Add secure powers generation for simd

[Gali-StarkWare]

No description provided.

[reviewable-StarkWare]

This change is 

[Gali-StarkWare]

• Add secure powers generation for simd #930 : 2 dependent PRs (#933 , #934 ) 👈 (View in Graphite)
• dev

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

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 96.55172% with 1 line in your changes missing coverage. Please review.

Project coverage is 91.49%. Comparing base (8550b7d) to head (3a203f6).
Report is 3 commits behind head on dev.

Files with missing lines	Patch %	Lines
crates/prover/src/core/backend/simd/utils.rs	96.55%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #930      +/-   ##
==========================================
+ Coverage   91.48%   91.49%   +0.01%     
==========================================
  Files          98       98              
  Lines       13733    13762      +29     
  Branches    13733    13762      +29     
==========================================
+ Hits        12563    12591      +28     
- Misses       1055     1056       +1     
  Partials      115      115              

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

[ohad-starkware]

Perhaps I didn't explain properly
the function should take alpha, n_powers and return Vec, that is the n_powers powers of alpha (exclusive) : 0, a, a^2....a^(n-1)
use larger vectors to perform multiple operations at a time

Reviewed all commit messages.
Reviewable status: 0 of 1 files reviewed, all discussions resolved (waiting on @shaharsamocha7)

[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: 0a2e9b9	Previous: cd8b37b	Ratio
iffts/simd ifft/22	13635674 ns/iter (± 219842)	6306399 ns/iter (± 210024)	2.16
iffts/simd ifft/25	133703364 ns/iter (± 845083)	65891527 ns/iter (± 1735211)	2.03
iffts/simd ifft/26	278573898 ns/iter (± 3255932)	136655216 ns/iter (± 2242165)	2.04
iffts/simd ifft/27	646251801 ns/iter (± 15536196)	312588285 ns/iter (± 9816879)	2.07
iffts/simd ifft/28	1351168377 ns/iter (± 36149701)	643771030 ns/iter (± 19147376)	2.10
simd rfft 20bit	3234196 ns/iter (± 21252)	1617070 ns/iter (± 66435)	2.00
merkle throughput/simd merkle	30000903 ns/iter (± 498754)	13712527 ns/iter (± 579195)	2.19

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

CC: @shaharsamocha7

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 14 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 62 at r2 (raw file):

// TODO(Gali): Remove #[allow(dead_code)].
#[allow(dead_code)]
pub fn generate_secure_powers_for_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {

Suggestion:

pub fn generate_secure_powers_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {

---

crates/prover/src/core/backend/simd/utils.rs line 63 at r2 (raw file):

#[allow(dead_code)]
pub fn generate_secure_powers_for_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {
    let step_arr: [SecureField; N_LANES] = (0..N_LANES)

that logic is already written in the CPU function, use it?

---

crates/prover/src/core/backend/simd/utils.rs line 69 at r2 (raw file):

            Some(res)
        })
        .collect::<Vec<SecureField>>()

less noise

Suggestion:

.collect_vec()

---

crates/prover/src/core/backend/simd/utils.rs line 71 at r2 (raw file):

        .collect::<Vec<SecureField>>()
        .try_into()
        .expect("Failed generating secure powers.");

I think it's alright to .unwrap() here, you won't fail the try_into...
non - blocking

---

crates/prover/src/core/backend/simd/utils.rs line 77 at r2 (raw file):

    let step_felt = step_arr[N_LANES - 1] * felt;

    let mut packed_powers_vec = Vec::new();

this is a good place to use vec::with_capacity

---

crates/prover/src/core/backend/simd/utils.rs line 80 at r2 (raw file):

    let mut curr_power: usize = 0;

    while curr_power < n_powers {

consider using scan land collect ike you did above, easier to read

---

crates/prover/src/core/backend/simd/utils.rs line 81 at r2 (raw file):

    while curr_power < n_powers {
        let base_packed_felt = PackedSecureField::from_array([base_felt; N_LANES]);

