⚡ Bolt: replace dict.setdefault with explicit membership checks in hot loops

.jules/bolt.md

@@ -0,0 +1,3 @@
+## 2024-05-24 - [Avoid `dict.setdefault` in hot loops]
+**Learning:** `dict.setdefault(key, default)` evaluates the `default` expression eagerly on *every* loop iteration before performing the membership check. In cases where the default value requires complex allocation or computation (e.g., `{"value_sum": [Fraction(0) for _ in value]}`), this results in significant performance degradation in hot loops.
+**Action:** Replace `dict.setdefault` in hot loops with explicit `if key not in dict:` membership checks and lazy assignment.

scripts/export_h2_bundle_lock_audit.py

@@ -4,7 +4,6 @@
 
 import json
 from pathlib import Path
-from typing import Any
 
 from utils import detect_runtime_environment
 

scripts/export_p3_paper_freeze.py

@@ -6,7 +6,6 @@
 import json
 from pathlib import Path
 import re
-from typing import Any
 
 from utils import detect_runtime_environment
 

scripts/export_p5_public_surface_sync.py

@@ -4,7 +4,6 @@
 
 import json
 from pathlib import Path
-from typing import Any
 
 from utils import detect_runtime_environment
 

scripts/export_r10_d0_same_endpoint_cost_attribution.py

@@ -21,7 +21,7 @@
     subroutine_braid_long_program,
     verify_program,
 )
-from exec_trace import ExecutionState, Program, TraceEvent, replay_trace, TraceInterpreter
+from exec_trace import Program, TraceEvent, replay_trace, TraceInterpreter
 from geometry import brute_force_hardmax_2d
 from model.exact_hardmax import encode_latest_write_query
 from model.free_running_executor import (

scripts/export_r26_d0_boundary_localization_execution_gate.py

@@ -4,7 +4,6 @@
 
 from pathlib import Path
 import sys
-from typing import Any
 
 SCRIPT_DIR = Path(__file__).resolve().parent
 if str(SCRIPT_DIR) not in sys.path:

scripts/export_r2_systems_baseline_gate.py

@@ -2,7 +2,7 @@
 
 from __future__ import annotations
 
-from collections import Counter, defaultdict
+from collections import defaultdict
 import csv
 import json
 from pathlib import Path

scripts/export_r3_d0_exact_execution_stress_gate.py

@@ -4,7 +4,6 @@
 
 import json
 from pathlib import Path
-from typing import Any
 
 from bytecode import (
     lower_program,

scripts/export_r47_origin_restricted_frontend_translation_gate.py

@@ -12,7 +12,6 @@
     BytecodeMemoryRegion,
     BytecodeOpcode,
     BytecodeProgram,
-    RestrictedFrontendTranslationCase,
     compile_restricted_frontend_program,
     first_divergence_step,
     lower_program,

scripts/export_r50_origin_restricted_tinyc_lowering_gate.py

@@ -12,7 +12,6 @@
     BytecodeMemoryRegion,
     BytecodeOpcode,
     BytecodeProgram,
-    RestrictedTinyCLoweringCase,
     compile_restricted_tinyc_program,
     first_divergence_step,
     lower_program,

scripts/export_release_worktree_hygiene_snapshot.py

@@ -6,7 +6,6 @@
 import os
 from pathlib import Path
 import subprocess
-from typing import Any
 
 from utils import detect_runtime_environment
 

src/geometry/hull_kv.py

@@ -17,7 +17,6 @@
     HardmaxResult,
     NumberLike,
     ValueLike,
-    _as_fraction,
     _coerce_key,
     _coerce_value,
     _normalize_number,
@@ -205,10 +204,11 @@ def _rebuild_if_needed(self) -> None:
         total_value_sum = [Fraction(0) for _ in range(self._value_width or 0)]
 
         for index, (key, value) in enumerate(self._entries):
-            bucket = aggregates.setdefault(
-                key,
-                {"value_sum": [Fraction(0) for _ in value], "count": 0, "entry_indices": []},
-            )
+            # Performance optimization: Replace `setdefault` with an explicit membership check
+            # to avoid eagerly allocating `[Fraction(0) for _ in value]` on every loop iteration.
+            if key not in aggregates:
+                aggregates[key] = {"value_sum": [Fraction(0) for _ in value], "count": 0, "entry_indices": []}
+            bucket = aggregates[key]
             for coord_index, coord in enumerate(value):
                 bucket["value_sum"][coord_index] += coord
                 total_value_sum[coord_index] += coord

src/model/free_running_executor.py

@@ -697,14 +697,15 @@ def evaluate_free_running_programs(
 
         outcomes.append(outcome)
         bucket = bucket_name(outcome.program_steps)
-        bucket_state = per_bucket.setdefault(
-            bucket,
-            {
+        # Performance optimization: Replace `setdefault` with an explicit check
+        # to avoid allocating new dictionary objects on every iteration.
+        if bucket not in per_bucket:
+            per_bucket[bucket] = {
                 "program_count": 0,
                 "exact_trace_count": 0,
                 "exact_final_state_count": 0,
-            },
-        )
+            }
+        bucket_state = per_bucket[bucket]
         bucket_state["program_count"] += 1
         bucket_state["exact_trace_count"] += int(outcome.exact_trace_match)
         bucket_state["exact_final_state_count"] += int(outcome.exact_final_state_match)

