAST v3 + PGO Phase 68) | Status |
+| Milestone | Target | Current (FAST v3 + PGO Phase 69) | Status |
 |-----------|--------|-----------------------------------|--------|
-| M1 | mimalloc の **50%** | 50.93% | 🟢 **EXCEEDED** (Phase 68 PGO, 10-run verified) |
-| M2 | mimalloc の **55%** | - | 🔴 未達（構造改造必要）|
+| M1 | mimalloc の **50%** | 51.77% | 🟢 **EXCEEDED** (Phase 69, Warm Pool Size=16, ENV-only) |
+| M2 | mimalloc の **55%** | - | 🔴 未達（残り +3.23pp、Phase 69+ 継続中）|
 | M3 | mimalloc の **60%** | - | 🔴 未達（構造改造必要）|
 | M4 | mimalloc の **65–70%** | - | 🔴 未達（構造改造必要）|

-**現状:** FAST v3 + PGO (Phase 68) = 61.614M ops/s = mimalloc の 50.93%（seed/WS diversified, 10-run 検証済み）
+**現状:** FAST v3 + PGO (Phase 69) = 62.63M ops/s = mimalloc の 51.77%（Warm Pool Size=16, ENV-only, 10-run 検証済み）

 **Phase 68 PGO 昇格（Phase 66 → Phase 68 upgrade）:**
 - Phase 66 baseline: 60.89M ops/s = 50.32% (+3.0% mean, 3-run stable)
@ -93,6 +94,26 @@ Notes:
 - More representative of production workload variance
 - Higher confidence in baseline stability

+**Phase 69 PGO 昇格（Phase 68 → Phase 69 upgrade）:**
+- Phase 68 baseline: 61.614M ops/s = 50.93% (+1.19% vs Phase 66, 10-run verified)
+- Phase 69 baseline: 62.63M ops/s = 51.77% (+3.26% vs Phase 68, 10-run verified)
+- Parameter change: Warm Pool Size 12 → 16 (ENV-only, zero code changes)
+- M1 (50%) achievement: **EXCEEDED** (+1.77pp above target, vs +0.93pp in Phase 68)
+- M2 (55%) progress: Gap reduced to +3.23pp (from +4.07pp in Phase 68)
+
+**Phase 69 Benefits Over Phase 68:**
+- +3.26% improvement from warm pool optimization (強GO threshold exceeded)
+- ENV-only change (zero layout tax risk, fully reversible)
+- Reduced registry O(N) scan overhead via larger warm pool
+- Non-additive with other optimizations (Warm Pool Size=16 alone is optimal)
+- Single strongest parameter improvement in refill tuning sweep
+
+**Phase 69 Implementation:**
+- Warm Pool Size: 12 → 16 SuperSlabs/class
+- ENV variable: `HAKMEM_WARM_POOL_SIZE=16` (default in MIXED_TINYV3_C7_SAFE preset)
+- Rollback: Set `HAKMEM_WARM_POOL_SIZE=12` or remove ENV variable
+- Results: `docs/analysis/PHASE69_REFILL_TUNING_1_RESULTS.md`
+
 ※注意: `mimalloc/system/jemalloc` の参照値は環境ドリフトでズレるため、定期的に再ベースラインする。
 - Phase 48 完了: `docs/analysis/PHASE48_REBASE_ALLOCATORS_AND_STABILITY_SUITE_RESULTS.md`
 - Phase 59 完了: `docs/analysis/PHASE59_50PERCENT_RECOVERY_BASELINE_REBASE_RESULTS.md`
--- a/docs/analysis/PHASE69_REFILL_TUNING_0_DESIGN.md
+++ b/docs/analysis/PHASE69_REFILL_TUNING_0_DESIGN.md
@ -33,7 +33,7 @@ malloc() → Unified Cache (C2/C3: 2048 slots, C5-C7: 128 slots)
            ↓
         Warm Pool (12 SuperSlabs per class)
            ↓ MISS
-         sll_refill_batch_from_ss(class_idx, max_take=16)
+         sll_refill_*_from_ss(class_idx, max_take=refill_count(class_idx))
            ↓
         trc_pop_from_freelist() OR trc_linear_carve()
            ↓
@ -44,12 +44,11 @@ malloc() → Unified Cache (C2/C3: 2048 slots, C5-C7: 128 slots)

