Mz/rename multi bit

backends/tfhe-cuda-backend/cuda/src/pbs/programmable_bootstrap_multibit.cuh

@@ -19,12 +19,12 @@
 
 template <typename Torus, class params>
 __device__ Torus calculates_monomial_degree(Torus *lwe_array_group,
-                                            uint32_t ggsw_idx,
+                                            uint32_t power_set_index,
                                             uint32_t grouping_factor) {
   Torus x = 0;
   for (int i = 0; i < grouping_factor; i++) {
     uint32_t mask_position = grouping_factor - (i + 1);
-    int selection_bit = (ggsw_idx >> mask_position) & 1;
+    int selection_bit = (power_set_index >> mask_position) & 1;
     x += selection_bit * lwe_array_group[i];
   }
 

tfhe/benches/core_crypto/dev_bench.rs

@@ -24,7 +24,7 @@ fn get_bench_params<Scalar: Numeric>() -> (
     DecompositionLevelCount,
     GlweDimension,
     PolynomialSize,
-    LweBskGroupingFactor,
+    MultiBitGroupingFactor,
     ThreadCount,
 ) {
     if Scalar::BITS == 64 {
@@ -35,7 +35,7 @@ fn get_bench_params<Scalar: Numeric>() -> (
             DecompositionLevelCount(5),
             GlweDimension(1),
             PolynomialSize(1024),
-            LweBskGroupingFactor(2),
+            MultiBitGroupingFactor(2),
             ThreadCount(5),
         )
     } else if Scalar::BITS == 32 {
@@ -46,7 +46,7 @@ fn get_bench_params<Scalar: Numeric>() -> (
             DecompositionLevelCount(1),
             GlweDimension(3),
             PolynomialSize(512),
-            LweBskGroupingFactor(2),
+            MultiBitGroupingFactor(2),
             ThreadCount(5),
         )
     } else {

tfhe/benches/core_crypto/pbs_bench.rs

@@ -104,8 +104,11 @@ fn throughput_benchmark_parameters<Scalar: UnsignedInteger>(
     }
 }
 
-fn multi_bit_benchmark_parameters<Scalar: UnsignedInteger + Default>(
-) -> Vec<(String, CryptoParametersRecord<Scalar>, LweBskGroupingFactor)> {
+fn multi_bit_benchmark_parameters<Scalar: UnsignedInteger + Default>() -> Vec<(
+    String,
+    CryptoParametersRecord<Scalar>,
+    MultiBitGroupingFactor,
+)> {
     if Scalar::BITS == 64 {
         let parameters = if cfg!(feature = "gpu") {
             vec![

tfhe/src/c_api/core_crypto/mod.rs

@@ -330,7 +330,7 @@ pub unsafe extern "C" fn core_crypto_lwe_multi_bit_bootstrapping_key_element_siz
         let output_glwe_sk_poly_size = PolynomialSize(output_glwe_sk_poly_size);
 
         let lwe_multi_bit_level_count = DecompositionLevelCount(lwe_multi_bit_level_count);
-        let lwe_multi_bit_grouping_factor = LweBskGroupingFactor(lwe_multi_bit_grouping_factor);
+        let lwe_multi_bit_grouping_factor = MultiBitGroupingFactor(lwe_multi_bit_grouping_factor);
 
         *result = lwe_multi_bit_bootstrap_key_size(
             input_lwe_sk_dim,
@@ -472,7 +472,7 @@ pub unsafe extern "C" fn core_crypto_par_generate_lwe_multi_bit_bootstrapping_ke
 
         let lwe_multi_bit_base_log = DecompositionBaseLog(lwe_multi_bit_base_log);
         let lwe_multi_bit_level_count = DecompositionLevelCount(lwe_multi_bit_level_count);
-        let lwe_multi_bit_grouping_factor = LweBskGroupingFactor(lwe_multi_bit_grouping_factor);
+        let lwe_multi_bit_grouping_factor = MultiBitGroupingFactor(lwe_multi_bit_grouping_factor);
 
         let lwe_multi_bit_slice_len = {
             let bsk = LweMultiBitBootstrapKeyOwned::new(

