diff --git a/ggml/src/ggml-cpu/arch-fallback.h b/ggml/src/ggml-cpu/arch-fallback.h
index 019fd056c8..6b36db2a36 100644
--- a/ggml/src/ggml-cpu/arch-fallback.h
+++ b/ggml/src/ggml-cpu/arch-fallback.h
@@ -37,8 +37,6 @@
 #define ggml_quantize_mat_q8_0_4x8_generic ggml_quantize_mat_q8_0_4x8
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
 #define ggml_quantize_mat_q8_K_4x8_generic ggml_quantize_mat_q8_K_4x8
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_0_8x8_q8_0_generic ggml_gemv_q4_0_8x8_q8_0
@@ -78,33 +76,15 @@
 #define ggml_gemv_iq4_nl_8x8_q8_0_generic ggml_gemv_iq4_nl_8x8_q8_0
 #define ggml_gemv_mxfp4_8x8_q8_0_generic ggml_gemv_mxfp4_8x8_q8_0
 #define ggml_gemv_q2_K_8x8_q8_K_generic ggml_gemv_q2_K_8x8_q8_K
-#define ggml_gemv_f16_1x16_f16_generic ggml_gemv_f16_1x16_f16
-#define ggml_gemv_f16_1x32_f16_generic ggml_gemv_f16_1x32_f16
-#define ggml_gemv_f16_1x64_f16_generic ggml_gemv_f16_1x64_f16
-#define ggml_gemv_f16_1x128_f16_generic ggml_gemv_f16_1x128_f16
-#define ggml_gemv_f32_1x16_f32_generic ggml_gemv_f32_1x16_f32
-#define ggml_gemv_f32_1x32_f32_generic ggml_gemv_f32_1x32_f32
-#define ggml_gemv_f32_1x64_f32_generic ggml_gemv_f32_1x64_f32
-#define ggml_gemv_f32_1x128_f32_generic ggml_gemv_f32_1x128_f32
 #define ggml_gemm_iq4_nl_8x8_q8_0_generic ggml_gemm_iq4_nl_8x8_q8_0
 #define ggml_gemm_mxfp4_8x8_q8_0_generic ggml_gemm_mxfp4_8x8_q8_0
 #define ggml_gemm_q2_K_8x8_q8_K_generic ggml_gemm_q2_K_8x8_q8_K
-#define ggml_gemm_f16_7x1x16_f16_generic ggml_gemm_f16_7x1x16_f16
-#define ggml_gemm_f16_7x1x32_f16_generic ggml_gemm_f16_7x1x32_f16
-#define ggml_gemm_f16_7x1x64_f16_generic ggml_gemm_f16_7x1x64_f16
-#define ggml_gemm_f16_7x1x128_f16_generic ggml_gemm_f16_7x1x128_f16
-#define ggml_gemm_f32_7x1x16_f32_generic ggml_gemm_f32_7x1x16_f32
-#define ggml_gemm_f32_7x1x32_f32_generic ggml_gemm_f32_7x1x32_f32
-#define ggml_gemm_f32_7x1x64_f32_generic ggml_gemm_f32_7x1x64_f32
-#define ggml_gemm_f32_7x1x128_f32_generic ggml_gemm_f32_7x1x128_f32
 #elif defined(__x86_64__) || defined(__i386__) || defined(_M_IX86) || defined(_M_X64)
 // quants.c
 #define ggml_vec_dot_nvfp4_q8_0_generic ggml_vec_dot_nvfp4_q8_0
 // repack.cpp
 #define ggml_quantize_mat_q8_0_4x4_generic ggml_quantize_mat_q8_0_4x4
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_K_8x4_q8_K_generic ggml_gemv_q4_K_8x4_q8_K
@@ -140,8 +120,6 @@
 #define ggml_quantize_mat_q8_0_4x8_generic ggml_quantize_mat_q8_0_4x8
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
 #define ggml_quantize_mat_q8_K_4x8_generic ggml_quantize_mat_q8_K_4x8
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_0_8x8_q8_0_generic ggml_gemv_q4_0_8x8_q8_0
@@ -187,8 +165,6 @@
 #define ggml_quantize_mat_q8_0_4x8_generic ggml_quantize_mat_q8_0_4x8
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
 #define ggml_quantize_mat_q8_K_4x8_generic ggml_quantize_mat_q8_K_4x8
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_0_8x8_q8_0_generic ggml_gemv_q4_0_8x8_q8_0
@@ -281,8 +257,6 @@
 #define ggml_quantize_mat_q8_0_4x8_generic ggml_quantize_mat_q8_0_4x8
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
 #define ggml_quantize_mat_q8_K_4x8_generic ggml_quantize_mat_q8_K_4x8
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_0_8x8_q8_0_generic ggml_gemv_q4_0_8x8_q8_0
@@ -336,8 +310,6 @@
 #define ggml_quantize_mat_q8_0_4x8_generic ggml_quantize_mat_q8_0_4x8
 #define ggml_quantize_mat_q8_K_4x4_generic ggml_quantize_mat_q8_K_4x4
 #define ggml_quantize_mat_q8_K_4x8_generic ggml_quantize_mat_q8_K_4x8
-#define ggml_repack_mat_f16_7x1_generic ggml_repack_mat_f16_7x1
-#define ggml_repack_mat_f32_7x1_generic ggml_repack_mat_f32_7x1
