Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test/metal/mont #30

Merged
merged 2 commits into from
Jan 4, 2025
Merged

Test/metal/mont #30

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a29e9c6
refactor(mont): test and benchmarks to use random BigInt values
moven0831 Jan 3, 2025
5350f18
tests(mont): replace ScalarField with BaseField in mont_mul benchmark…
moven0831 Jan 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions mopro-msm/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,7 @@ objc = { version = "=0.2.7" }
proptest = { version = "1.4.0" }
rayon = "1.5.1"
itertools = "0.13.0"
rand = "0.8.5"

[build-dependencies]
color-eyre = "0.6"
Expand Down
33 changes: 11 additions & 22 deletions mopro-msm/src/msm/metal_msm/tests/mont_backend/mont_benchmarks.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,10 +4,11 @@ use crate::msm::metal_msm::utils::limbs_conversion::{FromLimbs, ToLimbs};
use crate::msm::metal_msm::utils::mont_params::{
calc_bitwidth, calc_mont_radix, calc_nsafe, calc_num_limbs, calc_rinv_and_n0,
};
use ark_bn254::Fr as ScalarField;
use ark_bn254::Fq as BaseField;
use ark_ff::{BigInt, PrimeField};
use metal::*;
use num_bigint::BigUint;
use num_bigint::{BigUint, RandBigInt};
use rand::thread_rng;
use stopwatch::Stopwatch;

#[test]
Expand Down Expand Up @@ -53,26 +54,14 @@ fn expensive_computation(
}

pub fn benchmark(log_limb_size: u32, shader_file: &str) -> Result<i64, String> {
let p = BigUint::parse_bytes(
b"30644E72E131A029B85045B68181585D2833E84879B9709143E1F593F0000001",
16,
)
.unwrap();
assert!(p == ScalarField::MODULUS.try_into().unwrap());
let p: BigUint = BaseField::MODULUS.try_into().unwrap();

let p_bitwidth = calc_bitwidth(&p);
let num_limbs = calc_num_limbs(log_limb_size, p_bitwidth);

let a = BigUint::parse_bytes(
b"10ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let b = BigUint::parse_bytes(
b"11ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let mut rng = thread_rng();
let a = rng.gen_biguint_below(&p);
let b = rng.gen_biguint_below(&p);

let nsafe = calc_nsafe(log_limb_size);
if nsafe == 0 {
Expand All @@ -88,20 +77,20 @@ pub fn benchmark(log_limb_size: u32, shader_file: &str) -> Result<i64, String> {

let cost = 2u32.pow(16u32) as usize;
let expected = expensive_computation(cost, &a, &b, &p, &r);
let expected_limbs = ScalarField::from_bigint(expected.clone().try_into().unwrap())
let expected_limbs = BaseField::from_bigint(expected.clone().try_into().unwrap())
.unwrap()
.into_bigint()
.to_limbs(num_limbs, log_limb_size);

let ar_limbs = ScalarField::from_bigint(a_r.clone().try_into().unwrap())
let ar_limbs = BaseField::from_bigint(a_r.clone().try_into().unwrap())
.unwrap()
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
let br_limbs = ScalarField::from_bigint(b_r.clone().try_into().unwrap())
let br_limbs = BaseField::from_bigint(b_r.clone().try_into().unwrap())
.unwrap()
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
let p_limbs = &ScalarField::MODULUS.to_limbs(num_limbs, log_limb_size);
let p_limbs = &BaseField::MODULUS.to_limbs(num_limbs, log_limb_size);

let device = get_default_device();

Expand Down
30 changes: 12 additions & 18 deletions mopro-msm/src/msm/metal_msm/tests/mont_backend/mont_mul_cios.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,11 @@ use crate::msm::metal_msm::host::gpu::{
use crate::msm::metal_msm::host::shader::{compile_metal, write_constants};
use crate::msm::metal_msm::utils::limbs_conversion::{FromLimbs, ToLimbs};
use crate::msm::metal_msm::utils::mont_params::{calc_mont_radix, calc_nsafe, calc_rinv_and_n0};
use ark_bn254::Fr as ScalarField;
use ark_bn254::Fq as BaseField;
use ark_ff::{BigInt, PrimeField};
use metal::*;
use num_bigint::BigUint;
use num_bigint::{BigUint, RandBigInt};
use rand::thread_rng;

