diff --git a/kernels/portable/cpu/op_allclose.cpp b/kernels/portable/cpu/op_allclose.cpp
index a1b2cb87389..32d81f1ad80 100644
--- a/kernels/portable/cpu/op_allclose.cpp
+++ b/kernels/portable/cpu/op_allclose.cpp
@@ -1,7 +1,7 @@
 /*
  * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
- * Copyright 2025 Arm Limited and/or its affiliates.
+ * Copyright 2025-2026 Arm Limited and/or its affiliates.
  *
  * This source code is licensed under the BSD-style license found in the
  * LICENSE file in the root directory of this source tree.
@@ -41,8 +41,8 @@ bool data_is_close(
         return false;
       }
     } else {
-      auto allowed_error = atol + fabs(rtol * b[i]);
-      auto actual_error = fabs(a[i] - b[i]);
+      auto allowed_error = atol + std::fabs(rtol * b[i]);
+      auto actual_error = std::fabs(a[i] - b[i]);
       if (!std::isfinite(actual_error) || actual_error > allowed_error) {
         return false;
       }
diff --git a/kernels/portable/cpu/util/upsample_util.h b/kernels/portable/cpu/util/upsample_util.h
index 7ac1fec591f..6940c63dedb 100644
--- a/kernels/portable/cpu/util/upsample_util.h
+++ b/kernels/portable/cpu/util/upsample_util.h
@@ -1,6 +1,7 @@
 /*
  * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
+ * Copyright 2026 Arm Limited and/or its affiliates.
  *
  * This source code is licensed under the BSD-style license found in the
  * LICENSE file in the root directory of this source tree.
@@ -8,6 +9,8 @@
 
 #pragma once
 
+#include <cmath>
+
 #include <executorch/runtime/core/exec_aten/exec_aten.h>
 #include <executorch/runtime/core/exec_aten/util/tensor_util.h>
 #include <executorch/runtime/kernel/kernel_includes.h>
@@ -98,8 +101,8 @@ inline void guard_index_and_lambda(
     const int64_t& input_size,
     int64_t& input_index,
     scalar_t& lambda) {
-  input_index =
-      std::min(static_cast<int64_t>(floorf(real_input_index)), input_size - 1);
+  input_index = std::min(
+      static_cast<int64_t>(std::floor(real_input_index)), input_size - 1);
   lambda = std::min(
       std::max(real_input_index - input_index, static_cast<opmath_t>(0)),
       static_cast<opmath_t>(1));
@@ -140,8 +143,8 @@ inline int64_t nearest_neighbor_compute_source_index(
     int64_t input_size) {
   // Index computation matching OpenCV INTER_NEAREST
   // which is buggy and kept for BC
-  const int64_t src_index =
-      std::min(static_cast<int64_t>(floorf(dst_index * scale)), input_size - 1);
+  const int64_t src_index = std::min(
+      static_cast<int64_t>(std::floor(dst_index * scale)), input_size - 1);
   return src_index;
 }
 
diff --git a/kernels/portable/cpu/vec_ops.h b/kernels/portable/cpu/vec_ops.h
index bafc1b24879..0580ddc9336 100644
--- a/kernels/portable/cpu/vec_ops.h
+++ b/kernels/portable/cpu/vec_ops.h
@@ -1,7 +1,7 @@
 /*
  * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
- * Copyright 2024 Arm Limited and/or its affiliates.
+ * Copyright 2024,2026 Arm Limited and/or its affiliates.
  *
  * This source code is licensed under the BSD-style license found in the
  * LICENSE file in the root directory of this source tree.
@@ -201,7 +201,7 @@ inline void vec_softmax(T* ET_RESTRICT y, const U* ET_RESTRICT x, int n) {
   T sum = 0;
 
   for (const auto i : c10::irange(n)) {
-    y[i] = expf(x[i] - max_x);
+    y[i] = std::exp(x[i] - max_x);
     sum += y[i];
   }
 
@@ -230,7 +230,7 @@ inline void quantize_i8_f32(
     int32_t zero_point,
     size_t size) {
   for (const auto i : c10::irange(size)) {
-    float tmp = roundf(x[i] * scale + zero_point);
+    float tmp = std::round(x[i] * scale + zero_point);
     y[i] = internal::clamp(tmp, -128.f, 127.f);
   }
 }