Use SecureField in interaction.

[alonh5]

This change is 

[alonh5]

• Extract constraint code into functions. #644 : 2 dependent PRs (#646 , #654 )
• Use SecureField in interaction. #642 👈
• Modify mask_points for interaction. #639
• Modify trace_log_degree_bounds for interaction. #638
• Add wide fib lookup constraint. #637
• Add interaction to proof. #636
• Separate componenet and air trace writers. #651
• dev

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

Join @alonh5 and the rest of your teammates on Graphite

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 89.50%. Comparing base (1f38b04) to head (34a1233).

Additional details and impacted files

@@           Coverage Diff           @@
##              dev     #642   +/-   ##
=======================================
  Coverage   89.50%   89.50%           
=======================================
  Files          75       75           
  Lines        9640     9640           
  Branches     9640     9640           
=======================================
  Hits         8628     8628           
  Misses        932      932           
  Partials       80       80           

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

[spapinistarkware]

Reviewed 1 of 11 files at r1, 1 of 11 files at r2, all commit messages.
Reviewable status: 2 of 12 files reviewed, 3 unresolved discussions (waiting on @alonh5 and @shaharsamocha7)

---

crates/prover/src/core/utils.rs line 94 at r3 (raw file):

/// Securely combines the given values using the given random alpha and z.
/// Alpha and z should be secure field elements for soundness.
pub fn shifted_secure_combination<F: ExtensionOf<BaseField>>(

Is this needed?

---

crates/prover/src/core/air/mod.rs line 82 at r3 (raw file):

        trace: &ColumnVec<&CircleEvaluation<B, BaseField, BitReversedOrder>>,
        elements: &InteractionElements,
    ) -> ColumnVec<SecureEvaluation<B>>;

I don't think it is always a bunch of secure columns in general.
It should be exactly like trace, just a bunch of basefield columns.

Code quote:

ColumnVec<SecureEvaluation<B>>

---

crates/prover/src/core/prover/mod.rs line 74 at r3 (raw file):

    let interaction_elements = air.interaction_elements(channel);
    let interaction_trace_polys = air.interact(&trace, &interaction_elements);

1. This is not using the precomputed twiddles.
2. Do we need to change the interface? Can't interact() reutrn teh evaluation?

[alonh5]

Reviewable status: 2 of 12 files reviewed, 3 unresolved discussions (waiting on @shaharsamocha7 and @spapinistarkware)

---

crates/prover/src/core/utils.rs line 94 at r3 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

Is this needed?

Yes the values could be M31 (in the prover constraint evaluation) and QM31 (in the verifier oods constraint evaluation).

---

crates/prover/src/core/air/mod.rs line 82 at r3 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

I don't think it is always a bunch of secure columns in general.
It should be exactly like trace, just a bunch of basefield columns.

Done.

---

crates/prover/src/core/prover/mod.rs line 74 at r3 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

1. This is not using the precomputed twiddles.
2. Do we need to change the interface? Can't interact() reutrn teh evaluation?

Done.

[spapinistarkware]

Reviewed 1 of 11 files at r2, 9 of 9 files at r4, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @alonh5 and @shaharsamocha7)

---

crates/prover/src/examples/wide_fibonacci/component.rs line 135 at r4 (raw file):

        let trace_values = trace.iter().map(|eval| &eval.values[..]).collect_vec();
        let (alpha, z) = (elements[ALPHA_ID], elements[Z_ID]);
        let values = write_lookup_column(&trace_values, alpha, z);

For efficientcy, we probably want write_lookup_column() to already return a SecureColumn.
You can add a TODO for that.

---

crates/prover/src/examples/wide_fibonacci/component.rs line 139 at r4 (raw file):

        for (i, value) in values.into_iter().enumerate() {
            secure_column.set(i, value);
        }

I think we have a FromIterator impl for SecureColumn. so, you can use collect()

Code quote:

        for (i, value) in values.into_iter().enumerate() {
            secure_column.set(i, value);
        }

---

crates/prover/src/examples/wide_fibonacci/constraint_eval.rs line 95 at r4 (raw file):

            // Lookup constraints.
            let lookup_value =
                SecureCirclePoly::<CpuBackend>::eval_from_partial_evals(std::array::from_fn(|j| {

Both of us know what this function does, but I fear other people will be confused by it. Maybe we can consult with the team how to name this.

Code quote:

eval_from_partial_evals

[alonh5]

Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @shaharsamocha7 and @spapinistarkware)

---

crates/prover/src/examples/wide_fibonacci/component.rs line 135 at r4 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

For efficientcy, we probably want write_lookup_column() to already return a SecureColumn.
You can add a TODO for that.

Done.

---

crates/prover/src/examples/wide_fibonacci/component.rs line 139 at r4 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

I think we have a FromIterator impl for SecureColumn. so, you can use collect()

Done, thanks.

---

crates/prover/src/examples/wide_fibonacci/constraint_eval.rs line 95 at r4 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

Both of us know what this function does, but I fear other people will be confused by it. Maybe we can consult with the team how to name this.

Yes, I also think we should move it to be a method of secure field and then it will be more clear when it's next to from_m31_array(). I added a TODO also.

[alonh5]

Reviewable status: 11 of 13 files reviewed, 3 unresolved discussions (waiting on @shaharsamocha7 and @spapinistarkware)

---

crates/prover/src/examples/wide_fibonacci/constraint_eval.rs line 95 at r4 (raw file):

Previously, alonh5 (Alon Haramati) wrote…

Yes, I also think we should move it to be a method of secure field and then it will be more clear when it's next to from_m31_array(). I added a TODO also.

PR 679

[spapinistarkware]

Reviewed 2 of 2 files at r5, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @shaharsamocha7)

[alonh5]

Merge activity

• Jun 27, 8:38 AM EDT: @alonh5 started a stack merge that includes this pull request via Graphite.
• Jun 27, 8:38 AM EDT: @alonh5 merged this pull request with Graphite.