diff --git a/src/cpu/cpu_convolution_list.cpp b/src/cpu/cpu_convolution_list.cpp
index 573cee77702..5cdef8852d6 100644
--- a/src/cpu/cpu_convolution_list.cpp
+++ b/src/cpu/cpu_convolution_list.cpp
@@ -172,6 +172,40 @@ const std::map<pk_dt_impl_key_t, std::vector<impl_list_item_t>> &impl_list_map()
             CPU_INSTANCE(ref_fused_convolution_fwd_t)
             nullptr,
         }},
+        {{forward, f16, f16, f32}, {
+            CPU_INSTANCE_AVX512(brdgmm_dw_convolution_fwd_t)
+            CPU_INSTANCE_X64(ip_convolution_fwd_t)
+            CPU_INSTANCE_AMX(brgemm_1x1_convolution_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AMX(brgemm_convolution_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AMX(brgemm_convolution_fwd_t, avx512_core_amx_fp16, true)
+            CPU_INSTANCE_AVX512(brgemm_1x1_convolution_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX512(brgemm_convolution_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX512(brgemm_convolution_fwd_t, avx512_core_fp16, true)
+            CPU_INSTANCE_AVX2(brgemm_1x1_convolution_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE_AVX2(brgemm_convolution_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE_AVX2(brgemm_convolution_fwd_t, avx2_vnni_2, true)
+            CPU_INSTANCE(ref_convolution_fwd_t)
+            nullptr,
+        }},
+        {{forward, f16, f16, f16}, {
+            CPU_INSTANCE_AVX512(brdgmm_dw_convolution_fwd_t)
+            CPU_INSTANCE_X64(ip_convolution_fwd_t)
+            CPU_INSTANCE_AMX(brgemm_1x1_convolution_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AMX(brgemm_convolution_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AMX(brgemm_convolution_fwd_t, avx512_core_amx_fp16, true)
+            CPU_INSTANCE_AVX512(brgemm_1x1_convolution_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX512(brgemm_convolution_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX512(brgemm_convolution_fwd_t, avx512_core_fp16, true)
+            CPU_INSTANCE_AVX2(brgemm_1x1_convolution_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE_AVX2(brgemm_convolution_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE_AVX2(brgemm_convolution_fwd_t, avx2_vnni_2, true)
+            CPU_INSTANCE_ACL(acl_wino_convolution_fwd_t)
+            CPU_INSTANCE_ACL(acl_indirect_gemm_convolution_fwd_t)
+            CPU_INSTANCE_ACL(acl_gemm_convolution_fwd_t, f16)
+            CPU_INSTANCE(ref_convolution_fwd_t)
+            CPU_INSTANCE(ref_fused_convolution_fwd_t)
+            nullptr,
+        }},
         // BWD_D fp
         {{backward_data, f32, f32, f32}, REG_CONV_P({
             CPU_INSTANCE_X64(ip_convolution_bwd_data_t)
@@ -220,6 +254,22 @@ const std::map<pk_dt_impl_key_t, std::vector<impl_list_item_t>> &impl_list_map()
             CPU_INSTANCE(ref_convolution_bwd_data_t)
             nullptr,
         })},
+        {{backward_data, f32, f16, f16}, REG_BWD_D_PK({
+            CPU_INSTANCE_X64(ip_convolution_bwd_data_t)
+            CPU_INSTANCE_AMX(brgemm_convolution_bwd_strided_t<avx512_core_amx_fp16>)
+            CPU_INSTANCE_AVX512(brgemm_convolution_bwd_t<avx512_core_fp16>)