I think you can avoid ::from_array inside the loop, this is a load and is somewhat expensive
perhaps define the step as Simd([a^16; 16]), and multiply the current value with it each time (your step is a small register)

---

crates/prover/src/core/backend/simd/utils.rs line 81 at r2 (raw file):

    while curr_power < n_powers {
        let base_packed_felt = PackedSecureField::from_array([base_felt; N_LANES]);

also you can use broadcast to get [value; N_LANE] cheaper

---

crates/prover/src/core/backend/simd/utils.rs line 87 at r2 (raw file):

    }

    let powers_vec: Vec<SecureField> = packed_powers_vec

Suggestion:

    let secure_powers = packed_powers_vec

---

crates/prover/src/core/backend/simd/utils.rs line 88 at r2 (raw file):

    let powers_vec: Vec<SecureField> = packed_powers_vec
        .iter()

Suggestion:

.into_iter()

---

crates/prover/src/core/backend/simd/utils.rs line 90 at r2 (raw file):

        .iter()
        .flat_map(|x| x.to_array().to_vec())
        .collect();

Suggestion:

collect_vec();

---

crates/prover/src/core/backend/simd/utils.rs line 92 at r2 (raw file):

        .collect();

    powers_vec[0..n_powers].to_vec()

you're doing 2 copies here, I think you can reduce that to 0 (trim the front/back...?)

Code quote:

    let powers_vec: Vec<SecureField> = packed_powers_vec
        .iter()
        .flat_map(|x| x.to_array().to_vec())
        .collect();

    powers_vec[0..n_powers].to_vec()

---

crates/prover/src/core/backend/simd/utils.rs line 129 at r2 (raw file):

    fn generate_secure_powers_works() {
        let felt = qm31!(1, 2, 3, 4);
        let n_powers = 10;

if you had a bug (you don't) with n > 16 this test would not catch it, use a bigger number.

---

crates/prover/src/core/backend/simd/utils.rs line 134 at r2 (raw file):

        assert_eq!(powers.len(), n_powers);
        assert_eq!(powers[0], SecureField::one());

Suggestion:

        let cpu_powers = generate_secure_powers(felt, n_powers)
        
        let powers = generate_secure_powers_for_simd(felt, n_powers);

        assert_eq!(powers, cpu_powers);

---

crates/prover/src/core/backend/simd/utils.rs line 140 at r2 (raw file):

    #[test]
    fn generate_empty_secure_powers_works() {

I don't think this edge case is of concern, remove.

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 15 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 127 at r2 (raw file):

    #[test]
    fn generate_secure_powers_works() {

rename the tests

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 15 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 63 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

that logic is already written in the CPU function, use it?

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 69 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

less noise

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 77 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

this is a good place to use vec::with_capacity

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 80 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

consider using scan land collect ike you did above, easier to read

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 81 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

I think you can avoid ::from_array inside the loop, this is a load and is somewhat expensive
perhaps define the step as Simd([a^16; 16]), and multiply the current value with it each time (your step is a small register)

Done. Swapped base-step variables names as well

---

crates/prover/src/core/backend/simd/utils.rs line 81 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

also you can use broadcast to get [value; N_LANE] cheaper

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 92 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

you're doing 2 copies here, I think you can reduce that to 0 (trim the front/back...?)

Done. Not sure that I got it

---

crates/prover/src/core/backend/simd/utils.rs line 127 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

rename the tests

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 129 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

if you had a bug (you don't) with n > 16 this test would not catch it, use a bigger number.

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 140 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

I don't think this edge case is of concern, remove.

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 62 at r2 (raw file):

// TODO(Gali): Remove #[allow(dead_code)].
#[allow(dead_code)]
pub fn generate_secure_powers_for_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 87 at r2 (raw file):

    }

    let powers_vec: Vec<SecureField> = packed_powers_vec

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 88 at r2 (raw file):

    let powers_vec: Vec<SecureField> = packed_powers_vec
        .iter()

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 90 at r2 (raw file):

        .iter()
        .flat_map(|x| x.to_array().to_vec())
        .collect();

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 134 at r2 (raw file):

        assert_eq!(powers.len(), n_powers);
        assert_eq!(powers[0], SecureField::one());

Done.