 #define ggml_gemv_q4_0_4x4_q8_0_generic ggml_gemv_q4_0_4x4_q8_0
 #define ggml_gemv_q4_0_4x8_q8_0_generic ggml_gemv_q4_0_4x8_q8_0
 #define ggml_gemv_q4_0_8x8_q8_0_generic ggml_gemv_q4_0_8x8_q8_0
diff --git a/ggml/src/ggml-cpu/arch/riscv/repack.cpp b/ggml/src/ggml-cpu/arch/riscv/repack.cpp
index 6058c9b416..388076f9c0 100644
--- a/ggml/src/ggml-cpu/arch/riscv/repack.cpp
+++ b/ggml/src/ggml-cpu/arch/riscv/repack.cpp
@@ -665,6 +665,97 @@ void ggml_gemv_q2_K_16x1_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const v
         __riscv_vse32_v_f32m2(s + col_tile, v_sumf, vl);
     }
 }
+
+
+template<int ncols_interleaved>
+static inline void ggml_gemv_f16_1xM_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    GGML_UNUSED(bs);
+
+    const int nb = n / 1;
+
+    assert (nr == 1);
+    assert(n % 1 == 0);
+    assert(nc % ncols_interleaved == 0);
+
+    const _Float16 * a_ptr = (const _Float16 *) vy;
+    for (int x = 0; x < nc / ncols_interleaved; x++) {
+        const block_f16<ncols_interleaved, 1> * b_ptr = (const block_f16<ncols_interleaved, 1> *) vx + (x * nb);
+
+        // Accumulators
+        vfloat32m4_t sumf_0 = __riscv_vfmv_v_f_f32m4(0.0f, ncols_interleaved);
+
+        for (int l = 0; l < nb; l++) {
+            vfloat16m2_t b_0 = __riscv_vle16_v_f16m2((const _Float16 *)&b_ptr[l].d[0], ncols_interleaved);
+
+            sumf_0 = __riscv_vfwmacc_vf_f32m4(sumf_0, *(const _Float16*)(&a_ptr[l]), b_0, ncols_interleaved);
+        }
+
+        __riscv_vse32_v_f32m4(&s[x * ncols_interleaved], sumf_0, ncols_interleaved);
+    }
+
+    return;
+}
+
+void ggml_gemv_f16_1x16_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f16_1xM_f16<16>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f16_1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f16_1xM_f16<32>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f16_1x64_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f16_1xM_f16<64>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f16_1x128_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f16_1xM_f16<128>(n, s, bs, vx, vy, nr, nc);
+}
+
+template<int ncols_interleaved>
+static inline void ggml_gemv_f32_1xM_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    GGML_UNUSED(bs);
+
+    const int nb = n / 1;
+
+    assert (nr == 1);
+    assert(n % 1 == 0);
+    assert(nc % ncols_interleaved == 0);
+
+    const float * a_ptr = (const float *) vy;
+    for (int x = 0; x < nc / ncols_interleaved; x++) {
+        const block_f32<ncols_interleaved, 1> * b_ptr = (const block_f32<ncols_interleaved, 1> *) vx + (x * nb);
+
+        // Accumulators
+        vfloat32m4_t sumf_0 = __riscv_vfmv_v_f_f32m4(0.0f, ncols_interleaved);
+
+        for (int l = 0; l < nb; l++) {
+            vfloat32m4_t b_0 = __riscv_vle32_v_f32m4((const float *)&b_ptr[l].d[0], ncols_interleaved);
+
+            sumf_0 = __riscv_vfmacc_vf_f32m4(sumf_0, *(const float*)(&a_ptr[l]), b_0, ncols_interleaved);
+        }
+
+        __riscv_vse32_v_f32m4(&s[x * ncols_interleaved], sumf_0, ncols_interleaved);
+    }
+
+    return;
+}
+
+void ggml_gemv_f32_1x16_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f32_1xM_f32<16>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f32_1x32_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f32_1xM_f32<32>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f32_1x64_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f32_1xM_f32<64>(n, s, bs, vx, vy, nr, nc);
+}
+
+void ggml_gemv_f32_1x128_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
+    ggml_gemv_f32_1xM_f32<128>(n, s, bs, vx, vy, nr, nc);
+}
 #endif
 
