Fix batched impl

Author: Aidan

ggml-sycl.cpp

@@ -15272,8 +15272,8 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0,
     sycl_pool_alloc<sycl::half> dst_f16;
     char * dst_t;
 
-    dpct::library_data_t cu_compute_type = dpct::library_data_t::real_float;
-    dpct::library_data_t cu_data_type = dpct::library_data_t::real_float;
+    dpct::library_data_t cu_compute_type = dpct::library_data_t::real_half;
+    dpct::library_data_t cu_data_type = dpct::library_data_t::real_half;
 
     // dst strides
     size_t nbd2 = dst->nb[2];
@@ -15282,16 +15282,16 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0,
     const sycl::half alpha_f16 = 1.0f;
     const sycl::half beta_f16 = 0.0f;
 
-    const float alpha_f32 = 1.0f;
-    const float beta_f32  = 0.0f;
-
-    const void * alpha = &alpha_f32;
-    const void * beta  = &beta_f32;
+    const void * alpha = &alpha_f16;
+    const void * beta  = &beta_f16;
 
     // TODO: Renable (dst->op_params[0] =! GGML_PREC_DEFAULT) pathway
-    // oneMKL open source supports half, half, float, float: datatypes
+    // when oneMKL open source supports half, half, float, float: datatypes
+
+    dst_t = (char *) dst_f16.alloc(ne_dst);
 
-    dst_t = (char *) dst_ddf;
+    nbd2 /= sizeof(float) / sizeof(sycl::half);
+    nbd3 /= sizeof(float) / sizeof(sycl::half);
 
     GGML_ASSERT(ne12 % ne02 == 0);
     GGML_ASSERT(ne13 % ne03 == 0);
@@ -15377,6 +15377,8 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0,
     }
 #endif
 
+    const to_fp32_sycl_t to_fp32_sycl = ggml_get_to_fp32_sycl(GGML_TYPE_F16);
+    to_fp32_sycl(dst_f16.get(), dst_ddf, ne_dst, main_stream);
 }
 catch (sycl::exception const &exc) {
   std::cerr << exc.what() << "Exception caught at file:" << __FILE__