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 6 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 92 at r2 (raw file):

Previously, Gali-StarkWare wrote…

Done. Not sure that I got it

the collect_vec().into_iter() is redundant, you already have an iterator (scan), that collect is a copy for nothing

---

crates/prover/src/core/backend/simd/utils.rs line 64 at r3 (raw file):

#[allow(dead_code)]
pub fn generate_secure_powers_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {
    let base_arr: [SecureField; N_LANES] =

Suggestion:

 let base_arr =

---

crates/prover/src/core/backend/simd/utils.rs line 69 at r3 (raw file):

    let step_felt = base_arr[N_LANES - 1] * felt;
    let step_packed_felt = PackedSecureField::broadcast(step_felt);

style
non-blocking

Suggestion:

    let step = base_arr[N_LANES - 1] * felt;
    let step = PackedSecureField::broadcast(step_felt);

---

crates/prover/src/core/backend/simd/utils.rs line 75 at r3 (raw file):

    } else {
        (n_powers / N_LANES) + 1
    };

Suggestion:

    let packed_size = (n_powers + N_LANES - 1) / N_LANES;

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

        .collect_vec()
        .into_iter()
        .flat_map(|x| x.to_array().to_vec())

x.to_array() is a copy, use transmute?

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

        .collect_vec()
        .into_iter()
        .flat_map(|x| x.to_array().to_vec())

this invoked a heap allocation for every vector :(

Suggestion:

flat_map(|x| x.to_array())

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

    #[test]
    fn generate_secure_powers_simd_returns_the_same_as_cpu() {

Suggestion:

fn test_generate_secure_powers_simd() {

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 7 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

    fn generate_secure_powers_simd_returns_the_same_as_cpu() {
        let felt = qm31!(1, 2, 3, 4);
        let n_powers = 30;

bigger?

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 7 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 92 at r2 (raw file):

Previously, ohad-starkware (Ohad) wrote…

the collect_vec().into_iter() is redundant, you already have an iterator (scan), that collect is a copy for nothing

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

this invoked a heap allocation for every vector :(

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

x.to_array() is a copy, use transmute?

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

bigger?

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 64 at r3 (raw file):

#[allow(dead_code)]
pub fn generate_secure_powers_simd(felt: SecureField, n_powers: usize) -> Vec<SecureField> {
    let base_arr: [SecureField; N_LANES] =

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 75 at r3 (raw file):

    } else {
        (n_powers / N_LANES) + 1
    };

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

    #[test]
    fn generate_secure_powers_simd_returns_the_same_as_cpu() {

Done.

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 6 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

Previously, Gali-StarkWare wrote…

Done.

sorry the transmute is a lie I forgot were in the extension field

[shaharsamocha7]

Reviewable status: 0 of 1 files reviewed, 7 unresolved discussions (waiting on @Gali-StarkWare and @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 69 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

style
non-blocking

alternatively
let felt_pows = PackedSecureField::from_array(generate_secure_powers(felt, N_LANES).try_into().unwrap());
let felt_pow_n_lanes = PackedSecureField::broadcast(base_arr[N_LANES - 1] * felt);

---

crates/prover/src/core/backend/simd/utils.rs line 75 at r3 (raw file):

Previously, Gali-StarkWare wrote…

Done.

there is a function
n_powers.div_ceil(N_LANES)

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

Previously, Gali-StarkWare wrote…

Done.

To be more explicit:
Test name should describe what the test is checking rather than the implementation itself.

