Create MLE eval prover component

[andrewmilson]

This change is 

[andrewmilson]

• Create blake example that uses GKR for lookups #807
• Create MLE eval verifier component #805
• Create MLE eval prover component #804 👈
• Add build trace functions for MLE eval component #803
• Remove InteractionElements and LookupValues #802
• Pass entire mask to components #801
• Add MleCollection for accumulating MLEs #800
• Combine eq step constraints into single constraint #799
• Move all MLE eval related constraints into same module #798
• dev

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

Join @andrewmilson and the rest of your teammates on Graphite

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 95.00000% with 1 line in your changes missing coverage. Please review.

Project coverage is 92.05%. Comparing base (cdf66f3) to head (5b3001a).
Report is 1 commits behind head on dev.

Files with missing lines	Patch %	Lines
crates/prover/src/core/air/accumulation.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #804      +/-   ##
==========================================
+ Coverage   91.86%   92.05%   +0.19%     
==========================================
  Files          89       89              
  Lines       12080    12098      +18     
  Branches    12080    12098      +18     
==========================================
+ Hits        11097    11137      +40     
+ Misses        876      854      -22     
  Partials      107      107              

Flag	Coverage Δ
	92.05% <95.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[spapinistarkware]

Reviewed 1 of 3 files at r2.
Reviewable status: 1 of 5 files reviewed, 1 unresolved discussion (waiting on @andrewmilson and @shaharsamocha7)

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 211 at r2 (raw file):

                .values,
        );
        let is_second_lde = VeryPackedBaseColumn::from(

Do we really need is_second? Can you not just take is_first at a mask offset, 1?

[andrewmilson]

Reviewable status: 0 of 5 files reviewed, 1 unresolved discussion (waiting on @shaharsamocha7 and @spapinistarkware)

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 211 at r2 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

Do we really need is_second? Can you not just take is_first at a mask offset, 1?

Updated to use the offset but honestly might be slower due to the bad access patterns.

[andrewmilson]

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 211 at r2 (raw file):

Previously, andrewmilson (Andrew Milson) wrote…

Updated to use the offset but honestly might be slower due to the bad access patterns.

Actually I'm wrong never mind

[shaharsamocha7]

Reviewed 4 of 4 files at r5, all commit messages.
Reviewable status: 4 of 5 files reviewed, 5 unresolved discussions (waiting on @andrewmilson and @spapinistarkware)

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 68 at r5 (raw file):

impl<'twiddles, 'oracle, O: MleCoeffColumnOracle> MleEvalProverComponent<'twiddles, 'oracle, O> {
    // TODO(andrew): Some eval points may affect completeness. Document.

Can you document that?

Code quote:

    // TODO(andrew): Some eval points may affect completeness. Document.

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 89 at r5 (raw file):

        let trace_structure = mle_eval_info(interaction, n_variables).mask_offsets;
        let trace_locations = location_allocator.next_for_structure(&trace_structure);

Does TreeSubspan also useful for the constant columns?

Code quote:

 let trace_locations = location_allocator.next_for_structure(&trace_structure);

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 172 at r5 (raw file):

            carry_quotients_col_eval,
            is_first,
            is_second,

Should we get those as arguments?
Can we just get the constant column and then extract those values inside?

Code quote:

            is_first,
            is_second,

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 1009 at r5 (raw file):

            let log_size = mle.n_variables() as u32;
            let trace_domain = CanonicCoset::new(log_size).circle_domain();
            let mle_coeffs_col_by_coords = mle.clone().into_evals().into_secure_column_by_coords();

Is this copy the entire column? or what?

Code quote:

.clone().

[andrewmilson]

Reviewable status: 3 of 5 files reviewed, 5 unresolved discussions (waiting on @shaharsamocha7 and @spapinistarkware)

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 68 at r5 (raw file):

Previously, shaharsamocha7 wrote…

Can you document that?

Done.

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 89 at r5 (raw file):

Previously, shaharsamocha7 wrote…

Does TreeSubspan also useful for the constant columns?

It can as constant columns are now.
It treats the constant interaction tree just like any interaction tree.

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 172 at r5 (raw file):

Previously, shaharsamocha7 wrote…

Should we get those as arguments?
Can we just get the constant column and then extract those values inside?

Added a TODO

---

crates/prover/src/examples/xor/gkr_lookups/mle_eval.rs line 1009 at r5 (raw file):

Previously, shaharsamocha7 wrote…

Is this copy the entire column? or what?

Yeah, it's just a test function

[shaharsamocha7]

Reviewed 1 of 2 files at r3, 1 of 1 files at r7, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @spapinistarkware)