diff --git a/plonky2/src/gadgets/split_join.rs b/plonky2/src/gadgets/split_join.rs
index 2f35b94c77..d20c09bc5f 100644
--- a/plonky2/src/gadgets/split_join.rs
+++ b/plonky2/src/gadgets/split_join.rs
@@ -13,7 +13,6 @@ use crate::iop::target::{BoolTarget, Target};
 use crate::iop::witness::{PartitionWitness, Witness, WitnessWrite};
 use crate::plonk::circuit_builder::CircuitBuilder;
 use crate::plonk::circuit_data::CommonCircuitData;
-use crate::util::ceil_div_usize;
 use crate::util::serialization::{Buffer, IoResult, Read, Write};
 
 impl<F: RichField + Extendable<D>, const D: usize> CircuitBuilder<F, D> {
@@ -26,7 +25,7 @@ impl<F: RichField + Extendable<D>, const D: usize> CircuitBuilder<F, D> {
             return Vec::new();
         }
         let gate_type = BaseSumGate::<2>::new_from_config::<F>(&self.config);
-        let k = ceil_div_usize(num_bits, gate_type.num_limbs);
+        let k = num_bits.div_ceil(gate_type.num_limbs);
         let gates = (0..k)
             .map(|_| self.add_gate(gate_type, vec![]))
             .collect::<Vec<_>>();
diff --git a/plonky2/src/gates/lookup_table.rs b/plonky2/src/gates/lookup_table.rs
index 28f47c4d8a..e7ead9aff3 100644
--- a/plonky2/src/gates/lookup_table.rs
+++ b/plonky2/src/gates/lookup_table.rs
@@ -11,7 +11,6 @@ use std::sync::Arc;
 
 use itertools::Itertools;
 use keccak_hash::keccak;
-use plonky2_util::ceil_div_usize;
 
 use crate::field::extension::Extendable;
 use crate::field::packed::PackedField;
@@ -207,7 +206,7 @@ impl<F: RichField + Extendable<D>, const D: usize> SimpleGenerator<F, D> for Loo
     }
 
     fn run_once(&self, _witness: &PartitionWitness<F>, out_buffer: &mut GeneratedValues<F>) {
-        let first_row = self.last_lut_row + ceil_div_usize(self.lut.len(), self.num_slots) - 1;
+        let first_row = self.last_lut_row + self.lut.len().div_ceil(self.num_slots) - 1;
         let slot = (first_row - self.row) * self.num_slots + self.slot_nb;
 
         let slot_input_target =
diff --git a/plonky2/src/plonk/circuit_builder.rs b/plonky2/src/plonk/circuit_builder.rs
index de3ab16fdf..5f153c637e 100644
--- a/plonky2/src/plonk/circuit_builder.rs
+++ b/plonky2/src/plonk/circuit_builder.rs
@@ -9,7 +9,6 @@ use std::{collections::BTreeMap, sync::Arc, time::Instant};
 use hashbrown::{HashMap, HashSet};
 use itertools::Itertools;
 use log::{debug, info, warn, Level};
-use plonky2_util::ceil_div_usize;
 
 use crate::field::cosets::get_unique_coset_shifts;
 use crate::field::extension::{Extendable, FieldExtension};
@@ -1219,7 +1218,7 @@ impl<F: RichField + Extendable<D>, const D: usize> CircuitBuilder<F, D> {
             0
         } else {
             // There is 1 RE polynomial and multiple Sum/LDC polynomials.
-            ceil_div_usize(LookupGate::num_slots(&self.config), lookup_degree) + 1
+            LookupGate::num_slots(&self.config).div_ceil(lookup_degree) + 1
         };
         let constants_sigmas_cap = constants_sigmas_commitment.merkle_tree.cap.clone();
         let domain_separator = self.domain_separator.unwrap_or_default();
diff --git a/plonky2/src/plonk/prover.rs b/plonky2/src/plonk/prover.rs
index 400845bc6f..3001e60eb6 100644
--- a/plonky2/src/plonk/prover.rs
+++ b/plonky2/src/plonk/prover.rs
@@ -33,7 +33,7 @@ use crate::plonk::vars::EvaluationVarsBaseBatch;
 use crate::timed;
 use crate::util::partial_products::{partial_products_and_z_gx, quotient_chunk_products};
 use crate::util::timing::TimingTree;
-use crate::util::{ceil_div_usize, log2_ceil, transpose};
+use crate::util::{log2_ceil, transpose};
 