---

crates/prover/src/core/backend/simd/utils.rs line 125 at r3 (raw file):

        let cpu_powers = generate_secure_powers(felt, n_powers);
        let powers = generate_secure_powers_simd(felt, n_powers);
        assert_eq!(powers, cpu_powers);

Add a new line here
https://automationpanda.com/2020/07/07/arrange-act-assert-a-pattern-for-writing-good-tests/

Suggestion:

        let powers = generate_secure_powers_simd(felt, n_powers);
        
        assert_eq!(powers, cpu_powers);

[Gali-StarkWare]

Also changed some variables' names, LMK if it's better

Reviewable status: 0 of 1 files reviewed, 7 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 75 at r3 (raw file):

Previously, shaharsamocha7 wrote…

there is a function
n_powers.div_ceil(N_LANES)

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 85 at r3 (raw file):

Previously, ohad-starkware (Ohad) wrote…

sorry the transmute is a lie I forgot were in the extension field

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

Previously, shaharsamocha7 wrote…

To be more explicit:
Test name should describe what the test is checking rather than the implementation itself.

I thought that specifying that we want the same functionality as the of the other function is good for future changes (if we'll ever change one of them to do something different), but I guess there's no point doing that..

---

crates/prover/src/core/backend/simd/utils.rs line 125 at r3 (raw file):

Previously, shaharsamocha7 wrote…

Add a new line here
https://automationpanda.com/2020/07/07/arrange-act-assert-a-pattern-for-writing-good-tests/

Done.

[shaharsamocha7]

Reviewable status: 0 of 1 files reviewed, 7 unresolved discussions (waiting on @Gali-StarkWare and @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

Previously, Gali-StarkWare wrote…

I thought that specifying that we want the same functionality as the of the other function is good for future changes (if we'll ever change one of them to do something different), but I guess there's no point doing that..

The function purpose is to generate_secure_powers (with simd).
It will still exists if we deleted the cpu implementation.

The way you chose to test it is by checking equivalency to the cpu function.
Problem with adding it to the test name is that implementations tend to change (without changing the test name) while the purpose is more solid.

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, Gali-StarkWare wrote…

Done.

Actually why n_powers is so large?
we don't want this test to take too much time.

If so, maybe there are edge cases where n_powers is small.

---

crates/prover/src/core/backend/simd/utils.rs line 116 at r5 (raw file):

        let powers = generate_secure_powers_simd(felt, n_powers);

        assert_eq!(powers, cpu_powers);

consider to rename

Suggestion:

        let simd_powers = generate_secure_powers_simd(felt, n_powers);

        assert_eq!(simd_powers, cpu_powers);

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 3 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, shaharsamocha7 wrote…

Actually why n_powers is so large?
we don't want this test to take too much time.

If so, maybe there are edge cases where n_powers is small.

I thougth something like a 100, larger than two SIMD vectors ant not a multiple of 16
there is an edge case for <16 and 0, My thoughts were that you are already using the CPU function in your implementation, so comparing against that can be redundant when the function is simple.
I'm not sure so wait for Shahar's input on this.

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 2 unresolved discussions (waiting on @Gali-StarkWare and @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 119 at r3 (raw file):

Previously, shaharsamocha7 wrote…

The function purpose is to generate_secure_powers (with simd).
It will still exists if we deleted the cpu implementation.

The way you chose to test it is by checking equivalency to the cpu function.
Problem with adding it to the test name is that implementations tend to change (without changing the test name) while the purpose is more solid.

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, shaharsamocha7 wrote…

100 is fine IMO.
I would also check <16 but it's nice to have.
non-blocking

Should I add the <16 check in the same test or in a new one? Based on the test's naming and the discussion above it should be in the same test, just making sure

[shaharsamocha7]

Reviewable status: 0 of 1 files reviewed, 2 unresolved discussions (waiting on @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, Gali-StarkWare wrote…

Should I add the <16 check in the same test or in a new one? Based on the test's naming and the discussion above it should be in the same test, just making sure

The way I see it you can either adapt the test such that
n_powers_vec = [16, 100]

or split it to two tests
test_generate_secure_powers_simd_small
test_generate_secure_powers_simd_large

In this specific case I would go with option 1, but your call

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 2 unresolved discussions (waiting on @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r3 (raw file):

