diff --git a/samples/99_matrixexperiments/main.cpp b/samples/99_matrixexperiments/main.cpp
index 42b3cbe..044739e 100644
--- a/samples/99_matrixexperiments/main.cpp
+++ b/samples/99_matrixexperiments/main.cpp
@@ -174,7 +174,7 @@ static void go_naive(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -222,7 +222,7 @@ static void go_dpas_rowmajor(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -276,7 +276,7 @@ static void go_dpas_rowmajor_tiled(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -324,7 +324,7 @@ static void go_dpas_vnni(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -378,7 +378,7 @@ static void go_dpas_vnni_tiled(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -426,7 +426,7 @@ static void go_dpas_blockread_rowmajor(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -480,7 +480,7 @@ static void go_dpas_blockread_rowmajor_tiled(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -528,7 +528,7 @@ static void go_dpas_blockread_vnni(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
@@ -582,7 +582,7 @@ static void go_dpas_blockread_vnni_tiled(
         kernel.setArg(2, B);
         kernel.setArg(3, static_cast<cl_int>(K));
 
-        queue.enqueueFillBuffer(C, 0, 0, C_ref.size());
+        queue.enqueueFillBuffer(C, 0, 0, C_ref.size() * sizeof(C_ref[0]));
 
         float best = 999.0f;
         for (int test = 0; test < testIterations; test++) {
diff --git a/samples/99_matrixexperiments/matrix_helpers.cl b/samples/99_matrixexperiments/matrix_helpers.cl
index ef68fa4..04a63c5 100644
--- a/samples/99_matrixexperiments/matrix_helpers.cl
+++ b/samples/99_matrixexperiments/matrix_helpers.cl
@@ -464,19 +464,19 @@ uint8 intel_subgroup_block_read_u32_m8k16(const __global void* base_address, int
     return __builtin_IB_subgroup_block_read_flat_u32_m8k16v1(as_long(base_address), width - 1, height - 1, pitch - 1, coord);
 }
 
-void intel_subgroup_block_write_u32_m1k16v1(__global void* base_address, int width, int height, int pitch, int2 coord, uint data)
+void intel_subgroup_block_write_u32_m1k16(__global void* base_address, int width, int height, int pitch, int2 coord, uint data)
 {
     __builtin_IB_subgroup_block_write_flat_u32_m1k16v1(as_long(base_address), width - 1, height - 1, pitch - 1, coord, data);
 }
-void intel_subgroup_block_write_u32_m2k16v1(__global void* base_address, int width, int height, int pitch, int2 coord, uint2 data)
+void intel_subgroup_block_write_u32_m2k16(__global void* base_address, int width, int height, int pitch, int2 coord, uint2 data)
 {
     __builtin_IB_subgroup_block_write_flat_u32_m2k16v1(as_long(base_address), width - 1, height - 1, pitch - 1, coord, data);
 }
-void intel_subgroup_block_write_u32_m4k16v1(__global void* base_address, int width, int height, int pitch, int2 coord, uint4 data)
+void intel_subgroup_block_write_u32_m4k16(__global void* base_address, int width, int height, int pitch, int2 coord, uint4 data)
 {
     __builtin_IB_subgroup_block_write_flat_u32_m4k16v1(as_long(base_address), width - 1, height - 1, pitch - 1, coord, data);
 }
-void intel_subgroup_block_write_u32_m8k16v1(__global void* base_address, int width, int height, int pitch, int2 coord, uint8 data)
+void intel_subgroup_block_write_u32_m8k16(__global void* base_address, int width, int height, int pitch, int2 coord, uint8 data)
 {
     __builtin_IB_subgroup_block_write_flat_u32_m8k16v1(as_long(base_address), width - 1, height - 1, pitch - 1, coord, data);
 }
diff --git a/samples/99_matrixexperiments/matrix_kernel_tiled.cl b/samples/99_matrixexperiments/matrix_kernel_tiled.cl
index b363b14..a01e0f2 100644
--- a/samples/99_matrixexperiments/matrix_kernel_tiled.cl
+++ b/samples/99_matrixexperiments/matrix_kernel_tiled.cl
@@ -188,7 +188,7 @@ kernel void MM_KERNEL_NAME(bfloat16_dpas_blockread_rowmajor_tiled, 8, 16, MM, NN
 {
     const int tM = 8;
     const int tN = 16;
-    const int M = get_global_size(1) * tM;
+    const int M = get_global_size(1) * tM * MM;
     const int N = get_global_size(0) * NN;
     const int m = get_group_id(1) * tM * MM;
     const int n = get_group_id(0) * tN * NN;
@@ -220,7 +220,7 @@ kernel void MM_KERNEL_NAME(bfloat16_dpas_blockread_rowmajor_tiled, 8, 16, MM, NN
 
