chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237

Forostovec · 2025-10-19T11:20:25Z

Read a/b/div via Uint256::from_var_name(...).pack() in uint256_mul_div_mod
Use Uint256 in uint256_signed_nn for consistent access to a.high
Add Uint256::from_base_addr_with_offsets and use it in uint256_offseted_unsigned_div_rem (standard layout uses from_var_name, expanded layout uses offsets)
Form quotient_low/quotient_high via Uint512::split into two Uint256, and write remainder via Uint256

These changes remove duplicated packing logic, standardize error paths, and align uint256 hint implementations with the existing UintNNN helper patterns

…ia Uint512

JulianGCalderon

Hey @Forostovec, thanks for the contribution! Could you also update the changelog?

JulianGCalderon · 2025-10-20T13:44:36Z

vm/src/hint_processor/builtin_hint_processor/uint256_utils.rs

+    let a = Uint256::from_var_name("a", vm, ids_data, ap_tracking)?;
+    let a_high = a.high;
    //Main logic
    //memory[ap] = 1 if 0 <= (ids.a.high % PRIME) < 2 ** 127 else 0
    let result: Felt252 =
-        if *a_high >= Felt252::ZERO && a_high.as_ref() <= &Felt252::from(i128::MAX) {
+        if *a_high.as_ref() >= Felt252::ZERO && a_high.as_ref() <= &Felt252::from(i128::MAX) {


This would imply fetching a.low, but not using it, right?

JulianGCalderon · 2025-10-20T14:36:51Z

vm/src/hint_processor/builtin_hint_processor/uint256_utils.rs

+    let div = if div_offset_low == 0 && div_offset_high == 1 {
+        // Standard Uint256 layout
+        Uint256::from_var_name("div", vm, ids_data, ap_tracking)?
+    } else {
+        let div_addr = get_relocatable_from_var_name("div", vm, ids_data, ap_tracking)?;
+        Uint256::from_base_addr_with_offsets(div_addr, "div", vm, div_offset_low, div_offset_high)?
+    };
+    let div_low = div.low.as_ref();
+    let div_high = div.high.as_ref();


Here could we call from_base_addr_with_offsets directly and avoid the conditional? It would make the function easier to read.

Also, in this function we are packing the U256 into a bigint manually. Could you modify it to call pack? Given that this PR is for unifying U256 usage in general, we can include it.

gabrielbosio · 2025-12-02T14:23:20Z

Hi, @Forostovec, what's the status of this PR?

Forostovec · 2025-12-02T19:27:04Z

Hi, @Forostovec, what's the status of this PR?

Ohh, I'm sorry, gonna make chamges and update changelog

gabrielbosio · 2025-12-19T21:40:02Z

**Hyper Thereading Benchmark results**




hyperfine -r 2 -n "hyper_threading_main threads: 1" 'RAYON_NUM_THREADS=1 ./hyper_threading_main' -n "hyper_threading_pr threads: 1" 'RAYON_NUM_THREADS=1 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 1
  Time (mean ± σ):     22.606 s ±  0.006 s    [User: 21.754 s, System: 0.849 s]
  Range (min … max):   22.602 s … 22.610 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 1
  Time (mean ± σ):     22.661 s ±  0.049 s    [User: 21.821 s, System: 0.837 s]
  Range (min … max):   22.627 s … 22.695 s    2 runs
 
Summary
  hyper_threading_main threads: 1 ran
    1.00 ± 0.00 times faster than hyper_threading_pr threads: 1




hyperfine -r 2 -n "hyper_threading_main threads: 2" 'RAYON_NUM_THREADS=2 ./hyper_threading_main' -n "hyper_threading_pr threads: 2" 'RAYON_NUM_THREADS=2 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 2
  Time (mean ± σ):     12.191 s ±  0.034 s    [User: 21.941 s, System: 0.842 s]
  Range (min … max):   12.167 s … 12.215 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 2
  Time (mean ± σ):     12.121 s ±  0.009 s    [User: 21.813 s, System: 0.870 s]
  Range (min … max):   12.115 s … 12.128 s    2 runs
 