 void ggml_gemm_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
@@ -1700,125 +1791,7 @@ void ggml_gemm_q2_K_16x1_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const v
         }
     }
 }
-#endif
 
-template<int ncols_interleaved>
-static inline void ggml_gemv_f16_1xM_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-    const int nb = n / 1;
-
-    assert (nr == 1);
-    assert(n % 1 == 0);
-    assert(nc % ncols_interleaved == 0);
-
-    const _Float16 * a_ptr = (const _Float16 *) vy;
-    for (int x = 0; x < nc / ncols_interleaved; x++) {
-        const block_f16<ncols_interleaved, 1> * b_ptr = (const block_f16<ncols_interleaved, 1> *) vx + (x * nb);
-
-        // Accumulators
-        vfloat32m4_t sumf_0 = __riscv_vfmv_v_f_f32m4(0.0f, ncols_interleaved);
-
-        for (int l = 0; l < nb; l++) {
-            vfloat16m2_t b_0 = __riscv_vle16_v_f16m2((const _Float16 *)&b_ptr[l].d[0], ncols_interleaved);
-
-            sumf_0 = __riscv_vfwmacc_vf_f32m4(sumf_0, *(const _Float16*)(&a_ptr[l]), b_0, ncols_interleaved);
-        }
-
-        __riscv_vse32_v_f32m4(&s[x * ncols_interleaved], sumf_0, ncols_interleaved);
-    }
-
-    return;
-}
-
-void ggml_gemv_f16_1x16_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f16_1xM_f16<16>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f16_1x16_f16_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f16_1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f16_1xM_f16<32>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f16_1x32_f16_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f16_1x64_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f16_1xM_f16<64>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f16_1x64_f16_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f16_1x128_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f16_1xM_f16<128>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f16_1x128_f16_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-template<int ncols_interleaved>
-static inline void ggml_gemv_f32_1xM_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-    const int nb = n / 1;
-
-    assert (nr == 1);
-    assert(n % 1 == 0);
-    assert(nc % ncols_interleaved == 0);
-
-    const float * a_ptr = (const float *) vy;
-    for (int x = 0; x < nc / ncols_interleaved; x++) {
-        const block_f32<ncols_interleaved, 1> * b_ptr = (const block_f32<ncols_interleaved, 1> *) vx + (x * nb);
-
-        // Accumulators
-        vfloat32m4_t sumf_0 = __riscv_vfmv_v_f_f32m4(0.0f, ncols_interleaved);
-
-        for (int l = 0; l < nb; l++) {
-            vfloat32m4_t b_0 = __riscv_vle32_v_f32m4((const float *)&b_ptr[l].d[0], ncols_interleaved);
-
-            sumf_0 = __riscv_vfmacc_vf_f32m4(sumf_0, *(const float*)(&a_ptr[l]), b_0, ncols_interleaved);
-        }
-
-        __riscv_vse32_v_f32m4(&s[x * ncols_interleaved], sumf_0, ncols_interleaved);
-    }
-
-    return;
-}
-
-void ggml_gemv_f32_1x16_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f32_1xM_f32<16>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f32_1x16_f32_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f32_1x32_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f32_1xM_f32<32>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f32_1x32_f32_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f32_1x64_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f32_1xM_f32<64>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f32_1x64_f32_generic(n, s, bs, vx, vy, nr, nc);
-}
-
-void ggml_gemv_f32_1x128_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
-    ggml_gemv_f32_1xM_f32<128>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemv_f32_1x128_f32_generic(n, s, bs, vx, vy, nr, nc);
-}
 