 /// Set all the lookup gate wires (including multiplicities) and pad unused LU slots.
 /// Warning: rows are in descending order: the first gate to appear is the last LU gate, and
@@ -462,9 +462,9 @@ fn compute_lookup_polys<
     let degree = common_data.degree();
     let num_lu_slots = LookupGate::num_slots(&common_data.config);
     let max_lookup_degree = common_data.config.max_quotient_degree_factor - 1;
-    let num_partial_lookups = ceil_div_usize(num_lu_slots, max_lookup_degree);
+    let num_partial_lookups = num_lu_slots.div_ceil(max_lookup_degree);
     let num_lut_slots = LookupTableGate::num_slots(&common_data.config);
-    let max_lookup_table_degree = ceil_div_usize(num_lut_slots, num_partial_lookups);
+    let max_lookup_table_degree = num_lut_slots.div_ceil(num_partial_lookups);
 
     // First poly is RE, the rest are partial SLDCs.
     let mut final_poly_vecs = Vec::with_capacity(num_partial_lookups + 1);
@@ -652,10 +652,10 @@ fn compute_quotient_polys<
 
                 (LookupSelectors::StartEnd as usize..common_data.num_lookup_selectors)
                     .map(|r| {
-                        let lut_row_number = ceil_div_usize(
-                            common_data.luts[r - LookupSelectors::StartEnd as usize].len(),
-                            num_lut_slots,
-                        );
+                        let lut_row_number = common_data.luts
+                            [r - LookupSelectors::StartEnd as usize]
+                            .len()
+                            .div_ceil(num_lut_slots);
 
                         get_lut_poly(
                             common_data,
@@ -676,7 +676,7 @@ fn compute_quotient_polys<
         lut_re_poly_evals.iter().map(|v| v.as_slice()).collect();
 
     let points_batches = points.par_chunks(BATCH_SIZE);
-    let num_batches = ceil_div_usize(points.len(), BATCH_SIZE);
+    let num_batches = points.len().div_ceil(BATCH_SIZE);
 
     let quotient_values: Vec<Vec<F>> = points_batches
         .enumerate()
diff --git a/plonky2/src/plonk/vanishing_poly.rs b/plonky2/src/plonk/vanishing_poly.rs
index 9eb7252e1a..ef10108106 100644
--- a/plonky2/src/plonk/vanishing_poly.rs
+++ b/plonky2/src/plonk/vanishing_poly.rs
@@ -3,7 +3,6 @@ use alloc::{format, vec, vec::Vec};
 use core::cmp::min;
 
 use plonky2_field::polynomial::PolynomialCoeffs;
-use plonky2_util::ceil_div_usize;
 
 use super::circuit_builder::{LookupChallenges, NUM_COINS_LOOKUP};
 use super::vars::EvaluationVarsBase;
@@ -345,7 +344,7 @@ pub fn check_lookup_constraints<F: RichField + Extendable<D>, const D: usize>(
     let num_lut_slots = LookupTableGate::num_slots(&common_data.config);
     let lu_degree = common_data.quotient_degree_factor - 1;
     let num_sldc_polys = local_lookup_zs.len() - 1;
-    let lut_degree = ceil_div_usize(num_lut_slots, num_sldc_polys);
+    let lut_degree = num_lut_slots.div_ceil(num_sldc_polys);
 
     let mut constraints = Vec::with_capacity(4 + common_data.luts.len() + 2 * num_sldc_polys);
 
@@ -402,10 +401,9 @@ pub fn check_lookup_constraints<F: RichField + Extendable<D>, const D: usize>(
     // Check final RE constraints for each different LUT.
     for r in LookupSelectors::StartEnd as usize..common_data.num_lookup_selectors {
         let cur_ends_selector = lookup_selectors[r];
-        let lut_row_number = ceil_div_usize(
-            common_data.luts[r - LookupSelectors::StartEnd as usize].len(),
-            num_lut_slots,
-        );
+        let lut_row_number = common_data.luts[r - LookupSelectors::StartEnd as usize]
+            .len()
+            .div_ceil(num_lut_slots);
         let cur_function_eval = get_lut_poly(
             common_data,
             r - LookupSelectors::StartEnd as usize,
@@ -519,7 +517,7 @@ pub fn check_lookup_constraints_batch<F: RichField + Extendable<D>, const D: usi
     let num_lut_slots = LookupTableGate::num_slots(&common_data.config);
     let lu_degree = common_data.quotient_degree_factor - 1;
     let num_sldc_polys = local_lookup_zs.len() - 1;
-    let lut_degree = ceil_div_usize(num_lut_slots, num_sldc_polys);
+    let lut_degree = num_lut_slots.div_ceil(num_sldc_polys);
 
     let mut constraints = Vec::with_capacity(4 + common_data.luts.len() + 2 * num_sldc_polys);
 