#[test]
#[serial_test::serial]
Expand All @@ -24,34 +25,27 @@ pub fn test_mont_mul_15() {
}

pub fn do_test(log_limb_size: u32) {
let modulus_bits = ScalarField::MODULUS_BIT_SIZE as u32;
let modulus_bits = BaseField::MODULUS_BIT_SIZE as u32;
let num_limbs = ((modulus_bits + log_limb_size - 1) / log_limb_size) as usize;

let r = calc_mont_radix(num_limbs, log_limb_size);
let p: BigUint = ScalarField::MODULUS.try_into().unwrap();
let p: BigUint = BaseField::MODULUS.try_into().unwrap();
let nsafe = calc_nsafe(log_limb_size);

let res = calc_rinv_and_n0(&p, &r, log_limb_size);
let n0 = res.1;

let a = BigUint::parse_bytes(
b"10ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let b = BigUint::parse_bytes(
b"11ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let mut rng = thread_rng();
let a = rng.gen_biguint_below(&p);
let b = rng.gen_biguint_below(&p);

let a_r = &a * &r % &p;
let b_r = &b * &r % &p;
let expected = (&a * &b * &r) % &p;

let a_r_in_ark = ScalarField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = ScalarField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = ScalarField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let a_r_in_ark = BaseField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = BaseField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = BaseField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let expected_limbs = expected_in_ark
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
Expand All @@ -67,7 +61,7 @@ pub fn do_test(log_limb_size: u32) {
);
let p_buf = create_buffer(
&device,
&ScalarField::MODULUS.to_limbs(num_limbs, log_limb_size),
&BaseField::MODULUS.to_limbs(num_limbs, log_limb_size),
);
let result_buf = create_empty_buffer(&device, num_limbs);

Expand Down
30 changes: 12 additions & 18 deletions mopro-msm/src/msm/metal_msm/tests/mont_backend/mont_mul_modified.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,11 @@ use crate::msm::metal_msm::host::gpu::{
use crate::msm::metal_msm::host::shader::{compile_metal, write_constants};
use crate::msm::metal_msm::utils::limbs_conversion::{FromLimbs, ToLimbs};
use crate::msm::metal_msm::utils::mont_params::{calc_mont_radix, calc_nsafe, calc_rinv_and_n0};
use ark_bn254::Fr as ScalarField;
use ark_bn254::Fq as BaseField;
use ark_ff::{BigInt, PrimeField};
use metal::*;
use num_bigint::BigUint;
use num_bigint::{BigUint, RandBigInt};
use rand::thread_rng;

#[test]
#[serial_test::serial]
Expand All @@ -24,34 +25,27 @@ pub fn test_mont_mul_15() {
}

pub fn do_test(log_limb_size: u32) {
let modulus_bits = ScalarField::MODULUS_BIT_SIZE as u32;
let modulus_bits = BaseField::MODULUS_BIT_SIZE as u32;
let num_limbs = ((modulus_bits + log_limb_size - 1) / log_limb_size) as usize;

let r = calc_mont_radix(num_limbs, log_limb_size);
let p: BigUint = ScalarField::MODULUS.try_into().unwrap();
let p: BigUint = BaseField::MODULUS.try_into().unwrap();
let nsafe = calc_nsafe(log_limb_size);

let res = calc_rinv_and_n0(&p, &r, log_limb_size);
let n0 = res.1;

let a = BigUint::parse_bytes(
b"10ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let b = BigUint::parse_bytes(
b"11ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let mut rng = thread_rng();
let a = rng.gen_biguint_below(&p);
let b = rng.gen_biguint_below(&p);

let a_r = &a * &r % &p;
let b_r = &b * &r % &p;
let expected = (&a * &b * &r) % &p;

let a_r_in_ark = ScalarField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = ScalarField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = ScalarField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let a_r_in_ark = BaseField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = BaseField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = BaseField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let expected_limbs = expected_in_ark
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
Expand All @@ -67,7 +61,7 @@ pub fn do_test(log_limb_size: u32) {
);
let p_buf = create_buffer(
&device,
&ScalarField::MODULUS.to_limbs(num_limbs, log_limb_size),
&BaseField::MODULUS.to_limbs(num_limbs, log_limb_size),
);
let result_buf = create_empty_buffer(&device, num_limbs);

