llama_cpp 0.3.8 → 0.5.0

data/ext/llama_cpp/src/ggml-cuda.h

@@ -2,6 +2,14 @@
 
 #include "ggml.h"
 
+#ifdef GGML_USE_HIPBLAS
+#define GGML_CUDA_NAME "ROCm"
+#define GGML_CUBLAS_NAME "hipBLAS"
+#else
+#define GGML_CUDA_NAME "CUDA"
+#define GGML_CUBLAS_NAME "cuBLAS"
+#endif
+
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -16,9 +24,14 @@ GGML_API bool   ggml_cuda_can_mul_mat(const struct ggml_tensor * src0, const str
 GGML_API void   ggml_cuda_set_tensor_split(const float * tensor_split);
 GGML_API void   ggml_cuda_transform_tensor(void * data, struct ggml_tensor * tensor);
 GGML_API void   ggml_cuda_free_data(struct ggml_tensor * tensor);
+
 GGML_API void   ggml_cuda_assign_buffers(struct ggml_tensor * tensor);
 GGML_API void   ggml_cuda_assign_buffers_no_scratch(struct ggml_tensor * tensor);
 GGML_API void   ggml_cuda_assign_buffers_force_inplace(struct ggml_tensor * tensor);
+
+GGML_API void   ggml_cuda_assign_buffers_no_alloc(struct ggml_tensor * tensor);
+GGML_API void   ggml_cuda_assign_scratch_offset(struct ggml_tensor * tensor, size_t offset);
+
 GGML_API void   ggml_cuda_set_main_device(int main_device);
 GGML_API void   ggml_cuda_set_mul_mat_q(bool mul_mat_q);
 GGML_API void   ggml_cuda_set_scratch_size(size_t scratch_size);

data/ext/llama_cpp/src/ggml-metal.h

@@ -24,6 +24,7 @@
 
 // max memory buffers that can be mapped to the device
 #define GGML_METAL_MAX_BUFFERS 16
+#define GGML_METAL_MAX_COMMAND_BUFFERS 32
 
 struct ggml_tensor;
 struct ggml_cgraph;
@@ -38,6 +39,9 @@ struct ggml_metal_context;
 struct ggml_metal_context * ggml_metal_init(int n_cb);
 void ggml_metal_free(struct ggml_metal_context * ctx);
 
+void * ggml_metal_host_malloc(size_t n);
+void   ggml_metal_host_free  (void * data);
+
 // set the number of command buffers to use
 void ggml_metal_set_n_cb(struct ggml_metal_context * ctx, int n_cb);