Added total_ and claimed_sums as formal variables in ExprEvaluator.

[Alon-Ti]

No description provided.

[Alon-Ti]

• Added add_intermediate and add_secure_intermediate to eval API. #919
• Add safe simplify for expressions that compares random assignments before and after. #918
• Added test for expression simplifier. #908
• Added eval on exprs. #904
• Separated Base and Extension field expressions. #903
• Added From<G> in add_constraint. #901
• Simplify expressions. #894
• Exported shared code to temp variables in ExprEvaluator. #893
• Added total_ and claimed_sums as formal variables in ExprEvaluator. #892 👈 (View in Graphite)
• State machine constraint string. #889
• Expr formatting. #888
• Formal Expr logup variables. #887
• Initialize logups automatically when first used. #879 : 1 other dependent PR (#883 )
• Abstracted eval.write_fracs to eval.add_to_relations. #876
• Added LogupAtRow functionality to EvalAtRow. #875
• dev

This stack of pull requests is managed by Graphite. Learn more about stacking.

[reviewable-StarkWare]

This change is 

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 2.

Benchmark suite	Current: e49c0a4	Previous: cd8b37b	Ratio
iffts/simd ifft/22	12947109 ns/iter (± 265070)	6306399 ns/iter (± 210024)	2.05
merkle throughput/simd merkle	30514622 ns/iter (± 269483)	13712527 ns/iter (± 579195)	2.23

This comment was automatically generated by workflow using github-action-benchmark.

CC: @shaharsamocha7

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.83%. Comparing base (77de4d7) to head (159714d).

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #892      +/-   ##
==========================================
- Coverage   91.95%   91.83%   -0.13%     
==========================================
  Files          93       93              
  Lines       13097    13111      +14     
  Branches    13097    13111      +14     
==========================================
- Hits        12043    12040       -3     
- Misses        939      956      +17     
  Partials      115      115              

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

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[ohad-starkware]

Reviewable status: 0 of 3 files reviewed, 3 unresolved discussions (waiting on @Alon-Ti and @shaharsamocha7)

---

crates/prover/src/constraint_framework/expr.rs line 217 at r3 (raw file):

// (1 << 31) - 1 is an offset no column can reach, it signifies the variable
// offset, which is an input to the verifier.
const CLAIMED_SUM_DUMMY_OFFSET: usize = (1 << 31) - 1;

you can take m31::P

Code quote:

onst CLAIMED_SUM_DUMMY_OFFSET: usize = (1 << 31) - 1;

---

crates/prover/src/constraint_framework/expr.rs line 222 at r3 (raw file):

    pub fn new(interaction: usize, has_partial_sum: bool, log_size: u32) -> Self {
        let total_sum_name = "total_sum".to_string();
        let claimed_sum_name = "claimed_sum".to_string();

shouldnt they be specific per component?

Code quote:

        let total_sum_name = "total_sum".to_string();
        let claimed_sum_name = "claimed_sum".to_string();

---

crates/prover/src/constraint_framework/expr.rs line 229 at r3 (raw file):

            total_sum: Expr::Param(total_sum_name),
            claimed_sum: has_partial_sum
                .then_some((Expr::Param(claimed_sum_name), CLAIMED_SUM_DUMMY_OFFSET)),

if you dont care about the offest why is it here

[Alon-Ti]

Reviewable status: 0 of 3 files reviewed, 3 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

crates/prover/src/constraint_framework/expr.rs line 217 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

you can take m31::P

Done.

---

crates/prover/src/constraint_framework/expr.rs line 222 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

shouldnt they be specific per component?

There's no access to the component name here (nor really a notion of "component"), the way I think it would work is that every component will have a constraint evaluator function that will take "*_sum" as parameters, and the external glue code, that knows about components, will send them accordingly.

It's also possible that we may want to have some sort of "scoping" mechanism, but I'm saving that for when there are named columns because we can work around it here.

---

crates/prover/src/constraint_framework/expr.rs line 229 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

if you dont care about the offest why is it here

To reuse the rest of the logup code, there has to be some offset, this sends the dummy offset that the code on line 47 interprets as a param.

[shaharsamocha7]

Reviewed 1 of 3 files at r1, 1 of 1 files at r4, all commit messages.
Reviewable status: 2 of 3 files reviewed, 3 unresolved discussions (waiting on @ohad-starkware)

[ohad-starkware]

Reviewed 1 of 3 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @Alon-Ti)