 template<int ncols_interleaved>
 static inline void ggml_gemm_f16_7x1xM_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
@@ -1867,35 +1840,19 @@ static inline void ggml_gemm_f16_7x1xM_f16(int n, float * GGML_RESTRICT s, size_
 }
 
 void ggml_gemm_f16_7x1x16_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f16_7x1xM_f16<16>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f16_7x1x16_f16_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f16_7x1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f16_7x1xM_f16<32>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f16_7x1x32_f16_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f16_7x1x64_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f16_7x1xM_f16<64>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f16_7x1x64_f16_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f16_7x1x128_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f16_7x1xM_f16<128>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f16_7x1x128_f16_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 template<int ncols_interleaved>
@@ -1945,33 +1902,18 @@ static inline void ggml_gemm_f32_7x1xM_f32(int n, float * GGML_RESTRICT s, size_
 }
 
 void ggml_gemm_f32_7x1x16_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f32_7x1xM_f32<16>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f32_7x1x16_f32_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f32_7x1x32_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f32_7x1xM_f32<32>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f32_7x1x32_f32_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f32_7x1x64_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f32_7x1xM_f32<64>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f32_7x1x64_f32_generic(n, s, bs, vx, vy, nr, nc);
 }
 
 void ggml_gemm_f32_7x1x128_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
-#if defined __riscv_v_intrinsic
     ggml_gemm_f32_7x1xM_f32<128>(n, s, bs, vx, vy, nr, nc);
-    return;
-#endif
-    ggml_gemm_f32_7x1x128_f32_generic(n, s, bs, vx, vy, nr, nc);
 }
+#endif
diff --git a/ggml/src/ggml-cpu/repack.cpp b/ggml/src/ggml-cpu/repack.cpp
index 3f720a82ef..9a17c5cd65 100644
--- a/ggml/src/ggml-cpu/repack.cpp
+++ b/ggml/src/ggml-cpu/repack.cpp
@@ -164,6 +164,14 @@ void ggml_quantize_mat_q8_K_4x1_generic(const float * GGML_RESTRICT x, void * GG
         }
     }
 }
+
+void ggml_repack_mat_f16_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
+    ggml_repack_mat_f16_NxK_generic<7, 1>(x, vy, k);
+}
+
+void ggml_repack_mat_f32_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
+    ggml_repack_mat_f32_NxK_generic<7, 1>(x, vy, k);
+}
 #endif
 
 void ggml_quantize_mat_q8_0_4x4_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
@@ -344,16 +352,6 @@ void ggml_quantize_mat_q8_K_4x8_generic(const float * GGML_RESTRICT x, void * GG
     }
 }
 
