root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-03 23:34:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

metal : fix from ptr buffer name

Browse Source
This commit is contained in:
slaren 2024-11-06 01:10:53 +01:00
parent b8deef0ec0
commit d2c004fd27
2 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
ggml/src/ggml-metal.m
View File

@ -3844,7 +3844,7 @@ static ggml_backend_buffer_t ggml_backend_metal_device_buffer_from_ptr(ggml_back
}
}
return ggml_backend_buffer_init(ggml_backend_metal_buffer_type(), ggml_backend_metal_buffer_i, ctx, size);
return ggml_backend_buffer_init(ggml_backend_metal_buffer_from_ptr_type(), ggml_backend_metal_buffer_i, ctx, size);
}
static bool ggml_backend_metal_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) {
@ -3854,7 +3854,8 @@ static bool ggml_backend_metal_device_supports_op(ggml_backend_dev_t dev, const
}
static bool ggml_backend_metal_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) {
return buft->iface.get_name == ggml_backend_metal_buffer_type_get_name;
return buft->iface.get_name == ggml_backend_metal_buffer_type_get_name ||
buft->iface.get_name == ggml_backend_metal_buffer_from_ptr_type_get_name;
UNUSED(dev);
}

2
src/llama.cpp
View File

@ -9134,7 +9134,7 @@ static bool llm_load_tensors(
// print memory requirements per buffer type
for (auto & buf : model.bufs) {
LLAMA_LOG_INFO("%s: %10s model buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get()) / 1024.0 / 1024.0);
LLAMA_LOG_INFO("%s: %12s model buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get()) / 1024.0 / 1024.0);
}
// populate tensors_by_name