 | Component | Parameter | Current Value | Location |
 |-----------|-----------|---------------|----------|
-| **Refill Batch Size** | `TINY_REFILL_BATCH_SIZE` | 16 | `core/hakmem_tiny_config.h:87` |
+| **Refill Count (hot/mid/global)** | `HAKMEM_TINY_REFILL_COUNT_*` | hot=128, mid=96, global=64 (defaults) | `core/hakmem_tiny_init.inc:270` |
 | **Unified Cache C2/C3** | `unified_capacity(2/3)` | 2048 slots | `core/front/tiny_unified_cache.h:129` |
 | **Unified Cache C5-C7** | `unified_capacity(5-7)` | 128 slots | `core/front/tiny_unified_cache.h:131` |
 | **Unified Cache Others** | `unified_capacity(0/1/4)` | 64 slots | `core/front/tiny_unified_cache.h:128` |
 | **Warm Pool Size** | `TINY_WARM_POOL_MAX_PER_CLASS` | 12 | `core/front/tiny_warm_pool.h:46` |
-| **Drain Batch Size** | `TINY_DRAIN_BATCH_SIZE` | 16 | `core/hakmem_tiny_config.h:90` |

 ### 1.3 Refill Overhead Breakdown

@ -72,35 +71,31 @@ Total Refill Cost = (refill_count × fixed_overhead) + (blocks_carved × per_blo

 ## 2. Tunable Parameters (Phase 69 Sweep Plan)

-### 2.1 Refill Batch Size Sweep
+### 2.1 Refill Count Sweep (ENV-only)

-**Parameter**: `TINY_REFILL_BATCH_SIZE` (global default for all classes)
+**Parameter**:
+- `HAKMEM_TINY_REFILL_COUNT_HOT` (classes C0–C3)
+- `HAKMEM_TINY_REFILL_COUNT_MID` (classes C4–C7)
+- `HAKMEM_TINY_REFILL_COUNT` (fallback/global)

-**Current**: 16 (conservative, optimized for RSS)
-
-**Sweep Range**: [16, 32, 64, 128]
+**Current defaults** (when ENV unset):
+- hot=128, mid=96, global=64

 **Rationale**:
- 16 → 32: Expected +1-2% (fewer refill calls)
- 32 → 64: Expected +1-2% (diminishing returns, cache pressure increases)
- 64 → 128: Expected +0-1% (likely NO-GO due to cache thrashing)
+- Smaller counts reduce per-refill work but increase refill frequency.
+- Larger counts reduce refill frequency but may increase chain-building cost and TLS pressure.
+- Sweep is cheap (ENV-only) and reversible.

 **A/B Test Method**:
 ```bash
-# Baseline (16)
-make clean && make bench_random_mixed_hakmem_minimal_pgo
-ITERS=20000000 WS=400 RUNS=10 scripts/run_mixed_10_cleanenv.sh
+# Baseline
+RUNS=10 BENCH_BIN=./bench_random_mixed_hakmem_minimal_pgo scripts/run_mixed_10_cleanenv.sh

-# Treatment (32)
-sed -i 's/TINY_REFILL_BATCH_SIZE 16/TINY_REFILL_BATCH_SIZE 32/' core/hakmem_tiny_config.h
-make clean && make pgo-fast-full
-ITERS=20000000 WS=400 RUNS=10 scripts/run_mixed_10_cleanenv.sh
-
-# Compare results
+# Treatment examples (pick one axis at a time)
+HAKMEM_TINY_REFILL_COUNT_MID=128 RUNS=10 BENCH_BIN=./bench_random_mixed_hakmem_minimal_pgo scripts/run_mixed_10_cleanenv.sh
+HAKMEM_TINY_REFILL_COUNT_MID=64  RUNS=10 BENCH_BIN=./bench_random_mixed_hakmem_minimal_pgo scripts/run_mixed_10_cleanenv.sh
 ```

-**Expected Winner**: 32 (balance between refill frequency and cache locality)
-
 ---

