Fix circuit gas diff: Implement u96_limbs_less_than_guarantee_verify

[JulianGCalderon]

Fixes the gas issues related to circuits, by properly implementing u96_limbs_less_than_guarantee_verify

Types

The type CircuitOutputs used to contain an array of gate values, in integer form (u384). This PR changes two things:

• Now the gate values are stored in struct form (struct of 4 x u96). This was already being done in get_circuit_output, so I moved the logic to the eval_circuit libfunc.
• Now it also contains the circuit modulus, as it is required for the get_circuit_output libfunc.

The type U96LimbsLessThanGuarantee was a unit type as the related libfunc was not implemented. Now it contains the associated gate value, and the circuit modulus.

Libfuncs

The eval_circuit libfunc was changed to output the updated version of CircuitOutputs.

The get_circuit_output libfunc was changed to output the updated version of U96LimbsLessThanGuarantee.

The u96_limbs_less_than_guarantee_verify libfunc was implemented (it was a noop before this):

• compares the last limb of the circuit value with the last limb of the modulus.
• if they differ, it returns the difference.
• if it does not differ, it returns the received value (U96LimbsLessThanGuarantee), without the last limb.

[github-actions]

Benchmark results Main vs HEAD.

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_insert.cairo (JIT)	3.883 ± 0.027	3.854	3.926	1.02 ± 0.01
base dict_insert.cairo (AOT)	3.809 ± 0.025	3.771	3.840	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_insert.cairo (JIT)	3.815 ± 0.025	3.772	3.855	1.03 ± 0.01
head dict_insert.cairo (AOT)	3.713 ± 0.017	3.688	3.741	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_snapshot.cairo (JIT)	3.769 ± 0.021	3.733	3.802	1.03 ± 0.01
base dict_snapshot.cairo (AOT)	3.665 ± 0.021	3.636	3.702	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_snapshot.cairo (JIT)	3.706 ± 0.031	3.637	3.747	1.01 ± 0.02
head dict_snapshot.cairo (AOT)	3.666 ± 0.059	3.576	3.756	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base factorial_2M.cairo (JIT)	4.111 ± 0.019	4.079	4.140	1.01 ± 0.01
base factorial_2M.cairo (AOT)	4.077 ± 0.023	4.043	4.112	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head factorial_2M.cairo (JIT)	4.038 ± 0.039	3.992	4.097	1.01 ± 0.01
head factorial_2M.cairo (AOT)	4.011 ± 0.033	3.967	4.086	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base fib_2M.cairo (JIT)	3.668 ± 0.029	3.624	3.710	1.02 ± 0.01
base fib_2M.cairo (AOT)	3.599 ± 0.013	3.576	3.615	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head fib_2M.cairo (JIT)	3.602 ± 0.027	3.538	3.627	1.00 ± 0.02
head fib_2M.cairo (AOT)	3.600 ± 0.048	3.521	3.667	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base linear_search.cairo (JIT)	3.843 ± 0.026	3.805	3.898	1.04 ± 0.01
base linear_search.cairo (AOT)	3.705 ± 0.018	3.675	3.727	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head linear_search.cairo (JIT)	3.820 ± 0.051	3.776	3.917	1.04 ± 0.02
head linear_search.cairo (AOT)	3.662 ± 0.021	3.622	3.688	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base logistic_map.cairo (JIT)	3.926 ± 0.013	3.906	3.943	1.04 ± 0.01
base logistic_map.cairo (AOT)	3.768 ± 0.023	3.740	3.808	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head logistic_map.cairo (JIT)	3.943 ± 0.031	3.891	3.988	1.06 ± 0.01
head logistic_map.cairo (AOT)	3.729 ± 0.031	3.666	3.779	1.00

[github-actions]

Benchmarking results

Benchmark for program dict_insert

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	20.930 ± 0.192	20.786	21.437	5.53 ± 0.07
cairo-native (embedded AOT)	3.784 ± 0.030	3.725	3.833	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	3.838 ± 0.026	3.800	3.869	1.01 ± 0.01

Benchmark for program dict_snapshot

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	5.845 ± 0.037	5.762	5.886	1.57 ± 0.03
cairo-native (embedded AOT)	3.713 ± 0.060	3.636	3.811	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	3.758 ± 0.019	3.729	3.790	1.01 ± 0.02

Benchmark for program factorial_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	14.469 ± 0.074	14.374	14.627	3.58 ± 0.04
cairo-native (embedded AOT)	4.043 ± 0.039	3.994	4.113	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	4.077 ± 0.025	4.039	4.127	1.01 ± 0.01

Benchmark for program fib_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	14.443 ± 0.056	14.382	14.538	4.01 ± 0.03
cairo-native (embedded AOT)	3.600 ± 0.027	3.558	3.642	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	3.642 ± 0.038	3.594	3.726	1.01 ± 0.01

Benchmark for program linear_search

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	5.837 ± 0.028	5.794	5.866	1.59 ± 0.01
cairo-native (embedded AOT)	3.679 ± 0.021	3.649	3.723	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	3.796 ± 0.019	3.761	3.816	1.03 ± 0.01

Benchmark for program logistic_map

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	5.913 ± 0.054	5.814	5.996	1.54 ± 0.02
cairo-native (embedded AOT)	3.840 ± 0.017	3.820	3.880	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	4.033 ± 0.052	3.975	4.122	1.05 ± 0.01

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 99.41520% with 1 line in your changes missing coverage. Please review.

Project coverage is 81.05%. Comparing base (c4cde60) to head (3e609eb).

Files with missing lines	Patch %	Lines
src/compiler.rs	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1084      +/-   ##
==========================================
+ Coverage   80.93%   81.05%   +0.11%     
==========================================
  Files         108      108              
  Lines       29605    29726     +121     
==========================================
+ Hits        23961    24094     +133     
+ Misses       5644     5632      -12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.