tfhe/src/core_crypto/algorithms/lwe_multi_bit_bootstrap_key_generation.rs

@@ -25,7 +25,7 @@ use rayon::prelude::*;
 /// let polynomial_size = PolynomialSize(1024);
 /// let glwe_noise_distribution =
 ///     Gaussian::from_dispersion_parameter(StandardDev(0.00000000000000029403601535432533), 0.0);
-/// let grouping_factor = LweBskGroupingFactor(2);
+/// let grouping_factor = MultiBitGroupingFactor(2);
 /// let ciphertext_modulus = CiphertextModulus::new_native();
 ///
 /// // Create the PRNG
@@ -64,9 +64,9 @@ use rayon::prelude::*;
 ///     &mut encryption_generator,
 /// );
 ///
-/// let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+/// let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 ///
-/// for (mut ggsw_group, input_key_elements) in bsk.chunks_exact(ggsw_per_multi_bit_element.0).zip(
+/// for (mut ggsw_group, input_key_elements) in bsk.chunks_exact(multi_bit_power_set_size.0).zip(
 ///     input_lwe_secret_key
 ///         .as_ref()
 ///         .chunks_exact(grouping_factor.0),
@@ -137,10 +137,10 @@ pub fn generate_lwe_multi_bit_bootstrap_key<
         .unwrap();
 
     let output_grouping_factor = output.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     for ((mut ggsw_group, input_key_elements), mut loop_generator) in output
-        .chunks_exact_mut(ggsw_per_multi_bit_element.0)
+        .chunks_exact_mut(multi_bit_power_set_size.0)
         .zip(
             input_lwe_secret_key
                 .as_ref()
@@ -182,7 +182,7 @@ pub fn allocate_and_generate_new_lwe_multi_bit_bootstrap_key<
     output_glwe_secret_key: &GlweSecretKey<OutputKeyCont>,
     decomp_base_log: DecompositionBaseLog,
     decomp_level_count: DecompositionLevelCount,
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     noise_distribution: NoiseDistribution,
     ciphertext_modulus: CiphertextModulus<Scalar>,
     generator: &mut EncryptionRandomGenerator<Gen>,
@@ -232,7 +232,7 @@ where
 /// let polynomial_size = PolynomialSize(1024);
 /// let glwe_noise_distribution =
 ///     Gaussian::from_dispersion_parameter(StandardDev(0.00000000000000029403601535432533), 0.0);
-/// let grouping_factor = LweBskGroupingFactor(2);
+/// let grouping_factor = MultiBitGroupingFactor(2);
 /// let ciphertext_modulus = CiphertextModulus::new_native();
 ///
 /// // Create the PRNG
@@ -282,9 +282,9 @@ where
 ///
 /// par_convert_standard_lwe_multi_bit_bootstrap_key_to_fourier(&bsk, &mut multi_bit_bsk);
 ///
-/// let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+/// let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 ///
-/// for (mut ggsw_group, input_key_elements) in bsk.chunks_exact(ggsw_per_multi_bit_element.0).zip(
+/// for (mut ggsw_group, input_key_elements) in bsk.chunks_exact(multi_bit_power_set_size.0).zip(
 ///     input_lwe_secret_key
 ///         .as_ref()
 ///         .chunks_exact(grouping_factor.0),
@@ -354,10 +354,10 @@ pub fn par_generate_lwe_multi_bit_bootstrap_key<
         .unwrap();
 
     let output_grouping_factor = output.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     output
-        .par_chunks_exact_mut(ggsw_per_multi_bit_element.0)
+        .par_chunks_exact_mut(multi_bit_power_set_size.0)
         .zip(
             input_lwe_secret_key
                 .as_ref()
@@ -436,7 +436,7 @@ pub fn par_allocate_and_generate_new_lwe_multi_bit_bootstrap_key<
     output_glwe_secret_key: &GlweSecretKey<OutputKeyCont>,
     decomp_base_log: DecompositionBaseLog,
     decomp_level_count: DecompositionLevelCount,
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     noise_distribution: NoiseDistribution,
     ciphertext_modulus: CiphertextModulus<Scalar>,
     generator: &mut EncryptionRandomGenerator<Gen>,
@@ -559,10 +559,10 @@ pub fn generate_seeded_lwe_multi_bit_bootstrap_key<
         .unwrap();
 