Expand Down
40 changes: 17 additions & 23 deletions mopro-msm/src/msm/metal_msm/tests/mont_backend/mont_mul_optimised.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,10 +9,11 @@ use crate::msm::metal_msm::host::gpu::{
use crate::msm::metal_msm::host::shader::{compile_metal, write_constants};
use crate::msm::metal_msm::utils::limbs_conversion::{FromLimbs, ToLimbs};
use crate::msm::metal_msm::utils::mont_params::{calc_mont_radix, calc_rinv_and_n0};
use ark_bn254::Fr as ScalarField;
use ark_bn254::Fq as BaseField;
use ark_ff::{BigInt, PrimeField};
use metal::*;
use num_bigint::BigUint;
use num_bigint::{BigUint, RandBigInt};
use rand::thread_rng;

#[test]
#[serial_test::serial]
Expand All @@ -28,33 +29,26 @@ pub fn test_mont_mul_13() {

pub fn do_test(log_limb_size: u32) {
// Calculate num_limbs based on modulus size and limb size
let modulus_bits = ScalarField::MODULUS_BIT_SIZE as u32;
let modulus_bits = BaseField::MODULUS_BIT_SIZE as u32;
let num_limbs = ((modulus_bits + log_limb_size - 1) / log_limb_size) as usize;

let r = calc_mont_radix(num_limbs, log_limb_size);
let p: BigUint = ScalarField::MODULUS.try_into().unwrap();
let p: BigUint = BaseField::MODULUS.try_into().unwrap();

let res = calc_rinv_and_n0(&p, &r, log_limb_size);
let n0 = res.1;

let a = BigUint::parse_bytes(
b"10ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let b = BigUint::parse_bytes(
b"11ab655e9a2ca55660b44d1e5c37b00159aa76fed00000010a11800000000001",
16,
)
.unwrap();
let mut rng = thread_rng();
let a = rng.gen_biguint_below(&p);
let b = rng.gen_biguint_below(&p);

let a_r = &a * &r % &p;
let b_r = &b * &r % &p;
let expected = (&a * &b * &r) % &p;

let a_r_in_ark = ScalarField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = ScalarField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = ScalarField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let a_r_in_ark = BaseField::from_bigint(a_r.clone().try_into().unwrap()).unwrap();
let b_r_in_ark = BaseField::from_bigint(b_r.clone().try_into().unwrap()).unwrap();
let expected_in_ark = BaseField::from_bigint(expected.clone().try_into().unwrap()).unwrap();
let expected_limbs = expected_in_ark
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
Expand All @@ -70,7 +64,7 @@ pub fn do_test(log_limb_size: u32) {
);
let p_buf = create_buffer(
&device,
&ScalarField::MODULUS.to_limbs(num_limbs, log_limb_size),
&BaseField::MODULUS.to_limbs(num_limbs, log_limb_size),
);
let result_buf = create_empty_buffer(&device, num_limbs);

Expand Down Expand Up @@ -139,17 +133,17 @@ pub fn do_test(log_limb_size: u32) {
pub fn test_number_conversions() {
// Setup parameters
let log_limb_size = 12;
let modulus_bits = ScalarField::MODULUS_BIT_SIZE as u32;
let modulus_bits = BaseField::MODULUS_BIT_SIZE as u32;
let num_limbs = ((modulus_bits + log_limb_size - 1) / log_limb_size) as usize;

// Create test values using small numbers for clarity
let original_biguint = BigUint::parse_bytes(b"123456789", 10).unwrap();

// Convert BigUint to ScalarField
// Convert BigUint to BaseField
let scalar_field_value =
ScalarField::from_bigint(original_biguint.clone().try_into().unwrap()).unwrap();
BaseField::from_bigint(original_biguint.clone().try_into().unwrap()).unwrap();

// Convert ScalarField to limbs
// Convert BaseField to limbs
let limbs = scalar_field_value
.into_bigint()
.to_limbs(num_limbs, log_limb_size);
Expand All @@ -174,7 +168,7 @@ pub fn test_number_conversions() {
];

for value in test_values {
let scalar = ScalarField::from_bigint(value.clone().try_into().unwrap()).unwrap();
let scalar = BaseField::from_bigint(value.clone().try_into().unwrap()).unwrap();
let value_limbs = scalar.into_bigint().to_limbs(num_limbs, log_limb_size);
let converted: BigUint = BigInt::from_limbs(&value_limbs, log_limb_size)
.try_into()
Expand Down