@@ -931,7 +929,7 @@ pub fn check_lookup_constraints_circuit<F: RichField + Extendable<D>, const D: u
     let num_lut_slots = LookupTableGate::num_slots(&common_data.config);
     let lu_degree = common_data.quotient_degree_factor - 1;
     let num_sldc_polys = local_lookup_zs.len() - 1;
-    let lut_degree = ceil_div_usize(num_lut_slots, num_sldc_polys);
+    let lut_degree = num_lut_slots.div_ceil(num_sldc_polys);
 
     let mut constraints = Vec::with_capacity(4 + common_data.luts.len() + 2 * num_sldc_polys);
 
@@ -1023,10 +1021,9 @@ pub fn check_lookup_constraints_circuit<F: RichField + Extendable<D>, const D: u
     // Check final RE constraints for each different LUT.
     for r in LookupSelectors::StartEnd as usize..common_data.num_lookup_selectors {
         let cur_ends_selectors = lookup_selectors[r];
-        let lut_row_number = ceil_div_usize(
-            common_data.luts[r - LookupSelectors::StartEnd as usize].len(),
-            num_lut_slots,
-        );
+        let lut_row_number = common_data.luts[r - LookupSelectors::StartEnd as usize]
+            .len()
+            .div_ceil(num_lut_slots);
         let cur_function_eval = get_lut_poly_circuit(
             builder,
             common_data,
diff --git a/plonky2/src/recursion/dummy_circuit.rs b/plonky2/src/recursion/dummy_circuit.rs
index 501a475f1b..75b3929e82 100644
--- a/plonky2/src/recursion/dummy_circuit.rs
+++ b/plonky2/src/recursion/dummy_circuit.rs
@@ -8,7 +8,6 @@ use alloc::{
 use hashbrown::HashMap;
 use plonky2_field::extension::Extendable;
 use plonky2_field::polynomial::PolynomialCoeffs;
-use plonky2_util::ceil_div_usize;
 
 use crate::fri::proof::{FriProof, FriProofTarget};
 use crate::gadgets::polynomial::PolynomialCoeffsExtTarget;
@@ -103,7 +102,7 @@ pub(crate) fn dummy_circuit<
     // Number of `NoopGate`s to add to get a circuit of size `degree` in the end.
     // Need to account for public input hashing, a `PublicInputGate` and a `ConstantGate`.
     let degree = common_data.degree();
-    let num_noop_gate = degree - ceil_div_usize(common_data.num_public_inputs, 8) - 2;
+    let num_noop_gate = degree - common_data.num_public_inputs.div_ceil(8) - 2;
 
     let mut builder = CircuitBuilder::<F, D>::new(config);
     for _ in 0..num_noop_gate {
diff --git a/plonky2/src/util/partial_products.rs b/plonky2/src/util/partial_products.rs
index 1ceea7cab1..ac02188fd8 100644
--- a/plonky2/src/util/partial_products.rs
+++ b/plonky2/src/util/partial_products.rs
@@ -9,7 +9,6 @@ use crate::field::types::Field;
 use crate::hash::hash_types::RichField;
 use crate::iop::ext_target::ExtensionTarget;
 use crate::plonk::circuit_builder::CircuitBuilder;
-use crate::util::ceil_div_usize;
 
 pub(crate) fn quotient_chunk_products<F: Field>(
     quotient_values: &[F],
@@ -41,10 +40,10 @@ pub(crate) fn partial_products_and_z_gx<F: Field>(z_x: F, quotient_chunk_product
 pub(crate) fn num_partial_products(n: usize, max_degree: usize) -> usize {
     debug_assert!(max_degree > 1);
     let chunk_size = max_degree;
-    // We'll split the product into `ceil_div_usize(n, chunk_size)` chunks, but the last chunk will
+    // We'll split the product into `n.div_ceil( chunk_size)` chunks, but the last chunk will
     // be associated with Z(gx) itself. Thus we subtract one to get the chunks associated with
     // partial products.
-    ceil_div_usize(n, chunk_size) - 1
+    n.div_ceil(chunk_size) - 1
 }
 
 /// Checks the relationship between each pair of partial product accumulators. In particular, this
diff --git a/starky/src/cross_table_lookup.rs b/starky/src/cross_table_lookup.rs
index d0b020a6a0..73caf28cc8 100644
--- a/starky/src/cross_table_lookup.rs
+++ b/starky/src/cross_table_lookup.rs
@@ -45,7 +45,6 @@ use plonky2::iop::ext_target::ExtensionTarget;
 use plonky2::iop::target::Target;
 use plonky2::plonk::circuit_builder::CircuitBuilder;
 use plonky2::plonk::config::GenericConfig;
-use plonky2::util::ceil_div_usize;
 
 use crate::config::StarkConfig;
 use crate::constraint_consumer::{ConstraintConsumer, RecursiveConstraintConsumer};
@@ -125,7 +124,7 @@ impl<F: Field> CrossTableLookup<F> {
             let num_appearances = all_tables.filter(|twc| twc.table == table).count();
             let is_helpers = num_appearances > 1;
             if is_helpers {
-                num_helpers_by_ctl[i] = ceil_div_usize(num_appearances, constraint_degree - 1);
+                num_helpers_by_ctl[i] = num_appearances.div_ceil(constraint_degree - 1);
                 num_helpers += num_helpers_by_ctl[i];
             }
 
@@ -292,7 +291,7 @@ pub(crate) fn num_ctl_helper_columns_by_table<F: Field, const N: usize>(
             let sum = group.count();
             if sum > 1 {
                 // We only need helper columns if there are at least 2 columns.
-                num_by_table[table] = ceil_div_usize(sum, constraint_degree - 1);
+                num_by_table[table] = sum.div_ceil(constraint_degree - 1);
             }
         }
 
diff --git a/starky/src/lookup.rs b/starky/src/lookup.rs
index f63c4bd76c..2a90e6b93e 100644
--- a/starky/src/lookup.rs
+++ b/starky/src/lookup.rs
@@ -24,7 +24,6 @@ use plonky2::plonk::plonk_common::{
     reduce_with_powers, reduce_with_powers_circuit, reduce_with_powers_ext_circuit,
 };
 use plonky2::util::serialization::{Buffer, IoResult, Read, Write};
-use plonky2_util::ceil_div_usize;
 
 use crate::constraint_consumer::{ConstraintConsumer, RecursiveConstraintConsumer};
 use crate::evaluation_frame::StarkEvaluationFrame;
@@ -441,10 +440,11 @@ impl<F: Field> Lookup<F> {
     pub fn num_helper_columns(&self, constraint_degree: usize) -> usize {
         // One helper column for each column batch of size `constraint_degree-1`,
         // then one column for the inverse of `table + challenge` and one for the `Z` polynomial.
-        ceil_div_usize(
-            self.columns.len(),
-            constraint_degree.checked_sub(1).unwrap_or(1),
-        ) + 1
+
+        self.columns
+            .len()
+            .div_ceil(constraint_degree.checked_sub(1).unwrap_or(1))
+            + 1
     }
 }
 
@@ -757,10 +757,9 @@ pub(crate) fn get_helper_cols<F: Field>(
     challenge: GrandProductChallenge<F>,
     constraint_degree: usize,
 ) -> Vec<PolynomialValues<F>> {
-    let num_helper_columns = ceil_div_usize(
-        columns_filters.len(),
-        constraint_degree.checked_sub(1).unwrap_or(1),
-    );
+    let num_helper_columns = columns_filters
+        .len()
+        .div_ceil(constraint_degree.checked_sub(1).unwrap_or(1));
 
     let mut helper_columns = Vec::with_capacity(num_helper_columns);
 
diff --git a/util/src/lib.rs b/util/src/lib.rs
index 613bf6ef5a..687e9237f8 100644
--- a/util/src/lib.rs
+++ b/util/src/lib.rs
@@ -16,10 +16,6 @@ pub const fn bits_u64(n: u64) -> usize {
     (64 - n.leading_zeros()) as usize
 }
 
-pub const fn ceil_div_usize(a: usize, b: usize) -> usize {
-    (a + b - 1) / b
-}
-
 /// Computes `ceil(log_2(n))`.
 #[must_use]
 pub const fn log2_ceil(n: usize) -> usize {