Previously, shaharsamocha7 wrote…

The way I see it you can either adapt the test such that
n_powers_vec = [16, 100]

or split it to two tests
test_generate_secure_powers_simd_small
test_generate_secure_powers_simd_large

In this specific case I would go with option 1, but your call

Done. Also added 0

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 3 unresolved discussions (waiting on @Gali-StarkWare)

---

-- commits line 2 at r6:
please squash the commits

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @Gali-StarkWare)

---

crates/prover/src/core/backend/simd/utils.rs line 114 at r7 (raw file):

        let n_powers_vec = [0, 16, 100];

        let cpu_powers_vec: Vec<Vec<SecureField>> = n_powers_vec

you're doing 3 separate tests, have 3 corresponding asserts
so if one fails you know which of the edge cases are failing

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @Gali-StarkWare and @ohad-starkware)

---

-- commits line 2 at r6:

Previously, ohad-starkware (Ohad) wrote…

please squash the commits

Done.

---

crates/prover/src/core/backend/simd/utils.rs line 114 at r7 (raw file):

Previously, ohad-starkware (Ohad) wrote…

you're doing 3 separate tests, have 3 corresponding asserts
so if one fails you know which of the edge cases are failing

Done.

[shaharsamocha7]

Reviewable status: 0 of 1 files reviewed, 5 unresolved discussions (waiting on @Gali-StarkWare and @ohad-starkware)

---

crates/prover/src/core/backend/simd/utils.rs line 126 at r8 (raw file):

        assert_eq!(simd_powers_vec[0], cpu_powers_vec[0]);
        assert_eq!(simd_powers_vec[1], cpu_powers_vec[1]);
        assert_eq!(simd_powers_vec[2], cpu_powers_vec[2]);

That code is not robust in the sense that if you will want to add another test case you'll also have to make the change here.
We don't want to limit our code for no reason

I put a suggestion for refactor, LMK what you think

Suggestion:

        n_powers_vec.iter().for_each(|&n_powers| {
            let expected = generate_secure_powers(felt, n_powers);
            let actual = generate_secure_powers_simd(felt, n_powers);

            assert_eq!(expected, actual);
        });

[ohad-starkware]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @Gali-StarkWare and @shaharsamocha7)

---

-- commits line 2 at r6:

Previously, Gali-StarkWare wrote…

Done.

are you sure?
I expected a single line commit message

---

crates/prover/src/core/backend/simd/utils.rs line 126 at r8 (raw file):

Previously, shaharsamocha7 wrote…

That code is not robust in the sense that if you will want to add another test case you'll also have to make the change here.
We don't want to limit our code for no reason

I put a suggestion for refactor, LMK what you think

i'd also add n_powers to the panic message

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

crates/prover/src/core/backend/simd/utils.rs line 126 at r8 (raw file):

Previously, ohad-starkware (Ohad) wrote…

i'd also add n_powers to the panic message

I think so too. Added these asserts because Ohad requested to and I do agree with him that it will be easier to debug his way. Should have thought of a better way to do so. This way is nicer - it's both generic in terms of the number of test cases and will fail the specific test case.

[Gali-StarkWare]

Reviewable status: 0 of 1 files reviewed, 4 unresolved discussions (waiting on @ohad-starkware and @shaharsamocha7)

---

-- commits line 2 at r6:

Previously, ohad-starkware (Ohad) wrote…

are you sure?
I expected a single line commit message

Done.

[shaharsamocha7]

Reviewed 2 of 2 files at r13, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @ohad-starkware)

[ohad-starkware]

Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @Gali-StarkWare)

---

crates/prover/src/core/backend/simd/utils.rs line 121 at r13 (raw file):

            assert_eq!(
                expected, actual,
                "Error generating secure powers in n_powers = {}.",

[Gali-StarkWare]

Merge activity

• Dec 18, 5:58 AM EST: A user started a stack merge that includes this pull request via Graphite.
• Dec 18, 5:59 AM EST: A user merged this pull request with Graphite.