ggml-cpu: add RVV implementation for q1_0 x q8_0 vec dot

ggml/src/ggml-cpu/arch-fallback.h

@@ -203,7 +203,6 @@
 #elif defined(__riscv)
 // quants.c
 #define ggml_vec_dot_nvfp4_q8_0_generic ggml_vec_dot_nvfp4_q8_0
-#define ggml_vec_dot_q1_0_q8_0_generic ggml_vec_dot_q1_0_q8_0
 // repack.cpp
 #define ggml_quantize_mat_q8_0_4x1_generic ggml_quantize_mat_q8_0_4x1
 #define ggml_quantize_mat_q8_0_4x4_generic ggml_quantize_mat_q8_0_4x4

ggml/src/ggml-cpu/arch/riscv/quants.c

@@ -274,6 +274,55 @@ void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const voi
 #endif
 }
 
+void ggml_vec_dot_q1_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc) {
+#if defined(__riscv_v)
+    const int qk = QK1_0;
+    const int nb = n / qk;
+    // A single q8_0 block always contains 32 int8 values. With e8,m4 this
+    // maps to vl=32 on any spec-compliant RVV target (minimum VLEN is 64-bit).
+    const size_t vl = __riscv_vsetvl_e8m4(32);
+
+    assert(n % qk == 0);
+    assert(vl == 32);
+    assert(nrc == 1);
+    UNUSED(nrc);
+    UNUSED(bx);
+    UNUSED(by);
+    UNUSED(bs);
+
+    const block_q1_0 * GGML_RESTRICT x = vx;
+    const block_q8_0 * GGML_RESTRICT y = vy;
+    const vint16m1_t v_zero_sum = __riscv_vmv_v_x_i16m1(0, 1);
+
+    float sumf = 0.0f;
+
+    for (int ib = 0; ib < nb; ++ib) {
+        const float d0 = GGML_CPU_FP16_TO_FP32(x[ib].d);
+        float sumi = 0.0f;
+
+        for (int k = 0; k < 4; ++k) {
+            const block_q8_0 * GGML_RESTRICT yb = &y[ib * 4 + k];
+            const float d1 = GGML_CPU_FP16_TO_FP32(yb->d);
+            const uint8_t * GGML_RESTRICT bits = &x[ib].qs[k * 4];
+            vbool2_t m_positive = __riscv_vlm_v_b2(bits, vl);
+            vint8m4_t v_q8 = __riscv_vle8_v_i8m4(yb->qs, vl);
+            vint8m4_t v_q8_neg = __riscv_vsub_vv_i8m4(__riscv_vmv_v_x_i8m4(0, vl), v_q8, vl);
+            vint8m4_t v_q8_signed = __riscv_vmerge_vvm_i8m4(v_q8_neg, v_q8, m_positive, vl);
+            vint16m1_t v_sum = __riscv_vwredsum_vs_i8m4_i16m1(v_q8_signed, v_zero_sum, vl);
+            const int sumi_block = __riscv_vmv_x_s_i16m1_i16(v_sum);
+
+            sumi += d1 * sumi_block;
+        }
+
+        sumf += d0 * sumi;
+    }
+
+    *s = sumf;
+#else
+    ggml_vec_dot_q1_0_q8_0_generic(n, s, bs, vx, bx, vy, by, nrc);
+#endif
+}
+
 void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc) {
 #if defined(__riscv_v)
     const int qk = QK8_1;