[LLM]LLM windows load -api.dll (#8631)

* temp

* update

* revert setup.py

python/llm/src/bigdl/llm/ggml/model/bloom/bloom_cpp.py

@@ -81,6 +81,7 @@ def _load_shared_library(lib_base_name: str):
     # Add the library directory to the DLL search path on Windows (if needed)
     if sys.platform == "win32" and sys.version_info >= (3, 8):
         os.add_dll_directory(str(_base_path))
+        os.environ["PATH"] = str(_base_path) + ";" + os.environ["PATH"]
 
     # Try to load the shared library, handling potential errors
     for _lib_path in _lib_paths:

python/llm/src/bigdl/llm/ggml/model/gptneox/gptneox_cpp.py

@@ -81,6 +81,7 @@ def _load_shared_library(lib_base_name: str):
     # Add the library directory to the DLL search path on Windows (if needed)
     if sys.platform == "win32" and sys.version_info >= (3, 8):
         os.add_dll_directory(str(_base_path))
+        os.environ["PATH"] = str(_base_path) + ";" + os.environ["PATH"]
         cdll_args["winmode"] = 0
 
     # Try to load the shared library, handling potential errors

python/llm/src/bigdl/llm/ggml/model/llama/llama_cpp.py

@@ -81,6 +81,7 @@ def _load_shared_library(lib_base_name: str):
     # Add the library directory to the DLL search path on Windows (if needed)
     if sys.platform == "win32" and sys.version_info >= (3, 8):
         os.add_dll_directory(str(_base_path))
+        os.environ["PATH"] = str(_base_path) + ";" + os.environ["PATH"]
         if "CUDA_PATH" in os.environ:
             os.add_dll_directory(os.path.join(os.environ["CUDA_PATH"], "bin"))
             os.add_dll_directory(os.path.join(os.environ["CUDA_PATH"], "lib"))

python/llm/src/bigdl/llm/ggml/model/starcoder/starcoder_cpp.py

@@ -81,6 +81,7 @@ def _load_shared_library(lib_base_name: str):
     # Add the library directory to the DLL search path on Windows (if needed)
     if sys.platform == "win32" and sys.version_info >= (3, 8):
         os.add_dll_directory(str(_base_path))
+        os.environ["PATH"] = str(_base_path) + ";" + os.environ["PATH"]
 
     # Try to load the shared library, handling potential errors
     for _lib_path in _lib_paths:

python/llm/src/bigdl/llm/utils/utils.py

@@ -30,7 +30,8 @@ def get_cpu_flags():
         else:
             invalidOperationError(False, "Unsupported CPUFLAGS.")
     else:
-        flags = "_vnni" if check_avx_vnni() else ""
+        # flags = "_vnni" if check_avx_vnni() else ""
+        flags = "-api"
     return flags