+            CPU_INSTANCE_AVX2(brgemm_convolution_bwd_t<avx2_vnni_2>)
+            CPU_INSTANCE(ref_convolution_bwd_data_t)
+            nullptr,
+        })},
+        {{backward_data, f16, f16, f16}, REG_BWD_D_PK({
+            CPU_INSTANCE_X64(ip_convolution_bwd_data_t)
+            CPU_INSTANCE_AMX(brgemm_convolution_bwd_strided_t<avx512_core_amx_fp16>)
+            CPU_INSTANCE_AVX512(brgemm_convolution_bwd_t<avx512_core_fp16>)
+            CPU_INSTANCE_AVX2(brgemm_convolution_bwd_t<avx2_vnni_2>)
+            CPU_INSTANCE(ref_convolution_bwd_data_t)
+            nullptr,
+        })},
         // BWD_W fp
         {{backward_weights, f32, f32, f32}, REG_BWD_PK({
             CPU_INSTANCE_X64(ip_convolution_bwd_weights_t)
diff --git a/src/cpu/cpu_deconvolution_list.cpp b/src/cpu/cpu_deconvolution_list.cpp
index ae8867095eb..1eb0c5ec33d 100644
--- a/src/cpu/cpu_deconvolution_list.cpp
+++ b/src/cpu/cpu_deconvolution_list.cpp
@@ -49,8 +49,8 @@ using namespace dnnl::impl::prop_kind;
 const std::map<pk_impl_key_t, std::vector<impl_list_item_t>> &impl_list_map() {
     static const std::map<pk_impl_key_t, std::vector<impl_list_item_t>> the_map = REG_DECONV_P({
         {{forward}, {
-            CPU_INSTANCE_AMX(brgemm_deconvolution_fwd_t<avx512_core_amx_fp16>)
-            CPU_INSTANCE_AMX(brgemm_deconvolution_fwd_t<avx512_core_amx>)
+            CPU_INSTANCE_AMX(brgemm_deconvolution_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AMX(brgemm_deconvolution_fwd_t, avx512_core_amx)
             CPU_INSTANCE_AMX(jit_avx512_core_amx_deconvolution_fwd_t)
             CPU_INSTANCE_AVX512(jit_avx512_core_x8s8s32x_1x1_deconvolution_fwd_t)
             CPU_INSTANCE_AVX512(jit_avx512_core_x8s8s32x_deconvolution_fwd_t)
diff --git a/src/cpu/cpu_inner_product_list.cpp b/src/cpu/cpu_inner_product_list.cpp
index b37cc262c04..2259519eaa2 100644
--- a/src/cpu/cpu_inner_product_list.cpp
+++ b/src/cpu/cpu_inner_product_list.cpp
@@ -68,6 +68,21 @@ const std::map<pk_dt_impl_key_t, std::vector<impl_list_item_t>> &impl_list_map()
             CPU_INSTANCE(ref_inner_product_fwd_t)
             nullptr,
         }},
+        {{forward, f16, f16, f32}, {
+            CPU_INSTANCE_AMX(brgemm_inner_product_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AVX512(brgemm_inner_product_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX2(brgemm_inner_product_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE(ref_inner_product_fwd_t)
+            nullptr,
+        }},
+        {{forward, f16, f16, f16}, {
+            CPU_INSTANCE_AMX(brgemm_inner_product_fwd_t, avx512_core_amx_fp16)
+            CPU_INSTANCE_AVX512(brgemm_inner_product_fwd_t, avx512_core_fp16)
+            CPU_INSTANCE_AVX2(brgemm_inner_product_fwd_t, avx2_vnni_2)
+            CPU_INSTANCE_ACL(acl_inner_product_fwd_t)
+            CPU_INSTANCE(ref_inner_product_fwd_t)
+            nullptr,
+        }},
         {{backward_data, f32, f32, f32}, REG_BWD_PK({
             CPU_INSTANCE_AMX(brgemm_inner_product_bwd_data_t, avx512_core_amx) // bf32
             CPU_INSTANCE_AVX512(brgemm_inner_product_bwd_data_t, avx512_core)
