arena for dict

[TomerStarkware]

Move felt252_dict storage to the per-execution arena

Closes #NA

Moves FeltDict allocation off the system heap and onto the per-execution Bump arena, completing the arena migration for boxed values, nullables,
arrays, and dictionaries.

Before, each dict was Rc::into_raw(Rc::new(FeltDict { ... })) with its element buffer on the system heap (alloc/realloc/dealloc) and dup/drop
wired through runtime bindings + per-value drop overrides materialized via Felt252DictOverrides + symbol lookup. Now:

• cairo_native__dict_new arena-allocates the FeltDict struct and registers it in a thread-local DICT_REGISTRY.
• cairo_native__dict_get grows the elements buffer via EXECUTION_ARENA.alloc_layout + copy_nonoverlapping instead of realloc.
• Dict dup is the default (returns the same pointer twice) — safe because dict snapshots in Cairo cannot read inner values. Dict drop is a noop.
• On InvocationGuard::drop, the registry is drained and each FeltDict's HashMap is drop_in_place'd (the HashMap is the only system-heap
  allocation that survives in the FeltDict). The arena is then reset, freeing the FeltDict struct and elements buffer in one shot.

This removes:

• src/metadata/felt252_dict.rs (Felt252DictOverrides).
• The cairo_native__dict_drop / cairo_native__dict_dup runtime bindings.
• The custom Drop for FeltDict.
• The find_dict_drop_override closure threaded through every AbiArgument::to_bytes impl and through Value::to_ptr.

Net diff: -894 / +262.

Introduces Breaking Changes?

Yes.

• AotNativeExecutor::new and JitNativeExecutor::new no longer take a Felt252DictOverrides argument.
• AbiArgument::to_bytes no longer takes the find_dict_drop_override closure.
• Value::to_ptr no longer takes the find_dict_drop_override closure.
• FeltDict no longer has a drop_fn field, no longer impls Drop, and cairo_native__dict_new no longer takes a drop callback.
• cairo_native__dict_drop and cairo_native__dict_dup symbols are removed.

---

This change is 

[orizi]

@orizi reviewed 14 files and all commit messages, and made 1 comment.
Reviewable status: 14 of 16 files reviewed, 1 unresolved discussion (waiting on TomerStarkware).

---

src/executor.rs line 360 at r1 (raw file):

    blake_call_count: u64,
    execution_arena: Bump,
    dict_registry: Vec<*mut FeltDict>,

probably not really needed.

add todo

[TomerStarkware]

@TomerStarkware made 1 comment.
Reviewable status: 14 of 20 files reviewed, 1 unresolved discussion (waiting on orizi).

---

src/executor.rs line 360 at r1 (raw file):

Previously, orizi wrote…

probably not really needed.

add todo

Done.

[github-actions]

Benchmarking results

Benchmark for program dict_insert

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	11.251 ± 0.036	11.196	11.304	6.09 ± 0.08
cairo-native (embedded AOT)	1.848 ± 0.023	1.815	1.875	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	1.888 ± 0.018	1.862	1.922	1.02 ± 0.02

Benchmark for program dict_snapshot

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	555.0 ± 10.4	543.6	577.7	1.00
cairo-native (embedded AOT)	1682.2 ± 35.3	1626.4	1745.4	3.03 ± 0.09
cairo-native (embedded JIT using LLVM's ORC Engine)	1731.2 ± 10.8	1720.2	1756.7	3.12 ± 0.06

Benchmark for program factorial_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	4.890 ± 0.104	4.802	5.102	2.27 ± 0.05
cairo-native (embedded AOT)	2.153 ± 0.014	2.127	2.177	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	2.198 ± 0.020	2.179	2.237	1.02 ± 0.01

Benchmark for program fib_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	4.831 ± 0.073	4.756	4.949	2.91 ± 0.06
cairo-native (embedded AOT)	1.660 ± 0.022	1.630	1.689	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	1.705 ± 0.025	1.676	1.760	1.03 ± 0.02

