From a7b611bc933060cf2c051051cf5001c8f30cac36 Mon Sep 17 00:00:00 2001
From: "Yu, Zijun" <zijun.yu@intel.com>
Date: Thu, 14 Aug 2025 16:52:29 +0800
Subject: [PATCH] Minor updates for raising PR

---
 CMakePresets.json                       | 20 --------------------
 docs/build.md                           | 21 +++------------------
 ggml/src/ggml-openvino/ggml-decoder.cpp |  3 +--
 3 files changed, 4 insertions(+), 40 deletions(-)

diff --git a/CMakePresets.json b/CMakePresets.json
index 392c357f37..b5afeb3c0f 100644
--- a/CMakePresets.json
+++ b/CMakePresets.json
@@ -1,26 +1,6 @@
 {
   "version": 4,
   "configurePresets": [
-    {
-        "name": "ReleaseOV",
-        "generator": "Ninja",
-        "binaryDir": "${sourceDir}/build/${presetName}",
-        "installDir": "${sourceDir}/build/install/${presetName}",
-        "cacheVariables": {
-            "CMAKE_BUILD_TYPE": "Release",
-            "GGML_OPENVINO": true,
-            "OpenVINO_DIR": "$env{OPENVINO_LLAMA_PATH}/build/Release"
-        }
-    },
-    {
-        "name": "ReleaseCPU",
-        "generator": "Ninja",
-        "binaryDir": "${sourceDir}/build/${presetName}",
-        "installDir": "${sourceDir}/build/install/${presetName}",
-        "cacheVariables": {
-            "CMAKE_BUILD_TYPE": "Release"
-        }
-    },
     {
         "name":  "base",
         "hidden": true,
diff --git a/docs/build.md b/docs/build.md
index 1424a06508..9e44f18eae 100644
--- a/docs/build.md
+++ b/docs/build.md
@@ -698,7 +698,7 @@ To read documentation for how to build on IBM Z & LinuxONE, [click here](./build
 
 ## OpenVINO
 
-[OpenVINO](https://docs.openvino.ai/2025/index.html) is an open-source toolkit for optimizing and deploying high-performance AI inference, specifically designed for Intel hardware, including CPUs, GPUs, and NPUs, in the cloud, on-premises, and on the edge. 
+[OpenVINO](https://docs.openvino.ai/2025/index.html) is an open-source toolkit for optimizing and deploying high-performance AI inference, specifically designed for Intel hardware, including CPUs, GPUs, and NPUs, in the cloud, on-premises, and on the edge.
 The OpenVINO backend enhances performance by leveraging hardware-specific optimizations and can be enabled for use with llama.cpp.
 
 Follow the instructions below to install OpenVINO runtime and build llama.cpp with OpenVINO support.
@@ -800,9 +800,8 @@ export GGML_OPENVINO_CACHE_DIR=/tmp/ov_cache
 
 Control OpenVINO behavior using these environment variables:
 
--   **`GGML_OPENVINO_DEVICE`**: Specify the target device for OpenVINO inference.  If not set, automatically selects the first available device in priority order: GPU, CPU, NPU. When set to `NPU` to use Intel NPUs, it enables static compilation mode for optimal performance. 
--   **`GGML_OPENVINO_CACHE_DIR`**: Directory for model caching (recommended: `/tmp/ov_cache`). If set, enables model caching in OpenVINO. Note: Not supported when using NPU devices yet.  
--   **`GGML_OPENVINO_WEIGHT_AS_INPUT`**: Pass the weights as input to the OpenVINO model instead of creating Constant nodes for them.
+-   **`GGML_OPENVINO_DEVICE`**: Specify the target device for OpenVINO inference.  If not set, automatically selects the first available device in priority order: GPU, CPU, NPU. When set to `NPU` to use Intel NPUs, it enables static compilation mode for optimal performance.
+-   **`GGML_OPENVINO_CACHE_DIR`**: Directory for model caching (recommended: `/tmp/ov_cache`). If set, enables model caching in OpenVINO. Note: Not supported when using NPU devices yet.
 -   **`GGML_OPENVINO_PROFILING`**: Enable execution time profiling.
 -   **`GGML_OPENVINO_DUMP_CGRAPH`**: Save compute graph to `cgraph.txt`.
 -   **`GGML_OPENVINO_DUMP_IR`**: Export OpenVINO IR files with timestamps.
@@ -817,20 +816,6 @@ export GGML_OPENVINO_PROFILING=1
 
 ./build/ReleaseOV/bin/llama-simple -m ~/models/Llama-3.2-1B-Instruct.fp16.gguf -n 50 "The story of AI is "
 ```
-> **Note:** To apply your code changes, clear the `GGML_OPENVINO_CACHE_DIR` directory and rebuild the project.
-
-### Using Llama.cpp's Built-in CPU Backend (for Comparison)
-
-To compare performance with the default CPU backend:
-
-```bash
-# Build CPU-only version
-cmake --preset ReleaseCPU
-cmake --build build/ReleaseCPU --parallel
-
-# Run with the default CPU backend
-./build/ReleaseCPU/bin/llama-simple -m ~/models/Llama-3.2-1B-Instruct.fp16.gguf -n 50 "The story of AI is "
-```
 
 ## Notes about GPU-accelerated backends
 
diff --git a/ggml/src/ggml-openvino/ggml-decoder.cpp b/ggml/src/ggml-openvino/ggml-decoder.cpp
index 6bc2c253e8..09919c8505 100644
--- a/ggml/src/ggml-openvino/ggml-decoder.cpp
+++ b/ggml/src/ggml-openvino/ggml-decoder.cpp
@@ -57,8 +57,7 @@ GgmlOvDecoder::GgmlOvDecoder(struct ggml_cgraph* cgraph,
     }
 
     if (getenv("GGML_OPENVINO_DUMP_CGRAPH")) {
-        auto timestamp = (long long) ggml_time_us();
-        std::string filename = "cgraph_" + std::to_string(timestamp) + ".txt";
+        std::string filename = "cgraph.txt";
         dump_cgraph(cgraph, filename);
     }