Summary
  hyper_threading_pr threads: 2 ran
    1.01 ± 0.00 times faster than hyper_threading_main threads: 2




hyperfine -r 2 -n "hyper_threading_main threads: 4" 'RAYON_NUM_THREADS=4 ./hyper_threading_main' -n "hyper_threading_pr threads: 4" 'RAYON_NUM_THREADS=4 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 4
  Time (mean ± σ):      9.467 s ±  0.168 s    [User: 34.527 s, System: 1.078 s]
  Range (min … max):    9.348 s …  9.586 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 4
  Time (mean ± σ):      9.651 s ±  0.349 s    [User: 34.346 s, System: 1.034 s]
  Range (min … max):    9.403 s …  9.898 s    2 runs
 
Summary
  hyper_threading_main threads: 4 ran
    1.02 ± 0.04 times faster than hyper_threading_pr threads: 4




hyperfine -r 2 -n "hyper_threading_main threads: 6" 'RAYON_NUM_THREADS=6 ./hyper_threading_main' -n "hyper_threading_pr threads: 6" 'RAYON_NUM_THREADS=6 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 6
  Time (mean ± σ):      9.586 s ±  0.088 s    [User: 34.396 s, System: 1.042 s]
  Range (min … max):    9.524 s …  9.649 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 6
  Time (mean ± σ):      9.542 s ±  0.172 s    [User: 34.313 s, System: 1.024 s]
  Range (min … max):    9.420 s …  9.664 s    2 runs
 
Summary
  hyper_threading_pr threads: 6 ran
    1.00 ± 0.02 times faster than hyper_threading_main threads: 6




hyperfine -r 2 -n "hyper_threading_main threads: 8" 'RAYON_NUM_THREADS=8 ./hyper_threading_main' -n "hyper_threading_pr threads: 8" 'RAYON_NUM_THREADS=8 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 8
  Time (mean ± σ):      9.440 s ±  0.185 s    [User: 35.055 s, System: 1.108 s]
  Range (min … max):    9.310 s …  9.571 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 8
  Time (mean ± σ):      9.368 s ±  0.058 s    [User: 34.916 s, System: 1.068 s]
  Range (min … max):    9.327 s …  9.409 s    2 runs
 
Summary
  hyper_threading_pr threads: 8 ran
    1.01 ± 0.02 times faster than hyper_threading_main threads: 8




hyperfine -r 2 -n "hyper_threading_main threads: 16" 'RAYON_NUM_THREADS=16 ./hyper_threading_main' -n "hyper_threading_pr threads: 16" 'RAYON_NUM_THREADS=16 ./hyper_threading_pr'
Benchmark 1: hyper_threading_main threads: 16
  Time (mean ± σ):      9.521 s ±  0.163 s    [User: 35.080 s, System: 1.159 s]
  Range (min … max):    9.405 s …  9.636 s    2 runs
 
Benchmark 2: hyper_threading_pr threads: 16
  Time (mean ± σ):      9.651 s ±  0.099 s    [User: 34.908 s, System: 1.160 s]
  Range (min … max):    9.580 s …  9.721 s    2 runs
 
Summary
  hyper_threading_main threads: 16 ran
    1.01 ± 0.02 times faster than hyper_threading_pr threads: 16

gabrielbosio · 2025-12-19T21:41:02Z

Benchmark Results for unmodified programs 🚀

Command	Mean [s]	Min [s]	Max [s]	Relative
`base big_factorial`	1.975 ± 0.029	1.930	2.018	1.00 ± 0.02
`head big_factorial`	1.966 ± 0.012	1.948	1.988	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
`base big_fibonacci`	1.885 ± 0.009	1.867	1.901	1.00
`head big_fibonacci`	1.905 ± 0.026	1.881	1.968	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base blake2s_integration_benchmark`	6.626 ± 0.107	6.430	6.763	1.00
`head blake2s_integration_benchmark`	6.689 ± 0.069	6.607	6.820	1.01 ± 0.02