Benchmark for program linear_search

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	591.3 ± 8.4	579.3	607.9	1.00
cairo-native (embedded AOT)	1692.0 ± 17.5	1666.7	1712.5	2.86 ± 0.05
cairo-native (embedded JIT using LLVM's ORC Engine)	1784.7 ± 21.6	1765.8	1837.7	3.02 ± 0.06

Benchmark for program logistic_map

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	500.8 ± 4.1	495.8	509.2	1.00
cairo-native (embedded AOT)	1835.6 ± 14.5	1798.1	1849.7	3.67 ± 0.04
cairo-native (embedded JIT using LLVM's ORC Engine)	1999.3 ± 16.2	1977.8	2034.4	3.99 ± 0.05

[github-actions]

Benchmark results Main vs HEAD.

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_insert.cairo (JIT)	1.895 ± 0.009	1.884	1.913	1.04 ± 0.01
base dict_insert.cairo (AOT)	1.817 ± 0.022	1.780	1.841	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_insert.cairo (JIT)	1.843 ± 0.030	1.801	1.906	1.03 ± 0.02
head dict_insert.cairo (AOT)	1.796 ± 0.022	1.766	1.828	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_snapshot.cairo (JIT)	1.768 ± 0.021	1.740	1.807	1.05 ± 0.02
base dict_snapshot.cairo (AOT)	1.677 ± 0.021	1.648	1.710	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_snapshot.cairo (JIT)	1.696 ± 0.025	1.669	1.740	1.06 ± 0.02
head dict_snapshot.cairo (AOT)	1.604 ± 0.015	1.578	1.625	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base factorial_2M.cairo (JIT)	2.168 ± 0.024	2.134	2.198	1.03 ± 0.01
base factorial_2M.cairo (AOT)	2.112 ± 0.010	2.095	2.128	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head factorial_2M.cairo (JIT)	2.014 ± 0.028	1.962	2.055	1.01 ± 0.02
head factorial_2M.cairo (AOT)	1.988 ± 0.034	1.937	2.051	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base fib_2M.cairo (JIT)	1.714 ± 0.021	1.688	1.743	1.03 ± 0.02
base fib_2M.cairo (AOT)	1.672 ± 0.015	1.649	1.692	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head fib_2M.cairo (JIT)	1.656 ± 0.018	1.629	1.685	1.03 ± 0.02
head fib_2M.cairo (AOT)	1.610 ± 0.016	1.589	1.636	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base linear_search.cairo (JIT)	1.814 ± 0.017	1.780	1.833	1.06 ± 0.02
base linear_search.cairo (AOT)	1.704 ± 0.023	1.671	1.743	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head linear_search.cairo (JIT)	1.738 ± 0.033	1.690	1.790	1.04 ± 0.03
head linear_search.cairo (AOT)	1.669 ± 0.032	1.629	1.722	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base logistic_map.cairo (JIT)	2.003 ± 0.026	1.960	2.055	1.06 ± 0.02
base logistic_map.cairo (AOT)	1.890 ± 0.021	1.868	1.925	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head logistic_map.cairo (JIT)	1.972 ± 0.027	1.937	2.027	1.10 ± 0.02
head logistic_map.cairo (AOT)	1.796 ± 0.027	1.736	1.832	1.00

[orizi]

@orizi reviewed all commit messages and made 1 comment.
Reviewable status: 12 of 20 files reviewed, 2 unresolved discussions (waiting on TomerStarkware).

---

-- commits line 4 at r2:
rebase

[orizi]

@orizi reviewed 8 files and all commit messages, made 1 comment, and resolved 1 discussion.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on TomerStarkware).

---

src/executor.rs line 360 at r1 (raw file):

Previously, TomerStarkware wrote…

Done.

add the why

[orizi]

@orizi reviewed 1 file and all commit messages, made 1 comment, and resolved 1 discussion.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on TomerStarkware).