src/model/induced_causal.py

@@ -455,7 +455,11 @@ def fit_transition_library(
     examples = build_transition_examples(programs, interpreter=interpreter)
     by_opcode: dict[Opcode, list[TransitionExample]] = {}
     for example in examples:
-        by_opcode.setdefault(example.opcode, []).append(example)
+        # Performance optimization: Replace `setdefault` with an explicit check
+        # to avoid list allocation overhead on every loop iteration.
+        if example.opcode not in by_opcode:
+            by_opcode[example.opcode] = []
+        by_opcode[example.opcode].append(example)
 
     rules = tuple(_fit_rule_for_opcode(by_opcode[opcode], opcode) for opcode in sorted(by_opcode, key=str))
     return InducedTransitionLibrary(rules=rules)

src/model/r45_dual_mode.py

@@ -2,7 +2,7 @@
 
 from dataclasses import dataclass
 
-from bytecode import BoundedMemoryVMCase, lower_program, r43_bounded_memory_vm_cases
+from bytecode import lower_program, r43_bounded_memory_vm_cases
 from exec_trace import Program, TraceInterpreter
 
 from .exact_hardmax import extract_stack_slot_operations

src/model/softmax_baseline.py

@@ -582,15 +582,16 @@ def evaluate_teacher_forced_model(
                 total_correct += correct
 
                 bucket = baseline_bucket_name(example.program_steps)
-                bucket_state = per_bucket.setdefault(
-                    bucket,
-                    {
+                # Performance optimization: Replace `setdefault` with an explicit check
+                # to avoid dictionary allocation overhead on every loop iteration.
+                if bucket not in per_bucket:
+                    per_bucket[bucket] = {
                         "example_count": 0,
                         "token_count": 0,
                         "correct_tokens": 0,
                         "weighted_loss": 0.0,
-                    },
-                )
+                    }
+                bucket_state = per_bucket[bucket]
                 bucket_state["example_count"] = int(bucket_state["example_count"]) + 1
                 bucket_state["token_count"] = int(bucket_state["token_count"]) + token_count
                 bucket_state["correct_tokens"] = int(bucket_state["correct_tokens"]) + correct
@@ -753,7 +754,11 @@ def evaluate_free_running_rollout(
             )
 
             bucket = baseline_bucket_name(example.program_steps)
-            bucket_state = per_bucket.setdefault(bucket, {"example_count": 0, "exact_count": 0})
+            # Performance optimization: Replace `setdefault` with an explicit check
+            # to avoid dictionary allocation overhead on every loop iteration.
+            if bucket not in per_bucket:
+                per_bucket[bucket] = {"example_count": 0, "exact_count": 0}
+            bucket_state = per_bucket[bucket]
             bucket_state["example_count"] = int(bucket_state["example_count"]) + 1
             bucket_state["exact_count"] = int(bucket_state["exact_count"]) + int(exact)
 

src/model/trainable_latest_write.py

@@ -171,7 +171,11 @@ def exact_program_accuracy(scorer: TrainableLatestWriteScorer, samples: Sequence
         return 0.0
     per_program: dict[str, list[bool]] = {}
     for sample in samples:
-        per_program.setdefault(sample.program_name, []).append(scorer.predict_index(sample) == sample.target_index)
+        # Performance optimization: Replace `setdefault` with an explicit check
+        # to avoid list allocation overhead on every loop iteration.
+        if sample.program_name not in per_program:
+            per_program[sample.program_name] = []
+        per_program[sample.program_name].append(scorer.predict_index(sample) == sample.target_index)
     exact = sum(1 for outcomes in per_program.values() if all(outcomes))
     return exact / len(per_program)
 
@@ -198,12 +202,21 @@ def evaluate_scorer(
         correct_samples += int(correct)
 
         bucket = bucket_name(sample.program_steps)
-        bucket_state = per_bucket.setdefault(bucket, {"sample_count": 0, "sample_correct": 0, "programs": {}})
+        # Performance optimization: Replace `setdefault` with an explicit check
+        # to avoid dictionary allocation overhead on every loop iteration.
+        if bucket not in per_bucket:
+            per_bucket[bucket] = {"sample_count": 0, "sample_correct": 0, "programs": {}}
+        bucket_state = per_bucket[bucket]
         bucket_state["sample_count"] = int(bucket_state["sample_count"]) + 1
         bucket_state["sample_correct"] = int(bucket_state["sample_correct"]) + int(correct)
-        bucket_state["programs"].setdefault(sample.program_name, []).append(correct)
 
-        per_program.setdefault(sample.program_name, []).append(correct)
+        if sample.program_name not in bucket_state["programs"]:
+            bucket_state["programs"][sample.program_name] = []
+        bucket_state["programs"][sample.program_name].append(correct)
+
+        if sample.program_name not in per_program:
+            per_program[sample.program_name] = []
+        per_program[sample.program_name].append(correct)
         program_steps[sample.program_name] = sample.program_steps
 
     exact_programs = sum(1 for outcomes in per_program.values() if all(outcomes))

tests/test_bytecode_harness.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
 from bytecode import (
-    accumulator_loop_program,
     alternating_memory_loop_bytecode_program,
     arithmetic_smoke_program,
     branch_then_call_false_program,

tests/test_bytecode_memory_surfaces.py

@@ -7,7 +7,6 @@
 from bytecode import (
     analyze_memory_surfaces,
     call_frame_roundtrip_program,
-    countdown_helper_call_program,
     memory_surface_cases,
     memory_surface_negative_programs,
     run_memory_surface_case,

tests/test_model_softmax_baseline.py

@@ -13,7 +13,6 @@
     evaluate_free_running_rollout,
     evaluate_teacher_forced_model,
     require_torch,
-    serialize_event_tokens,
     SoftmaxBaselineConfig,
     SoftmaxTrainingConfig,
     summarize_trace_sequences,