     let output_grouping_factor = output.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     for ((mut ggsw_group, input_key_elements), mut loop_generator) in output
-        .chunks_exact_mut(ggsw_per_multi_bit_element.0)
+        .chunks_exact_mut(multi_bit_power_set_size.0)
         .zip(
             input_lwe_secret_key
                 .as_ref()
@@ -612,7 +612,7 @@ pub fn allocate_and_generate_new_seeded_lwe_multi_bit_bootstrap_key<
     decomp_base_log: DecompositionBaseLog,
     decomp_level_count: DecompositionLevelCount,
     noise_distribution: NoiseDistribution,
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     ciphertext_modulus: CiphertextModulus<Scalar>,
     noise_seeder: &mut NoiseSeeder,
 ) -> SeededLweMultiBitBootstrapKeyOwned<Scalar>
@@ -708,10 +708,10 @@ pub fn par_generate_seeded_lwe_multi_bit_bootstrap_key<
         .unwrap();
 
     let output_grouping_factor = output.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     output
-        .par_chunks_exact_mut(ggsw_per_multi_bit_element.0)
+        .par_chunks_exact_mut(multi_bit_power_set_size.0)
         .zip(
             input_lwe_secret_key
                 .as_ref()
@@ -766,7 +766,7 @@ pub fn par_allocate_and_generate_new_seeded_lwe_multi_bit_bootstrap_key<
     decomp_base_log: DecompositionBaseLog,
     decomp_level_count: DecompositionLevelCount,
     noise_distribution: NoiseDistribution,
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     ciphertext_modulus: CiphertextModulus<Scalar>,
     noise_seeder: &mut NoiseSeeder,
 ) -> SeededLweMultiBitBootstrapKeyOwned<Scalar>

tfhe/src/core_crypto/algorithms/lwe_multi_bit_programmable_bootstrapping.rs

@@ -22,14 +22,14 @@ pub use super::lwe_programmable_bootstrapping::generate_programmable_bootstrap_g
 
 pub fn modulus_switch_multi_bit<Scalar>(
     ciphertext_modulus_log: CiphertextModulusLog,
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     lwe_mask_elements: &[Scalar],
 ) -> impl Iterator<Item = usize> + '_
 where
     Scalar: UnsignedInteger + CastInto<usize> + CastFrom<usize>,
 {
     // Start at 1, the first ggsw is not rotated
-    (1..grouping_factor.ggsw_per_multi_bit_element().0).map(move |power_set_index| {
+    (1..grouping_factor.multi_bit_power_set_size().0).map(move |power_set_index| {
         let mut monomial_degree = Scalar::ZERO;
         for (&mask_element, selection_bit) in lwe_mask_elements
             .iter()
@@ -46,10 +46,10 @@ where
 // Returns an iterator of booleans (as usize), corresponding to successive mask group elements
 // to indicate if they must be used at the given power_set_index
 pub(crate) fn selection_bit(
-    grouping_factor: LweBskGroupingFactor,
+    grouping_factor: MultiBitGroupingFactor,
     power_set_index: usize,
 ) -> impl Iterator<Item = usize> {
-    debug_assert!(power_set_index < grouping_factor.ggsw_per_multi_bit_element().0);
+    debug_assert!(power_set_index < grouping_factor.multi_bit_power_set_size().0);
 
     (0..grouping_factor.0).map(move |mask_idx| {
         let mask_position = grouping_factor.0 - (mask_idx + 1);
@@ -72,7 +72,7 @@ pub struct StandardMultiBitModulusSwitchedCt<
     C: Container<Element = Scalar> + Sync,
 > {
     pub input: &'a LweCiphertext<C>,
-    pub grouping_factor: LweBskGroupingFactor,
+    pub grouping_factor: MultiBitGroupingFactor,
     pub log_modulus: CiphertextModulusLog,
 }
 
