diff --git a/kernels/zen/3/sup/bli_gemmsup_rd_zen_asm_s6x16.c b/kernels/zen/3/sup/bli_gemmsup_rd_zen_asm_s6x16.c
index 96bc92749..c309c8c0c 100644
--- a/kernels/zen/3/sup/bli_gemmsup_rd_zen_asm_s6x16.c
+++ b/kernels/zen/3/sup/bli_gemmsup_rd_zen_asm_s6x16.c
@@ -3,7 +3,7 @@
    An object-based framework for developing high-performance BLAS-like
    libraries.
    Copyright (C) 2014, The University of Texas at Austin
-   Copyright (C) 2020, Advanced Micro Devices, Inc.
+   Copyright (C) 2020 - 2022 , Advanced Micro Devices, Inc.
    Redistribution and use in source and binary forms, with or without
    modification, are permitted provided that the following conditions are
    met:
@@ -516,7 +516,8 @@ void bli_sgemmsup_rd_zen_asm_1x16
     je(.SBETAZERO)                     // if ZF = 1, jump to beta == 0 case
 
     label(.SROWSTORED)    
-    vfmadd231ps(mem(rcx), ymm3, ymm4)
+    vmovups(mem(rcx), xmm0)
+    vfmadd231ps(xmm0, xmm3, xmm4)
     vmovups(xmm4, mem(rcx))
     jmp(.SDONE)                        // jump to end.
 
diff --git a/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16.c b/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16.c
index 752a0a01c..507ff5a71 100644
--- a/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16.c
+++ b/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16.c
@@ -5,7 +5,7 @@
    libraries.
 
    Copyright (C) 2014, The University of Texas at Austin
-   Copyright (C) 2020, Advanced Micro Devices, Inc.
+   Copyright (C) 2020-2022, Advanced Micro Devices, Inc.
 
    Redistribution and use in source and binary forms, with or without
    modification, are permitted provided that the following conditions are
@@ -8048,15 +8048,18 @@ void bli_sgemmsup_rv_zen_asm_3x2
 
     label(.SROWSTORED)
         
-    vfmadd231ps(mem(rcx), xmm3, xmm4)
+    vmovsd(mem(rcx), xmm0)////a0a1
+    vfmadd231ps(xmm0, xmm3, xmm4)
     vmovsd(xmm4, mem(rcx))
     add(rdi, rcx)
 
-    vfmadd231ps(mem(rcx), xmm3, xmm6)
+    vmovsd(mem(rcx), xmm0)////a0a1
+    vfmadd231ps(xmm0, xmm3, xmm6)
     vmovsd(xmm6, mem(rcx))
     add(rdi, rcx)
 
-    vfmadd231ps(mem(rcx), xmm3, xmm8)
+    vmovsd(mem(rcx), xmm0)////a0a1
+    vfmadd231ps(xmm0, xmm3, xmm8)
     vmovsd(xmm8, mem(rcx))
     
     jmp(.SDONE)                        // jump to end.
@@ -8329,11 +8332,13 @@ void bli_sgemmsup_rv_zen_asm_2x2
     label(.SROWSTORED)
         
 
-    vfmadd231ps(mem(rcx), xmm3, xmm4)
+    vmovsd(mem(rcx), xmm0)////a0a1
+    vfmadd231ps(xmm0, xmm3, xmm4)
     vmovsd(xmm4, mem(rcx))
     add(rdi, rcx)
         
-    vfmadd231ps(mem(rcx), xmm3, xmm6)
+    vmovsd(mem(rcx), xmm0)////a0a1
+    vfmadd231ps(xmm0, xmm3, xmm6)
     vmovsd(xmm6, mem(rcx))
     
     jmp(.SDONE)                        // jump to end.
@@ -8577,7 +8582,8 @@ void bli_sgemmsup_rv_zen_asm_1x2
 
     label(.SROWSTORED)
     
-    vfmadd231ps(mem(rcx), xmm3, xmm4)
+    vmovsd(mem(rcx), xmm0)
+    vfmadd231ps(xmm0, xmm3, xmm4)
     vmovsd(xmm4, mem(rcx))
 
     jmp(.SDONE)                        // jump to end.
diff --git a/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16m.c b/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16m.c
index 41dbbd699..e6ecd47f4 100644
--- a/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16m.c
+++ b/kernels/zen/3/sup/bli_gemmsup_rv_zen_asm_s6x16m.c
@@ -5,7 +5,7 @@
    libraries.
 
    Copyright (C) 2014, The University of Texas at Austin
-   Copyright (C) 2020, Advanced Micro Devices, Inc.
+   Copyright (C) 2020-2022, Advanced Micro Devices, Inc.
 
    Redistribution and use in source and binary forms, with or without
    modification, are permitted provided that the following conditions are
@@ -2231,22 +2231,28 @@ void bli_sgemmsup_rv_zen_asm_6x2m
 
 	label(.SROWSTORED)
 		
-	vfmadd231ps(mem(rcx), xmm3, xmm4)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm4)
 	vmovlpd(xmm4, mem(rcx))
 	add(rdi, rcx)	
-	vfmadd231ps(mem(rcx), xmm3, xmm6)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm6)
 	vmovlpd(xmm6, mem(rcx))
 	add(rdi, rcx)	
-	vfmadd231ps(mem(rcx), xmm3, xmm8)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm8)
 	vmovlpd(xmm8, mem(rcx))
 	add(rdi, rcx)	
-	vfmadd231ps(mem(rcx), xmm3, xmm10)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm10)
 	vmovlpd(xmm10, mem(rcx))
 	add(rdi, rcx)	
-	vfmadd231ps(mem(rcx), xmm3, xmm12)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm12)
 	vmovlpd(xmm12, mem(rcx))
 	add(rdi, rcx)	
-	vfmadd231ps(mem(rcx), xmm3, xmm14)
+	vmovsd(mem(rcx), xmm0)
+	vfmadd231ps(xmm0, xmm3, xmm14)
 	vmovlpd(xmm14, mem(rcx))
 		
 	jmp(.SDONE)                        // jump to end.