 ### 2.2 Unified Cache Capacity Sweep (C5-C7 Focus)
@ -183,27 +178,18 @@ HAKMEM_WARM_POOL_SIZE=24  # Treatment 2
   - Expected: +1-2% (miss rate reduction for mid-size allocations)
   - Risk: Low (ENV-only change)

-3. **Refill Batch Size** (requires recompile + PGO):
-   ```bash
-   for BATCH in 16 32 64; do
-     sed -i "s/TINY_REFILL_BATCH_SIZE .*/TINY_REFILL_BATCH_SIZE $BATCH/" core/hakmem_tiny_config.h
-     make pgo-fast-full
-     RUNS=10 scripts/run_mixed_10_cleanenv.sh
-   done
-   ```
-   - Expected: +1-3% (refill frequency reduction)
-   - Risk: Medium (requires PGO rebuild, potential layout tax)
+3. **Refill Batch Size**（削除, 2025-12 audit）:
+   `TINY_REFILL_BATCH_SIZE` は現行 Tiny front に接続されておらず、knob として成立していない。
+   実際の refill 量は `HAKMEM_TINY_REFILL_COUNT_*` で制御する（ENV-only）。

 ### 3.2 Phase 69-2: Combined Optimization (Best Settings)

 After identifying winners from Phase 69-1, combine them:

 ```bash
-# Example: batch=32, C5-C7=256, warm_pool=16
+# Example: C5-C7=256, warm_pool=16
 HAKMEM_WARM_POOL_SIZE=16 \
 HAKMEM_TINY_UNIFIED_C5=256 HAKMEM_TINY_UNIFIED_C6=256 HAKMEM_TINY_UNIFIED_C7=256 \
-  make pgo-fast-full  # with TINY_REFILL_BATCH_SIZE=32
-
  RUNS=10 scripts/run_mixed_10_cleanenv.sh
 ```

@ -289,8 +275,9 @@ atomic_fetch_add(&g_refill_count_total, 1, memory_order_relaxed);
   - Warm pool size: 12 → 16 → 24
   - Unified cache C5-C7: 128 → 256 → 512

-2. **Batch Size Sweep** (requires PGO rebuild):
-   - TINY_REFILL_BATCH_SIZE: 16 → 32 → 64
+2. **Refill Count Sweep** (ENV-only):
+   - `HAKMEM_TINY_REFILL_COUNT_MID`: 64 → 96 → 128 → 160 (example)
+   - `HAKMEM_TINY_REFILL_COUNT_HOT`: 96 → 128 → 160 (example)

 ### Follow-Up (Phase 69-2)

@ -300,7 +287,6 @@ atomic_fetch_add(&g_refill_count_total, 1, memory_order_relaxed);

 4. **Baseline Promotion** (if Strong GO):
   - Update `pgo_fast_profile_config.sh` with winning ENV vars
-   - Update `core/hakmem_tiny_config.h` with winning batch size
   - Re-run `make pgo-fast-full` to bake optimizations into baseline

 ---
--- a/scripts/run_mixed_10_cleanenv.sh
+++ b/scripts/run_mixed_10_cleanenv.sh
@ -39,6 +39,8 @@ export HAKMEM_FASTLANE_DIRECT=${HAKMEM_FASTLANE_DIRECT:-1}
 # NOTE: Phase 9/10 are promoted (bench_profile defaults to 1). Keep cleanenv aligned by default.
 export HAKMEM_FREE_TINY_FAST_MONO_DUALHOT=${HAKMEM_FREE_TINY_FAST_MONO_DUALHOT:-1}
 export HAKMEM_FREE_TINY_FAST_MONO_LEGACY_DIRECT=${HAKMEM_FREE_TINY_FAST_MONO_LEGACY_DIRECT:-1}
+# NOTE: Phase 69-1 winner (Warm Pool Size=16, +3.26% Strong GO, ENV-only)
+export HAKMEM_WARM_POOL_SIZE=${HAKMEM_WARM_POOL_SIZE:-16}

 for i in $(seq 1 "${runs}"); do
  echo "=== Run ${i}/${runs} ==="