From 091ab2bd594110c5027af5b467bc82a29b39d60d Mon Sep 17 00:00:00 2001
From: binbin Deng <108676127+plusbang@users.noreply.github.com>
Date: Tue, 4 Mar 2025 10:41:39 +0800
Subject: [PATCH] [NPU] Add troubleshooting in portable zip doc (#12924)

---
 .../llama_cpp_npu_portable_zip_quickstart.md           |  9 ++++++++-
 .../llama_cpp_npu_portable_zip_quickstart.zh-CN.md     | 10 +++++++++-
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.md b/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.md
index c5e225dc..9212ec0f 100644
--- a/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.md
+++ b/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.md
@@ -17,6 +17,7 @@ IPEX-LLM provides llama.cpp support for running GGUF models on Intel NPU. This g
 - [Step 2: Setup](#step-2-setup)
 - [Step 3: Run GGUF Model](#step-3-run-gguf-model)
 - [More details](npu_quickstart.md)
+- [Troubleshooting](#troubleshooting)
 
 
 ## Prerequisites
@@ -34,7 +35,7 @@ Then, extract the zip file to a folder.
 
 ## Step 2: Setup
 
-- Open "Command Prompt" (cmd), and enter the extracted folder through `cd /d PATH\TO\EXTRACTED\FOLDER`
+- Open **"Command Prompt" (cmd)**, and enter the extracted folder through `cd /d PATH\TO\EXTRACTED\FOLDER`
 - Runtime configuration based on your device:
   - For **Intel Core™ Ultra Processors (Series 2) with processor number 2xxV (code name Lunar Lake)**:
 
@@ -63,3 +64,9 @@ You could then use cli tool to run GGUF models on Intel NPU through running `lla
 ```cmd
 llama-cli-npu.exe -m DeepSeek-R1-Distill-Qwen-7B-Q6_K.gguf -n 32 --prompt "What is AI?"
 ```
+
+## Troubleshooting
+
+### `L0 pfnCreate2 result: ZE_RESULT_ERROR_INVALID_ARGUMENT, code 0x78000004` error
+
+First, verify that your NPU driver version meets the requirement. Then, check the runtime configuration based on your device. And please attention the difference between **Command Prompt** and **Windows PowerShell**. Take Arrow Lake for example, you need to use `set IPEX_LLM_NPU_ARL=1` in **Command Prompt** while `$env:IPEX_LLM_NPU_ARL = "1"` in **Windows PowerShell**.
diff --git a/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.zh-CN.md b/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.zh-CN.md
index 1e093569..df5a012b 100644
--- a/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.zh-CN.md
+++ b/docs/mddocs/Quickstart/llama_cpp_npu_portable_zip_quickstart.zh-CN.md
@@ -17,6 +17,7 @@ IPEX-LLM 提供了 llama.cpp 的相关支持以在 Intel NPU 上运行 GGUF 模
 - [步骤 2：启动](#步骤-2启动)
 - [步骤 3：运行 GGUF 模型](#步骤-3运行-gguf-模型)
 - [更多信息](npu_quickstart.md)
+- [故障排除](#故障排除)
 
 
 ## 系统环境准备
@@ -34,7 +35,7 @@ IPEX-LLM 提供了 llama.cpp 的相关支持以在 Intel NPU 上运行 GGUF 模
 
 ## 步骤 2：启动
 
-- 打开命令提示符（cmd），并通过在命令行输入指令 "cd /d PATH\TO\EXTRACTED\FOLDER" 进入解压缩后的文件夹
+- 打开**命令提示符（cmd）**，并通过在命令行输入指令 "cd /d PATH\TO\EXTRACTED\FOLDER" 进入解压缩后的文件夹
 - 根据你的设备完成运行配置：
   - 对于 **处理器为 2xxV 的 Intel Core™ Ultra Processors (Series 2) (代号 Lunar Lake)**:
 
@@ -63,3 +64,10 @@ IPEX-LLM 提供了 llama.cpp 的相关支持以在 Intel NPU 上运行 GGUF 模
 ```cmd
 llama-cli-npu.exe -m DeepSeek-R1-Distill-Qwen-7B-Q6_K.gguf -n 32 --prompt "What is AI?"
 ```
+
+## 故障排除
+
+### `L0 pfnCreate2 result: ZE_RESULT_ERROR_INVALID_ARGUMENT, code 0x78000004` 报错
+
+首先确认你的 NPU 驱动版本是否符合要求，然后根据你的设备检查运行时配置，请注意 **命令提示符** 和 **Windows PowerShell** 的区别。
+以 Arrow Lake 为例，在 **命令提示符** 中需要设置 `set IPEX_LLM_NPU_ARL=1`，而在 **Windows PowerShell** 中是 `$env:IPEX_LLM_NPU_ARL = "1"`。