Update GGML_HIP_UMA (#473)

Add UMA config for higher speed like in (ggerganov/llama.cpp#7414)
but made 2 changes:

- Remove UMA build option
- Use it in all case if hipalloc failed with 'not have enough memory'

Another change is look for 'hipcc' on linux and not 'amdclang++'

llama.cpp/ggml-cuda.cu

@@ -159,13 +159,7 @@
 #define cudaHostRegisterReadOnly hipHostRegisterReadOnly
 #define cudaHostUnregister hipHostUnregister
 #define cudaLaunchHostFunc hipLaunchHostFunc
-#ifdef GGML_HIP_UMA
-#define cudaMalloc hipMallocManaged
-#define cudaMallocHost(ptr, size) hipHostMalloc(ptr, size)
-#else
-#define cudaMalloc hipMalloc
 #define cudaMallocHost(ptr, size) hipHostMalloc(ptr, size, hipHostMallocDefault)
-#endif
 #define cudaMemcpy hipMemcpy
 #define cudaMemcpyAsync hipMemcpyAsync
 #define cudaMemcpyPeerAsync hipMemcpyPeerAsync
@@ -10866,6 +10860,25 @@ int ggml_cuda_get_device() {
     return id;
 }
 
+static inline cudaError_t ggml_cuda_device_malloc(void ** ptr, size_t size, int device) {
+#if defined(GGML_USE_HIPBLAS)
+    auto res = hipMalloc(ptr, size);
+    // if Not enough space on VRAM => try with UMA
+    if (res == hipErrorOutOfMemory) {
+        GGML_CUDA_LOG_INFO("  Device %d: can not alloc %d MB on VRAM try alloc on HMM\n", device, (uint32_t)(size / 1024 / 1024));
+        res = hipMallocManaged(ptr, size);
+        if (res == hipSuccess) {
+            // Config the memory for best speed (It's not supposed to fail)
+            CUDA_CHECK(hipMemAdvise(*ptr, size, hipMemAdviseSetCoarseGrain, device));
+            GGML_CUDA_LOG_INFO("    => success\n");
+        }
+    }
+    return res;
+#else
+    return cudaMalloc(ptr, size);
+#endif
+}
+
 static ggml_cuda_device_info ggml_cuda_init() {
 #ifdef __HIP_PLATFORM_AMD__
     // Workaround for a rocBLAS bug when using multiple graphics cards:
@@ -11020,7 +11033,7 @@ struct ggml_cuda_pool_leg : public ggml_cuda_pool {
         size_t look_ahead_size = (size_t) (1.05 * size);
         look_ahead_size = 256 * ((look_ahead_size + 255)/256);
         ggml_cuda_set_device(device);
-        CUDA_CHECK(cudaMalloc((void **) &ptr, look_ahead_size));
+        CUDA_CHECK(ggml_cuda_device_malloc(&ptr, look_ahead_size, device));
         *actual_size = look_ahead_size;
         pool_size += look_ahead_size;
 #ifdef DEBUG_CUDA_MALLOC
@@ -11286,7 +11299,7 @@ GGML_CALL static ggml_backend_buffer_t ggml_backend_cuda_buffer_type_alloc_buffe
     size = std::max(size, (size_t)1); // cudaMalloc returns null for size 0
 
     void * dev_ptr;
-    cudaError_t err = cudaMalloc(&dev_ptr, size);
+    cudaError_t err = ggml_cuda_device_malloc(&dev_ptr, size, buft_ctx->device);
     if (err != cudaSuccess) {
         // clear the error
         cudaGetLastError();
@@ -11547,7 +11560,7 @@ GGML_CALL static void ggml_backend_cuda_split_buffer_init_tensor(ggml_backend_bu
         // currently, init_tensor cannot fail, it needs to be fixed in ggml-backend first
         ggml_cuda_set_device(id);
         char * buf;
-        CUDA_CHECK(cudaMalloc(&buf, size));
+        CUDA_CHECK(ggml_cuda_device_malloc((void**)&buf, size, id));
 
         // set padding to 0 to avoid possible NaN values
         if (size > original_size) {

llamafile/cuda.c

@@ -778,9 +778,9 @@ static bool import_cuda_impl(void) {
 
     char dso[PATH_MAX];
     char bindir[PATH_MAX];
-    const char *compiler_path;
+    const char *compiler_path = NULL;
     char compiler_path_buf[PATH_MAX];
-    const char *library_path;
+    const char *library_path = NULL;
     char library_path_buf[PATH_MAX];
 
     // Attempt to load AMD GPU support.
@@ -791,15 +791,21 @@ static bool import_cuda_impl(void) {
 
         // Get some essential paths.
         // ROCm SDK puts BLAS DLLs in same folder as clang++
-        if (get_rocm_bin_path(compiler_path_buf, "amdclang++") ||
-            get_rocm_bin_path(compiler_path_buf, "clang++")) {
-            strcpy(library_path_buf, compiler_path_buf);
-            dirname(library_path_buf);
-            compiler_path = compiler_path_buf;
-            library_path = library_path_buf;
+        if (!IsWindows()) {
+            if (get_rocm_bin_path(compiler_path_buf, "hipcc")) {
+                strcpy(library_path_buf, compiler_path_buf);
+                dirname(library_path_buf);
+                compiler_path = compiler_path_buf;
+                library_path = library_path_buf;
+            }
         } else {
-            compiler_path = 0;
-            library_path = 0;
+            if (get_rocm_bin_path(compiler_path_buf, "amdclang++") ||
+                get_rocm_bin_path(compiler_path_buf, "clang++")) {
+                strcpy(library_path_buf, compiler_path_buf);
+                dirname(library_path_buf);
+                compiler_path = compiler_path_buf;
+                library_path = library_path_buf;
+            }
         }
 
         // Get path of GGML DSO for AMD.