     for (int mm = 0; mm < MM; mm++) {
         for (int nn = 0; nn < NN; nn++) {
-            intel_subgroup_block_write_u32_m8k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n + nn * tN, m + mm * tM), as_uint8(sum[mm][nn]));
+            intel_subgroup_block_write_u32_m8k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n + nn * tN, m + mm * tM), as_uint8(sum[mm][nn]));
         }
     }
 }
@@ -230,7 +230,7 @@ kernel void MM_KERNEL_NAME(bfloat16_dpas_blockread_vnni_tiled, 8, 16, MM, NN)(gl
 {
     const int tM = 8;
     const int tN = 16;
-    const int M = get_global_size(1) * tM;
+    const int M = get_global_size(1) * tM * MM;
     const int N = get_global_size(0) * NN;
     const int m = get_group_id(1) * tM * MM;
     const int n = get_group_id(0) * tN * NN;
@@ -262,7 +262,7 @@ kernel void MM_KERNEL_NAME(bfloat16_dpas_blockread_vnni_tiled, 8, 16, MM, NN)(gl
 
     for (int mm = 0; mm < MM; mm++) {
         for (int nn = 0; nn < NN; nn++) {
-            intel_subgroup_block_write_u32_m8k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n + nn * tN, m + mm * tM), as_uint8(sum[mm][nn]));
+            intel_subgroup_block_write_u32_m8k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n + nn * tN, m + mm * tM), as_uint8(sum[mm][nn]));
         }
     }
 }
diff --git a/samples/99_matrixexperiments/matrix_kernels.cl b/samples/99_matrixexperiments/matrix_kernels.cl
index a310db6..efc8f4f 100644
--- a/samples/99_matrixexperiments/matrix_kernels.cl
+++ b/samples/99_matrixexperiments/matrix_kernels.cl
@@ -366,7 +366,7 @@ kernel void bfloat16_dpas_blockread_rowmajor_m1_n16(global float* C, global usho
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m1k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint(sum));
+    intel_subgroup_block_write_u32_m1k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -386,7 +386,7 @@ kernel void bfloat16_dpas_blockread_rowmajor_m2_n16(global float* C, global usho
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m2k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint2(sum));
+    intel_subgroup_block_write_u32_m2k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint2(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -406,7 +406,7 @@ kernel void bfloat16_dpas_blockread_rowmajor_m4_n16(global float* C, global usho
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m4k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint4(sum));
+    intel_subgroup_block_write_u32_m4k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint4(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -426,7 +426,7 @@ kernel void bfloat16_dpas_blockread_rowmajor_m8_n16(global float* C, global usho
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m8k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint8(sum));
+    intel_subgroup_block_write_u32_m8k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint8(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -446,7 +446,7 @@ kernel void bfloat16_dpas_blockread_vnni_m1_n16(global float* C, global ushort*
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m1k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint(sum));
+    intel_subgroup_block_write_u32_m1k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -466,7 +466,7 @@ kernel void bfloat16_dpas_blockread_vnni_m2_n16(global float* C, global ushort*
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m2k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint2(sum));
+    intel_subgroup_block_write_u32_m2k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint2(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -486,7 +486,7 @@ kernel void bfloat16_dpas_blockread_vnni_m4_n16(global float* C, global ushort*
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m4k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint4(sum));
+    intel_subgroup_block_write_u32_m4k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint4(sum));
 }
 
 __attribute__((intel_reqd_sub_group_size(16))) __attribute__((reqd_work_group_size(16, 1, 1)))
@@ -506,7 +506,7 @@ kernel void bfloat16_dpas_blockread_vnni_m8_n16(global float* C, global ushort*
         sum = mat_mul_sg16(aData, bData, sum);
     }
 
-    intel_subgroup_block_write_u32_m8k16v1(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint8(sum));
+    intel_subgroup_block_write_u32_m8k16(C, N * sizeof(float), M, N * sizeof(float), (int2)(n, m), as_uint8(sum));
 }
 
 #endif // cl_intel_subgroup_extended_block_read