Command	Mean [s]	Min [s]	Max [s]	Relative
`base compare_arrays_200000`	2.022 ± 0.017	1.990	2.040	1.00
`head compare_arrays_200000`	2.024 ± 0.014	2.009	2.044	1.00 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base dict_integration_benchmark`	1.350 ± 0.006	1.341	1.360	1.01 ± 0.01
`head dict_integration_benchmark`	1.337 ± 0.006	1.332	1.350	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
`base field_arithmetic_get_square_benchmark`	1.128 ± 0.010	1.117	1.144	1.00
`head field_arithmetic_get_square_benchmark`	1.130 ± 0.009	1.119	1.146	1.00 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base integration_builtins`	6.745 ± 0.072	6.677	6.891	1.00
`head integration_builtins`	6.824 ± 0.069	6.724	6.963	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base keccak_integration_benchmark`	6.771 ± 0.078	6.664	6.923	1.00
`head keccak_integration_benchmark`	6.819 ± 0.054	6.732	6.940	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base linear_search`	2.010 ± 0.020	1.990	2.049	1.00
`head linear_search`	2.021 ± 0.016	2.001	2.044	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base math_cmp_and_pow_integration_benchmark`	1.436 ± 0.007	1.426	1.447	1.00 ± 0.01
`head math_cmp_and_pow_integration_benchmark`	1.434 ± 0.008	1.425	1.455	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
`base math_integration_benchmark`	1.393 ± 0.009	1.376	1.405	1.00 ± 0.01
`head math_integration_benchmark`	1.392 ± 0.007	1.381	1.403	1.00

Command	Mean [s]	Min [s]	Max [s]	Relative
`base memory_integration_benchmark`	1.133 ± 0.005	1.124	1.139	1.00
`head memory_integration_benchmark`	1.142 ± 0.009	1.131	1.160	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base operations_with_data_structures_benchmarks`	1.468 ± 0.009	1.454	1.479	1.00 ± 0.01
`head operations_with_data_structures_benchmarks`	1.464 ± 0.008	1.452	1.477	1.00

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`base pedersen`	508.9 ± 2.6	504.9	512.1	1.00 ± 0.01
`head pedersen`	508.0 ± 1.6	506.3	510.8	1.00

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`base poseidon_integration_benchmark`	592.8 ± 2.4	588.3	595.8	1.00
`head poseidon_integration_benchmark`	596.6 ± 1.9	593.0	599.3	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base secp_integration_benchmark`	1.722 ± 0.012	1.706	1.743	1.01 ± 0.01
`head secp_integration_benchmark`	1.706 ± 0.008	1.699	1.726	1.00

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`base set_integration_benchmark`	646.6 ± 2.4	643.4	649.6	1.00
`head set_integration_benchmark`	650.6 ± 3.0	644.7	655.2	1.01 ± 0.01

Command	Mean [s]	Min [s]	Max [s]	Relative
`base uint256_integration_benchmark`	3.869 ± 0.046	3.797	3.955	1.00 ± 0.01
`head uint256_integration_benchmark`	3.854 ± 0.022	3.823	3.907	1.00

chore(vm): unify Uint256 usage across u256 hints and split quotient v…

0b57587

…ia Uint512

JulianGCalderon reviewed Oct 21, 2025

View reviewed changes

Forostovec added 3 commits December 2, 2025 22:33

refactor

5db1cce

Merge branch 'main' into feat/unify-uint256-hints

48fad48

Update CHANGELOG.md

2cbd0b8

JulianGCalderon mentioned this pull request Dec 22, 2025

[DNM] chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237 #2284

Draft

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237

chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237

Uh oh!

Forostovec commented Oct 19, 2025

Uh oh!

JulianGCalderon left a comment

Uh oh!

JulianGCalderon Oct 20, 2025

Uh oh!

JulianGCalderon Oct 20, 2025

Uh oh!

gabrielbosio commented Dec 2, 2025

Uh oh!

Forostovec commented Dec 2, 2025

Uh oh!

gabrielbosio commented Dec 19, 2025

Uh oh!

gabrielbosio commented Dec 19, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237

Are you sure you want to change the base?

chore(vm): unify Uint256 usage across u256 hints and split quotient via Uint512 #2237

Uh oh!

Conversation

Forostovec commented Oct 19, 2025

Uh oh!

JulianGCalderon left a comment

Choose a reason for hiding this comment

Uh oh!

JulianGCalderon Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

JulianGCalderon Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

gabrielbosio commented Dec 2, 2025

Uh oh!

Forostovec commented Dec 2, 2025

Uh oh!

gabrielbosio commented Dec 19, 2025

Uh oh!

gabrielbosio commented Dec 19, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants