From 29769816f64d2cf4671b3f0181e72a3906728902 Mon Sep 17 00:00:00 2001
From: Cheng <git@zcbenz.com>
Date: Thu, 26 Mar 2026 00:49:18 -0700
Subject: [PATCH] Remove no longer needed const_cast

---
 mlx/backend/cuda/reduce/all_reduce.cu | 14 +++-----------
 mlx/backend/cuda/reduce/col_reduce.cu |  8 ++------
 mlx/backend/cuda/reduce/row_reduce.cu |  9 +++++++--
 3 files changed, 12 insertions(+), 19 deletions(-)

diff --git a/mlx/backend/cuda/reduce/all_reduce.cu b/mlx/backend/cuda/reduce/all_reduce.cu
index dd4980a971..0659504c5b 100644
--- a/mlx/backend/cuda/reduce/all_reduce.cu
+++ b/mlx/backend/cuda/reduce/all_reduce.cu
@@ -98,9 +98,7 @@ void all_reduce(
   size_t block_step;
   size_t insize = in.size();
   Dtype dt = in.dtype();
-
-  // Cub doesn't like const pointers for load (sigh).
-  void* indata = const_cast<void*>(gpu_ptr<void>(in));
+  void* indata = gpu_ptr<void>(in);
 
   // Large array so allocate an intermediate and accumulate there
   std::tie(blocks, threads, block_step) = get_args(insize, N_READS);
@@ -120,7 +118,7 @@ void all_reduce(
             kernel,
             blocks,
             threads,
-            static_cast<T*>(indata),
+            indata,
             gpu_ptr<U>(intermediate),
             block_step,
             insize);
@@ -143,13 +141,7 @@ void all_reduce(
       using U = typename cu::ReduceResult<OP, T>::type;
       auto kernel = cu::all_reduce<T, U, OP, N_READS>;
       encoder.add_kernel_node(
-          kernel,
-          blocks,
-          threads,
-          static_cast<T*>(indata),
-          gpu_ptr<U>(out),
-          block_step,
-          insize);
+          kernel, blocks, threads, indata, gpu_ptr<U>(out), block_step, insize);
     });
   });
 }
diff --git a/mlx/backend/cuda/reduce/col_reduce.cu b/mlx/backend/cuda/reduce/col_reduce.cu
index b3b5a93fd7..8a6327bf38 100644
--- a/mlx/backend/cuda/reduce/col_reduce.cu
+++ b/mlx/backend/cuda/reduce/col_reduce.cu
@@ -282,8 +282,6 @@ void col_reduce_looped(
         using OP = MLX_GET_TYPE(reduce_type_tag);
         using T = cuda_type_t<MLX_GET_TYPE(type_tag)>;
         using U = typename cu::ReduceResult<OP, T>::type;
-        // Cub doesn't like const pointers for vectorized loads. (sigh)
-        T* indata = const_cast<T*>(gpu_ptr<T>(in));
 
         constexpr int N_READS = 4;
         constexpr int BM = 32;
@@ -296,7 +294,7 @@ void col_reduce_looped(
             kernel,
             grid,
             blocks,
-            indata,
+            gpu_ptr<T>(in),
             gpu_ptr<U>(out),
             static_cast<cu::ColReduceArgs>(args),
             out.size() / args.reduction_stride);
@@ -389,8 +387,6 @@ void col_reduce_two_pass(
         using OP = MLX_GET_TYPE(reduce_type_tag);
         using T = cuda_type_t<MLX_GET_TYPE(type_tag)>;
         using U = typename cu::ReduceResult<OP, T>::type;
-        // Cub doesn't like const pointers for vectorized loads. (sigh)
-        T* indata = const_cast<T*>(gpu_ptr<T>(in));
 
         constexpr int N_READS = 4;
         constexpr int BM = 32;
@@ -403,7 +399,7 @@ void col_reduce_two_pass(
             kernel,
             grid,
             blocks,
-            indata,
+            gpu_ptr<T>(in),
             gpu_ptr<U>(intermediate),
             static_cast<cu::ColReduceArgs>(args),
             out.size() / args.reduction_stride);
diff --git a/mlx/backend/cuda/reduce/row_reduce.cu b/mlx/backend/cuda/reduce/row_reduce.cu
index d09071669c..1dd40d14bd 100644
--- a/mlx/backend/cuda/reduce/row_reduce.cu
+++ b/mlx/backend/cuda/reduce/row_reduce.cu
@@ -268,10 +268,15 @@ void row_reduce_simple(
         kernel = cu::row_reduce_simple<T, U, OP, N_READS, 2>;
       }
 
-      T* indata = const_cast<T*>(gpu_ptr<T>(in));
       int size = plan.shape.back();
       encoder.add_kernel_node(
-          kernel, grid, block, indata, gpu_ptr<U>(out), out.size(), size);
+          kernel,
+          grid,
+          block,
+          gpu_ptr<T>(in),
+          gpu_ptr<U>(out),
+          out.size(),
+          size);
     });
   });
 }