@@ -165,7 +165,7 @@ pub fn prepare_multi_bit_ggsw_mem_optimized<GgswBufferCont, GgswGroupCont, Fouri
 ///     Gaussian::from_dispersion_parameter(StandardDev(0.00000000000000029403601535432533), 0.0);
 /// let pbs_base_log = DecompositionBaseLog(23);
 /// let pbs_level = DecompositionLevelCount(1);
-/// let grouping_factor = LweBskGroupingFactor(2); // Group bits in pairs
+/// let grouping_factor = MultiBitGroupingFactor(2); // Group bits in pairs
 /// let ciphertext_modulus = CiphertextModulus::new_native();
 ///
 /// // Request the best seeder possible, starting with hardware entropy sources and falling back to
@@ -417,7 +417,7 @@ pub fn multi_bit_non_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCo
     let ggsw_vec: Vec<_> = multi_bit_bsk.ggsw_iter().collect();
 
     let grouping_factor = multi_bit_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 
     let input_lwe_dimension = multi_bit_bsk.input_lwe_dimension();
 
@@ -473,8 +473,8 @@ pub fn multi_bit_non_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCo
                 let switched_degrees =
                     switched_modulus_input.switched_modulus_input_mask_per_group(work_index);
 
-                let ggsw_group = &ggsw_vec[work_index * ggsw_per_multi_bit_element.0
-                    ..(work_index + 1) * ggsw_per_multi_bit_element.0];
+                let ggsw_group = &ggsw_vec[work_index * multi_bit_power_set_size.0
+                    ..(work_index + 1) * multi_bit_power_set_size.0];
 
                 let mut ready_for_consumer = ready_for_consumer_lock.lock().unwrap();
 
@@ -648,7 +648,7 @@ pub fn multi_bit_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCont>(
     let ggsw_vec: Vec<_> = multi_bit_bsk.ggsw_iter().collect();
 
     let grouping_factor = multi_bit_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 
     let input_lwe_dimension = multi_bit_bsk.input_lwe_dimension();
 
@@ -699,8 +699,8 @@ pub fn multi_bit_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCont>(
                 let switched_degrees =
                     switched_modulus_input.switched_modulus_input_mask_per_group(work_index);
 
-                let ggsw_group = &ggsw_vec[work_index * ggsw_per_multi_bit_element.0
-                    ..(work_index + 1) * ggsw_per_multi_bit_element.0];
+                let ggsw_group = &ggsw_vec[work_index * multi_bit_power_set_size.0
+                    ..(work_index + 1) * multi_bit_power_set_size.0];
 
                 let mut ready_for_consumer = ready_for_consumer_lock.lock().unwrap();
 
@@ -843,7 +843,7 @@ pub fn multi_bit_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCont>(
 ///     Gaussian::from_dispersion_parameter(StandardDev(0.00000000000000029403601535432533), 0.0);
 /// let pbs_base_log = DecompositionBaseLog(23);
 /// let pbs_level = DecompositionLevelCount(1);
-/// let grouping_factor = LweBskGroupingFactor(2); // Group bits in pairs
+/// let grouping_factor = MultiBitGroupingFactor(2); // Group bits in pairs
 /// let ciphertext_modulus = CiphertextModulus::new_native();
 ///
 /// // Request the best seeder possible, starting with hardware entropy sources and falling back to
@@ -1236,7 +1236,7 @@ pub fn std_multi_bit_non_deterministic_blind_rotate_assign<Scalar, OutputCont, K
     let ggsw_vec: Vec<_> = multi_bit_bsk.iter().collect();
 
     let grouping_factor = multi_bit_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 
     let input_lwe_dimension = multi_bit_bsk.input_lwe_dimension();
 
@@ -1312,8 +1312,8 @@ pub fn std_multi_bit_non_deterministic_blind_rotate_assign<Scalar, OutputCont, K
                 let switched_degrees =
                     switched_modulus_input.switched_modulus_input_mask_per_group(work_index);
 
-                let ggsw_group = &ggsw_vec[work_index * ggsw_per_multi_bit_element.0
-                    ..(work_index + 1) * ggsw_per_multi_bit_element.0];
+                let ggsw_group = &ggsw_vec[work_index * multi_bit_power_set_size.0
+                    ..(work_index + 1) * multi_bit_power_set_size.0];
 
                 let mut ready_for_consumer = ready_for_consumer_lock.lock().unwrap();
 
@@ -1502,7 +1502,7 @@ pub fn std_multi_bit_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCo
     let ggsw_vec: Vec<_> = multi_bit_bsk.iter().collect();
 
     let grouping_factor = multi_bit_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = grouping_factor.multi_bit_power_set_size();
 
     let input_lwe_dimension = multi_bit_bsk.input_lwe_dimension();
 