-#if defined __riscv_zvfh
-void ggml_repack_mat_f16_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
-    ggml_repack_mat_f16_NxK_generic<7, 1>(x, vy, k);
-}
-
-void ggml_repack_mat_f32_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) {
-    ggml_repack_mat_f32_NxK_generic<7, 1>(x, vy, k);
-}
-#endif
-
 } // extern "C"
 
 template <int64_t NB_ROWS, int64_t INTER_SIZE, ggml_type PARAM_TYPE>
@@ -384,20 +382,18 @@ template <> void ggml_repack_mat_t<4, 8, GGML_TYPE_Q8_K>(const float * GGML_REST
 }
 
 #if defined __riscv_zvfh
-template <> void ggml_quantize_mat_t<1, GGML_TYPE_Q8_0>(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row) {
+template <> void ggml_repack_mat_t<4, 1, GGML_TYPE_Q8_0>(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row) {
     assert(nrow == 4);
     UNUSED(nrow);
     ggml_quantize_mat_q8_0_4x1(x, vy, n_per_row);
 }
 
-template <> void ggml_quantize_mat_t<1, GGML_TYPE_Q8_K>(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row) {
+template <> void ggml_repack_mat_t<4, 1, GGML_TYPE_Q8_K>(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row) {
     assert(nrow == 4);
     UNUSED(nrow);
     ggml_quantize_mat_q8_K_4x1(x, vy, n_per_row);
 }
-#endif
 
-#if defined __riscv_zvfh
 template <> void ggml_repack_mat_t<7, 1, GGML_TYPE_F16>(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row) {
     assert(nrow == 7);
     UNUSED(nrow);
@@ -1845,9 +1841,7 @@ void ggml_gemv_q2_K_16x1_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs,
         }
     }
 }
-#endif
 
-#if defined __riscv_zvfh
 void ggml_gemv_f16_1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
     ggml_gemv_f16_KxM_f16_generic<1, 16>(n, s, bs, vx, vy, nr, nc);
 }
@@ -2944,9 +2938,7 @@ void ggml_gemm_q2_K_16x1_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs,
         }
     }
 }
-#endif
 
-#if defined __riscv_zvfh
 void ggml_gemm_f16_7x1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) {
     ggml_gemm_f16_NxKxM_f16_generic<7, 1, 16>(n, s, bs, vx, vy, nr, nc);
 }
@@ -4286,9 +4278,7 @@ template <> int repack<block_q8_0, 1, 16>(struct ggml_tensor * t, const void * d
 template <> int repack<block_q2_K, 1, 16>(struct ggml_tensor * t, const void * data, size_t data_size) {
     return repack_q2_K_to_q2_K_16_bl(t, 1, data, data_size);
 }
-#endif
 
-#if defined __riscv_zvfh
 template <> int repack<ggml_half, 1, 16>(struct ggml_tensor * t, const void * data, size_t data_size) {
     return repack_f16_to_f16_MxK_bl<16, 1>(t, data, data_size);
 }
@@ -4411,9 +4401,7 @@ template <> void gemv<block_q8_0, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_
 template <> void gemv<block_q2_K, 1, 16, GGML_TYPE_Q8_K>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemv_q2_K_16x1_q8_K(n, s, bs, vx, vy, nr, nc);
 }
-#endif
 
-#if defined __riscv_zvfh
 template <> void gemv<ggml_half, 1, 16, GGML_TYPE_F16>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemv_f16_1x16_f16(n, s, bs, vx, vy, nr, nc);
 }
@@ -4523,28 +4511,26 @@ template <> void gemm<block_q8_0, 4, 8, 4, GGML_TYPE_Q8_0>(int n, float * s, siz
 }
 
 #if defined __riscv_zvfh
-template <> void gemm<block_q4_0, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
+template <> void gemm<block_q4_0, 4, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_q4_0_16x1_q8_0(n, s, bs, vx, vy, nr, nc);
 }
 
-template <> void gemm<block_q4_K, 1, 16, GGML_TYPE_Q8_K>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
+template <> void gemm<block_q4_K, 4, 1, 16, GGML_TYPE_Q8_K>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_q4_K_16x1_q8_K(n, s, bs, vx, vy, nr, nc);
 }
 
