From 8744d58e94ed5e653c725b0a88496a243a705c24 Mon Sep 17 00:00:00 2001
From: Torre Zuk <42548444+TorreZuk@users.noreply.github.com>
Date: Thu, 7 May 2026 16:07:14 -0600
Subject: [PATCH 1/2] adding OpenBLAS axpy replacements for WIN32

---
 .../clients/common/cblas_interface.cpp        | 69 +++++++++++++++++++
 .../clients/include/cblas_interface.hpp       |  2 +
 2 files changed, 71 insertions(+)

diff --git a/projects/rocblas/clients/common/cblas_interface.cpp b/projects/rocblas/clients/common/cblas_interface.cpp
index ef31c586681..00e7bba0262 100644
--- a/projects/rocblas/clients/common/cblas_interface.cpp
+++ b/projects/rocblas/clients/common/cblas_interface.cpp
@@ -93,6 +93,75 @@ void ref_axpy<rocblas_half>(
     // cblas_saxpy(n, alpha, x_float, incx, y_float, incy);
 }
 
+#if defined(WIN32) && !defined(BLIS_ENABLE_CBLAS)
+
+template <typename T>
+void ref_axpy(int64_t n, T alpha, T* x, int64_t incx, T* y, int64_t incy)
+{
+    // Handle negative increments
+    x += incx < 0 ? incx * (1 - n) : 0;
+    y += incy < 0 ? incy * (1 - n) : 0;
+
+    for(int64_t i = 0; i < n; i++)
+    {
+        y[i * incy] += alpha * x[i * incx];
+    }
+}
+
+template void
+    ref_axpy<float>(int64_t n, float alpha, float* x, int64_t incx, float* y, int64_t incy);
+
+template void
+    ref_axpy<double>(int64_t n, double alpha, double* x, int64_t incx, double* y, int64_t incy);
+
+template void ref_axpy<rocblas_float_complex>(int64_t                n,
+                                              rocblas_float_complex  alpha,
+                                              rocblas_float_complex* x,
+                                              int64_t                incx,
+                                              rocblas_float_complex* y,
+                                              int64_t                incy);
+
+template void ref_axpy<rocblas_double_complex>(int64_t                 n,
+                                               rocblas_double_complex  alpha,
+                                               rocblas_double_complex* x,
+                                               int64_t                 incx,
+                                               rocblas_double_complex* y,
+                                               int64_t                 incy);
+
+template <>
+void ref_axpy<rocblas_bfloat16>(int64_t            n,
+                                rocblas_bfloat16   alpha,
+                                rocblas_bfloat16*  x,
+                                int64_t            incx,
+                                rocblas_bfloat16*  y,
+                                int64_t            incy)
+{
+    // Handle negative increments
+    int64_t abs_incx = incx < 0 ? -incx : incx;
+    int64_t abs_incy = incy < 0 ? -incy : incy;
+
+    // Convert to float
+    host_vector<float> x_float(n * abs_incx);
+    host_vector<float> y_float(n * abs_incy);
+
+    for(int64_t i = 0; i < n; i++)
+    {
+        x_float[i * abs_incx] = float(x[i * abs_incx]);
+        y_float[i * abs_incy] = float(y[i * abs_incy]);
+    }
+
+    // Compute in float precision
+    ref_axpy<float>(n, float(alpha), x_float, incx, y_float, incy);
+
+    // Convert back to bfloat16
+    for(int64_t i = 0; i < n; i++)
+    {
+        y[i * abs_incy] = rocblas_bfloat16(y_float[i * abs_incy]);
+    }
+}
+
+#endif
+
 template <>
 void ref_asum<float>(int64_t n, const float* x, int64_t incx, float* result)
 {
diff --git a/projects/rocblas/clients/include/cblas_interface.hpp b/projects/rocblas/clients/include/cblas_interface.hpp
index 50820fa83f3..e10f8fa446f 100644
--- a/projects/rocblas/clients/include/cblas_interface.hpp
+++ b/projects/rocblas/clients/include/cblas_interface.hpp
@@ -94,6 +94,7 @@ inline void ref_asum(int64_t n, const rocblas_double_complex* x, int64_t incx, d
 template <typename T>
 void ref_axpy(int64_t n, T alpha, T* x, int64_t incx, T* y, int64_t incy);
 
+#if !(defined(WIN32) && !defined(BLIS_ENABLE_CBLAS)) // windows OpenBLAS bug override
 template <>
 inline void ref_axpy(int64_t n, float alpha, float* x, int64_t incx, float* y, int64_t incy)
 {
@@ -127,6 +128,7 @@ inline void ref_axpy(int64_t                 n,
 {
     cblas_zaxpy(n, &alpha, x, incx, y, incy);
 }
+#endif
 
 // copy
 template <typename T>

From 755ad379056d470535b19a66d4fab84bb3f20402 Mon Sep 17 00:00:00 2001
From: Torre Zuk <Torre.Zuk@amd.com>
Date: Thu, 7 May 2026 16:19:17 -0600
Subject: [PATCH 2/2] clang format

---
 projects/rocblas/clients/common/cblas_interface.cpp | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/projects/rocblas/clients/common/cblas_interface.cpp b/projects/rocblas/clients/common/cblas_interface.cpp
index 00e7bba0262..e79e1723da1 100644
--- a/projects/rocblas/clients/common/cblas_interface.cpp
+++ b/projects/rocblas/clients/common/cblas_interface.cpp
@@ -129,12 +129,12 @@ template void ref_axpy<rocblas_double_complex>(int64_t                 n,
                                                int64_t                 incy);
 
 template <>
-void ref_axpy<rocblas_bfloat16>(int64_t            n,
-                                rocblas_bfloat16   alpha,
-                                rocblas_bfloat16*  x,
-                                int64_t            incx,
-                                rocblas_bfloat16*  y,
-                                int64_t            incy)
+void ref_axpy<rocblas_bfloat16>(int64_t           n,
+                                rocblas_bfloat16  alpha,
+                                rocblas_bfloat16* x,
+                                int64_t           incx,
+                                rocblas_bfloat16* y,
+                                int64_t           incy)
 {
     // Handle negative increments
     int64_t abs_incx = incx < 0 ? -incx : incx;