1. add "integrated" in ggml_cuda_device_info for distinguish whether it is Intergrate_gpu or discrete_gpu

2. Adjust the func:"ggml_backend_cuda_device_supports_buft" for this new feature

Author: Yangxiaoz

ggml/src/ggml-cuda/common.cuh

@@ -635,6 +635,7 @@ struct ggml_cuda_device_info {
         int     nsm;                // number of streaming multiprocessors
         size_t  smpb;               // max. shared memory per block
         size_t  smpbo;              // max. shared memory per block (with opt-in)
+        bool    integrated;         // Device is integrated as opposed to discrete
         bool    vmm;                // virtual memory support
         size_t  vmm_granularity;    // granularity of virtual memory
         size_t  total_vram;

ggml/src/ggml-cuda/ggml-cuda.cu

@@ -243,7 +243,7 @@ static ggml_cuda_device_info ggml_cuda_init() {
 
         info.default_tensor_split[id] = total_vram;
         total_vram += prop.totalGlobalMem;
-
+        info.devices[id].integrated= prop.integrated;
         info.devices[id].nsm       = prop.multiProcessorCount;
         info.devices[id].smpb      = prop.sharedMemPerBlock;
         info.devices[id].warp_size = prop.warpSize;
@@ -1065,6 +1065,10 @@ static const char * ggml_backend_cuda_host_buffer_type_name(ggml_backend_buffer_
     GGML_UNUSED(buft);
 }
 
+static bool ggml_backend_buft_is_cuda_host(ggml_backend_buffer_type_t buft) {
+    return buft->iface.get_name == ggml_backend_cuda_host_buffer_type_name;
+}
+
 static void ggml_backend_cuda_host_buffer_free_buffer(ggml_backend_buffer_t buffer) {
     CUDA_CHECK(cudaFreeHost(buffer->context));
 }
@@ -3263,7 +3267,13 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
 }
 
 static bool ggml_backend_cuda_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) {
-    return (ggml_backend_buft_is_cuda(buft) || ggml_backend_buft_is_cuda_split(buft)) && buft->device == dev;
+    ggml_backend_cuda_device_context * dev_ctx = (ggml_backend_cuda_device_context *) dev->context;
+    const int integrated = ggml_cuda_info().devices[dev_ctx->device].integrated;
+    if(integrated){
+        return (ggml_backend_buft_is_cuda(buft) || ggml_backend_buft_is_cuda_split(buft) ||ggml_backend_buft_is_cuda_host(buft)) && buft->device == dev;
+    }else{
+        return (ggml_backend_buft_is_cuda(buft) || ggml_backend_buft_is_cuda_split(buft)) && buft->device == dev;
+    }
 }
 
 static int64_t get_op_batch_size(const ggml_tensor * op) {