-template <> void gemm<block_iq4_nl, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
+template <> void gemm<block_iq4_nl, 4, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_iq4_nl_16x1_q8_0(n, s, bs, vx, vy, nr, nc);
 }
 
-template <> void gemm<block_q8_0, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
+template <> void gemm<block_q8_0, 4, 1, 16, GGML_TYPE_Q8_0>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_q8_0_16x1_q8_0(n, s, bs, vx, vy, nr, nc);
 }
 
-template <> void gemm<block_q2_K, 1, 16, GGML_TYPE_Q8_K>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
+template <> void gemm<block_q2_K, 4, 1, 16, GGML_TYPE_Q8_K>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_q2_K_16x1_q8_K(n, s, bs, vx, vy, nr, nc);
 }
-#endif
 
-#if defined __riscv_zvfh
 template <> void gemm<ggml_half, 7, 1, 16, GGML_TYPE_F16>(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) {
     ggml_gemm_f16_7x1x16_f16(n, s, bs, vx, vy, nr, nc);
 }
@@ -4991,23 +4977,19 @@ static const ggml::cpu::tensor_traits * ggml_repack_get_optimal_repack_type(cons
     // These implement outer-product style matrix multiplication kernels with
     // an interleave of 1.
 #if defined __riscv_zvfh
-    static const ggml::cpu::repack::tensor_traits<block_q4_0, 1, 16, GGML_TYPE_Q8_0> q4_0_16x1_q8_0;
-    static const ggml::cpu::repack::tensor_traits<block_q4_K, 1, 16, GGML_TYPE_Q8_K> q4_K_16x1_q8_K;
-    static const ggml::cpu::repack::tensor_traits<block_iq4_nl, 1, 16, GGML_TYPE_Q8_0> iq4_nl_16x1_q8_0;
-    static const ggml::cpu::repack::tensor_traits<block_q8_0, 1, 16, GGML_TYPE_Q8_0> q8_0_16x1_q8_0;
-    static const ggml::cpu::repack::tensor_traits<block_q2_K, 1, 16, GGML_TYPE_Q8_K> q2_K_16x1_q8_K;
-#endif
+    static const ggml::cpu::repack::tensor_traits<block_q4_0, 4, 1, 16, GGML_TYPE_Q8_0> q4_0_16x1_q8_0;
+    static const ggml::cpu::repack::tensor_traits<block_q4_K, 4, 1, 16, GGML_TYPE_Q8_K> q4_K_16x1_q8_K;
+    static const ggml::cpu::repack::tensor_traits<block_iq4_nl, 4, 1, 16, GGML_TYPE_Q8_0> iq4_nl_16x1_q8_0;
+    static const ggml::cpu::repack::tensor_traits<block_q8_0, 4, 1, 16, GGML_TYPE_Q8_0> q8_0_16x1_q8_0;
+    static const ggml::cpu::repack::tensor_traits<block_q2_K, 4, 1, 16, GGML_TYPE_Q8_K> q2_K_16x1_q8_K;
 
     // instance for F16
-#if defined __riscv_zvfh
     static const ggml::cpu::repack::tensor_traits<ggml_half, 7, 1, 16, GGML_TYPE_F16>  f16_7x16x1_f16;
     static const ggml::cpu::repack::tensor_traits<ggml_half, 7, 1, 32, GGML_TYPE_F16>  f16_7x32x1_f16;
     static const ggml::cpu::repack::tensor_traits<ggml_half, 7, 1, 64, GGML_TYPE_F16>  f16_7x64x1_f16;
     static const ggml::cpu::repack::tensor_traits<ggml_half, 7, 1, 128, GGML_TYPE_F16> f16_7x128x1_f16;
-#endif
 