@@ -1573,8 +1573,8 @@ pub fn std_multi_bit_deterministic_blind_rotate_assign<Scalar, OutputCont, KeyCo
                 let switched_degrees =
                     switched_modulus_input.switched_modulus_input_mask_per_group(work_index);
 
-                let ggsw_group = &ggsw_vec[work_index * ggsw_per_multi_bit_element.0
-                    ..(work_index + 1) * ggsw_per_multi_bit_element.0];
+                let ggsw_group = &ggsw_vec[work_index * multi_bit_power_set_size.0
+                    ..(work_index + 1) * multi_bit_power_set_size.0];
 
                 let mut ready_for_consumer = ready_for_consumer_lock.lock().unwrap();
 

tfhe/src/core_crypto/algorithms/seeded_lwe_multi_bit_bootstrap_key_decompression.rs

@@ -35,15 +35,15 @@ pub fn decompress_seeded_lwe_multi_bit_bootstrap_key_with_existing_generator<
 
     // Forking logic must match multi bit BSK generation
     let output_grouping_factor = output_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     let forking_config = input_bsk.decompression_fork_config(Uniform);
 
     let gen_iter = generator.try_fork_from_config(forking_config).unwrap();
 
     for ((mut output_ggsw_group, input_ggsw_group), mut loop_generator) in output_bsk
-        .chunks_exact_mut(ggsw_per_multi_bit_element.0)
-        .zip(input_bsk.chunks_exact(ggsw_per_multi_bit_element.0))
+        .chunks_exact_mut(multi_bit_power_set_size.0)
+        .zip(input_bsk.chunks_exact(multi_bit_power_set_size.0))
         .zip(gen_iter)
     {
         let group_forking_config = input_ggsw_group.decompression_fork_config(Uniform);
@@ -121,15 +121,15 @@ pub fn par_decompress_seeded_lwe_multi_bit_bootstrap_key_with_existing_generator
 
     // Forking logic must match multi bit BSK generation
     let output_grouping_factor = output_bsk.grouping_factor();
-    let ggsw_per_multi_bit_element = output_grouping_factor.ggsw_per_multi_bit_element();
+    let multi_bit_power_set_size = output_grouping_factor.multi_bit_power_set_size();
 
     let forking_config = input_bsk.decompression_fork_config(Uniform);
 
     let gen_iter = generator.par_try_fork_from_config(forking_config).unwrap();
 
     output_bsk
-        .par_chunks_exact_mut(ggsw_per_multi_bit_element.0)
-        .zip(input_bsk.par_chunks_exact(ggsw_per_multi_bit_element.0))
+        .par_chunks_exact_mut(multi_bit_power_set_size.0)
+        .zip(input_bsk.par_chunks_exact(multi_bit_power_set_size.0))
         .zip(gen_iter)
         .for_each(
             |((mut output_ggsw_group, input_ggsw_group), mut loop_generator)| {

tfhe/src/core_crypto/algorithms/test/lwe_multi_bit_bootstrap_key_generation.rs

@@ -6,8 +6,8 @@ use crate::core_crypto::commons::math::random::{
 };
 use crate::core_crypto::commons::math::torus::UnsignedTorus;
 use crate::core_crypto::commons::parameters::{
-    CiphertextModulus, DecompositionBaseLog, DecompositionLevelCount, GlweDimension,
-    LweBskGroupingFactor, LweDimension, PolynomialSize,
+    CiphertextModulus, DecompositionBaseLog, DecompositionLevelCount, GlweDimension, LweDimension,
+    MultiBitGroupingFactor, PolynomialSize,
 };
 use crate::core_crypto::commons::test_tools::new_secret_random_generator;
 use crate::core_crypto::entities::*;
@@ -36,7 +36,7 @@ fn test_parallel_and_seeded_multi_bit_bsk_gen_equivalence<
         let base_log = DecompositionBaseLog(
             crate::core_crypto::commons::test_tools::random_usize_between(2..5),
         );
-        let grouping_factor = LweBskGroupingFactor(
+        let grouping_factor = MultiBitGroupingFactor(
             crate::core_crypto::commons::test_tools::random_usize_between(2..4),
         );
         let mask_seed = Seed(crate::core_crypto::commons::test_tools::any_usize() as u128);