diff --git a/src/cpu/cpu_pooling_list.cpp b/src/cpu/cpu_pooling_list.cpp
index 60c00966e4d..6a8f5230e19 100644
--- a/src/cpu/cpu_pooling_list.cpp
+++ b/src/cpu/cpu_pooling_list.cpp
@@ -54,9 +54,11 @@ const std::map<pk_impl_key_t, std::vector<impl_list_item_t>> &impl_list_map() {
     static const std::map<pk_impl_key_t, std::vector<impl_list_item_t>> the_map = REG_POOLING_P({
         {{forward}, {
             /* fp */
+            CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx512_core_fp16, f16)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx512_core, bf16)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx512_core, f32)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx2_vnni_2, bf16)
+            CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx2_vnni_2, f16)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx2, f32)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, avx, f32)
             CPU_INSTANCE_X64(jit_uni_pooling_fwd_t, sse41, f32)
diff --git a/src/cpu/matmul/cpu_matmul_list.cpp b/src/cpu/matmul/cpu_matmul_list.cpp
index 8279505e305..de544a03e0c 100644
--- a/src/cpu/matmul/cpu_matmul_list.cpp
+++ b/src/cpu/matmul/cpu_matmul_list.cpp
@@ -68,6 +68,7 @@ using namespace dnnl::impl::cpu::matmul;
 // constexpr impl_list_item_t impl_list[] = REG_MATMUL_P({
 const impl_list_item_t impl_list[] = REG_MATMUL_P({
         CPU_INSTANCE_ACL(acl_matmul_t)
+        CPU_INSTANCE_AMX(brgemm_matmul_t, avx512_core_amx_fp16)
         CPU_INSTANCE_AMX(brgemm_matmul_t, avx512_core_amx)
         CPU_INSTANCE_AVX512(brgemm_matmul_t, avx512_core)
         CPU_INSTANCE(gemm_f32_matmul_t)
@@ -78,7 +79,7 @@ const impl_list_item_t impl_list[] = REG_MATMUL_P({
         CPU_INSTANCE_AVX2(brgemm_matmul_t, avx2_vnni_2)
         CPU_INSTANCE_AVX2(brgemm_matmul_t, avx2_vnni)
         CPU_INSTANCE(gemm_x8s8s32x_matmul_t)
-        CPU_INSTANCE_AVX512(brgemm_matmul_t<avx512_core_fp16>)
+        CPU_INSTANCE_AVX512(brgemm_matmul_t, avx512_core_fp16)
         CPU_INSTANCE(ref_matmul_t)
         CPU_INSTANCE(ref_matmul_int8_t)
         // These implementations are enabled only when DNNL_EXPERIMENTAL_SPARSE
diff --git a/src/cpu/reorder/cpu_reorder.cpp b/src/cpu/reorder/cpu_reorder.cpp
index 11a4fea2fd2..44007099608 100644
--- a/src/cpu/reorder/cpu_reorder.cpp
+++ b/src/cpu/reorder/cpu_reorder.cpp
@@ -26,12 +26,14 @@ static const std::map<reorder_impl_key_t, const void *> &
 regular_impl_list_map() {
     static const std::map<reorder_impl_key_t, const void *> the_map = {
             {{f32, bf16, 0}, &regular_f32_bf16_impl_list_map()},
+            {{f32, f16, 0}, &regular_f32_f16_impl_list_map()},
             {{f32, f32, 0}, &regular_f32_f32_impl_list_map()},
             {{f32, s32, 0}, &regular_f32_s32_impl_list_map()},
             {{f32, s8, 0}, &regular_f32_s8_impl_list_map()},
             {{f32, u8, 0}, &regular_f32_u8_impl_list_map()},
             {{f32, bin, 0}, &regular_f32_bin_impl_list_map()},
             {{bf16, data_type::undef, 0}, &regular_bf16_impl_list_map()},
+            {{f16, data_type::undef, 0}, &regular_f16_impl_list_map()},
             {{s32, data_type::undef, 0}, &regular_s32_impl_list_map()},