     // instance for F32
-#if defined __riscv_zvfh
     static const ggml::cpu::repack::tensor_traits<float, 7, 1, 16, GGML_TYPE_F32> f32_7x16x1_f32;
     static const ggml::cpu::repack::tensor_traits<float, 7, 1, 32, GGML_TYPE_F32> f32_7x32x1_f32;
     static const ggml::cpu::repack::tensor_traits<float, 7, 1, 64, GGML_TYPE_F32> f32_7x64x1_f32;
diff --git a/ggml/src/ggml-cpu/repack.h b/ggml/src/ggml-cpu/repack.h
index 16d5a2f313..fcaba7b63e 100644
--- a/ggml/src/ggml-cpu/repack.h
+++ b/ggml/src/ggml-cpu/repack.h
@@ -149,6 +149,7 @@ struct block_f32 {
 
 using block_f32_32x1 = block_f32<32, 1>;
 using block_f32_7x1  = block_f32<7, 1>;
+using block_f32_4x1  = block_f32<4, 1>;
 
 #if defined(__cplusplus)
 extern "C" {
@@ -190,6 +191,8 @@ void ggml_gemm_mxfp4_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const v
 void ggml_gemm_mxfp4_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q8_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q8_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+
+// RISC-V
 #if defined __riscv_zvfh
 void ggml_quantize_mat_q8_0_4x1(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
 void ggml_quantize_mat_q8_K_4x1(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
@@ -203,6 +206,28 @@ void ggml_gemm_q4_K_16x1_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const v
 void ggml_gemm_iq4_nl_16x1_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q8_0_16x1_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q2_K_16x1_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+
+// FP16
+void ggml_repack_mat_f16_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
+void ggml_gemv_f16_1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f16_1x32_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f16_1x64_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f16_1x128_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f16_7x1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f16_7x1x32_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f16_7x1x64_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f16_7x1x128_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+
+// FP32
+void ggml_repack_mat_f32_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
+void ggml_gemv_f32_1x16_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f32_1x32_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f32_1x64_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemv_f32_1x128_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f32_7x1x16_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f32_7x1x32_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f32_7x1x64_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+void ggml_gemm_f32_7x1x128_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 #endif
 
 // Native implementations
@@ -242,6 +267,8 @@ void ggml_gemm_mxfp4_4x4_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs,
 void ggml_gemm_mxfp4_8x8_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q8_0_4x4_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q8_0_4x8_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+
+// RISC-V
 #if defined __riscv_zvfh
 void ggml_quantize_mat_q8_0_4x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
 void ggml_quantize_mat_q8_K_4x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
@@ -255,42 +282,19 @@ void ggml_gemm_q4_K_16x1_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs,
 void ggml_gemm_q8_0_16x1_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_q2_K_16x1_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_iq4_nl_16x1_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-#endif
 
 // FP16
-void ggml_repack_mat_f16_4x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
-void ggml_repack_mat_f16_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
-void ggml_gemv_f16_1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f16_1x32_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f16_1x64_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f16_1x128_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_4x1x32_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_7x1x16_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_7x1x32_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_7x1x64_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_7x1x128_f16_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_repack_mat_f16_4x1(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
 void ggml_repack_mat_f16_7x1(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
 void ggml_gemv_f16_1x16_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemv_f16_1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemv_f16_1x64_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemv_f16_1x128_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f16_4x1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f16_7x1x16_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f16_7x1x32_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f16_7x1x64_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f16_7x1x128_f16(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 
 // FP32
-void ggml_repack_mat_f32_7x1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
-void ggml_gemv_f32_1x16_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f32_1x32_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f32_1x64_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemv_f32_1x128_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f32_7x1x16_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f32_7x1x32_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f32_7x1x64_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
-void ggml_gemm_f32_7x1x128_f32_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_repack_mat_f32_7x1(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
 void ggml_gemv_f32_1x16_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemv_f32_1x32_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
@@ -300,6 +304,7 @@ void ggml_gemm_f32_7x1x16_f32(int n, float * GGML_RESTRICT s, size_t bs, const v
 void ggml_gemm_f32_7x1x32_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f32_7x1x64_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
 void ggml_gemm_f32_7x1x128_f32(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
+#endif
 
 #if defined(__cplusplus)
 } // extern "C"