             {{s8, data_type::undef, 0}, &regular_s8_impl_list_map()},
             {{u8, data_type::undef, 0}, &regular_u8_impl_list_map()},
diff --git a/src/cpu/reorder/cpu_reorder_regular_f16.cpp b/src/cpu/reorder/cpu_reorder_regular_f16.cpp
index b83887ef0df..2392098d07a 100644
--- a/src/cpu/reorder/cpu_reorder_regular_f16.cpp
+++ b/src/cpu/reorder/cpu_reorder_regular_f16.cpp
@@ -27,8 +27,8 @@ const impl_list_map_t &regular_f16_impl_list_map() {
         // f16 ->
         {{f16, data_type::undef, 0}, {
             DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64::brgemm_matmul_matrix_B_reorder_t))
-            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64::jit_blk_reorder_t))
-            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64::jit_uni_reorder_t))
+            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64_jit_blk_reorder_t))
+            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64_jit_uni_reorder_t))
             REG_SR(f16, any, f16, any, fmt_order_any, spec_reference)
             REG_SR(f16, any, f32, any, fmt_order_any, spec_reference)
             REG_SR(f16, any, s8, any, fmt_order_any, spec_reference)
diff --git a/src/cpu/reorder/cpu_reorder_regular_f32_f16.cpp b/src/cpu/reorder/cpu_reorder_regular_f32_f16.cpp
index a5d79f8a79a..02cd9cd7c66 100644
--- a/src/cpu/reorder/cpu_reorder_regular_f32_f16.cpp
+++ b/src/cpu/reorder/cpu_reorder_regular_f32_f16.cpp
@@ -26,8 +26,8 @@ const impl_list_map_t &regular_f32_f16_impl_list_map() {
     static const impl_list_map_t the_map = REG_REORDER_P({
         // f32 -> f16
         {{f32, f16, 0}, {
-            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64::jit_blk_reorder_t))
-            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64::jit_uni_reorder_t))
+            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64_jit_blk_reorder_t))
+            DNNL_X64_ONLY(CPU_REORDER_INSTANCE(x64_jit_uni_reorder_t))
 
             nullptr,
         }},
diff --git a/src/cpu/x64/jit_uni_pool_kernel.cpp b/src/cpu/x64/jit_uni_pool_kernel.cpp
index 6be92a61370..c378e3e8d92 100644
--- a/src/cpu/x64/jit_uni_pool_kernel.cpp
+++ b/src/cpu/x64/jit_uni_pool_kernel.cpp
@@ -1638,6 +1638,7 @@ template struct jit_uni_pool_kernel<avx>;
 template struct jit_uni_pool_kernel<avx2>;
 template struct jit_uni_pool_kernel<avx2_vnni_2>;
 template struct jit_uni_pool_kernel<avx512_core>;
+template struct jit_uni_pool_kernel<avx512_core_fp16>;
 
 } // namespace x64
 } // namespace cpu
diff --git a/src/cpu/x64/jit_uni_pooling.cpp b/src/cpu/x64/jit_uni_pooling.cpp
index 977158156d8..ca8967dfec2 100644
--- a/src/cpu/x64/jit_uni_pooling.cpp
+++ b/src/cpu/x64/jit_uni_pooling.cpp
@@ -1279,6 +1279,7 @@ template struct jit_uni_pooling_fwd_t<avx512_core, data_type::f32>;
 template struct jit_uni_pooling_bwd_t<avx512_core, data_type::f32>;
 template struct jit_uni_pooling_fwd_t<avx512_core, data_type::bf16>;
 template struct jit_uni_pooling_bwd_t<avx512_core, data_type::bf16>;
+template struct jit_uni_pooling_fwd_t<avx512_core_fp16, data_type::f16>;
 
 } // namespace x64
 } // namespace cpu