Fixing github actions tests

IITH-Compilers · Oct 21, 2023 · cb4c6a3 · cb4c6a3
1 parent f19afc1
commit cb4c6a3
Show file tree

Hide file tree

Showing 140 changed files with 140 additions and 84,805 deletions.
diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json
@@ -3,7 +3,7 @@
 {
 	"name": "LLVM Manylinux",
 	// Or use a Dockerfile or Docker Compose file. More info: https://containers.dev/guide/dockerfile
-	"image": "ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm12"
+	"image": "ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm14"
 
 	// Features to add to the dev container. More info: https://containers.dev/features.
 	// "features": {},

diff --git a/.github/workflows/upload-pypi.yml b/.github/workflows/upload-pypi.yml
@@ -22,7 +22,7 @@ jobs:
 
   build_sdist:
     runs-on: ubuntu-latest
-    container: ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm12
+    container: ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm14
     steps:
       - uses: actions/checkout@v3
 

diff --git a/Manylinux2014_Compliant_Source/pkg/IR2Vec/core.cpp b/Manylinux2014_Compliant_Source/pkg/IR2Vec/core.cpp
@@ -101,9 +101,9 @@ PyObject *IR2Vec_generateEmbeddings(PyObject *self, PyObject *args) {
                                // Thus , need to add one more .cpp in .so
   const char *mode = "\0";
 
-  string vocab_path = seed_emb_path + "/seedEmbeddingVocab-300-llvm12.txt";
+  string vocab_path = seed_emb_path + "/seedEmbeddingVocab-llvm14.txt";
   // const char*
-  // vocab_path=(seed_emb_path+"/seedEmbeddingVocab-300-llvm12.txt").c_str(); //
+  // vocab_path=(seed_emb_path+"/seedEmbeddingVocab-300-llvm14.txt").c_str(); //
   // this should be exact path till .txt . *** NOT TAKING THIS AS USER
   // DEFINED/PROVIDED ARGS ***
   const char *level = "\0"; // remember that ir2vec accepts a char type for this

diff --git a/Manylinux2014_Compliant_Source/pkg/Manifest.in b/Manylinux2014_Compliant_Source/pkg/Manifest.in
@@ -1 +1 @@
-include ./seedEmbeddingVocab-300-llvm12.txt
+include ./seedEmbeddingVocab-llvm14.txt
diff --git a/Manylinux2014_Compliant_Source/pkg/regen-oracle.sh b/Manylinux2014_Compliant_Source/pkg/regen-oracle.sh
@@ -6,19 +6,19 @@ cd src/test-suite
 
 rm -rf oracle
 
-SEED_VERSION="llvm12"
+SEED_VERSION="llvm14"
 SRC_WD="PE-benchmarks"
 DEST_FOLDER_LL="PE-benchmarks-llfiles-${SEED_VERSION}"
 DEST_FOLDER_SYM="oracle/SYM_${SEED_VERSION}_f"
 DEST_FOLDER_FA="oracle/FA_${SEED_VERSION}_f"
-DEST_FOLDER_SYM_P="oracle/SYM_${SEED_VERSION}"
-DEST_FOLDER_FA_P="oracle/FA_${SEED_VERSION}"
+DEST_FOLDER_SYM_P="oracle/SYM_${SEED_VERSION}_p"
+DEST_FOLDER_FA_P="oracle/FA_${SEED_VERSION}_p"
 
 mkdir -p ${DEST_FOLDER_LL}
 
 for d in ${SRC_WD}/*.c ${SRC_WD}/*.cpp ${SRC_WD}/*.cc; do
 	echo "Compiling ${d} to IR"
-	name=$(basename ${d}) && oname=${name%.*} && clang -S -emit-llvm -Xclang -disable-O0-optnone ${d} -o ${DEST_FOLDER_LL}/${oname}.ll &
+	name=$(basename ${d}) && oname=${name%.*} && clang-14 -S -emit-llvm -Xclang -disable-O0-optnone ${d} -o ${DEST_FOLDER_LL}/${oname}.ll &
 done
 wait
 
@@ -28,8 +28,7 @@ mkdir -p ${DEST_FOLDER_SYM_P}
 mkdir -p ${DEST_FOLDER_FA_P}
 
 IR2VEC_PATH=../../build/bin/ir2vec
-
-VOCAB_PATH="../../vocabulary/seedEmbeddingVocab-300-${SEED_VERSION}.txt"
+VOCAB_PATH="../../vocabulary/seedEmbeddingVocab-${SEED_VERSION}.txt"
 
 while IFS= read -r d; do
 	echo "Generating embeddings for ${d}"

diff --git a/Manylinux2014_Compliant_Source/pkg/tests/test_ir2vec.py b/Manylinux2014_Compliant_Source/pkg/tests/test_ir2vec.py
@@ -56,8 +56,10 @@ def test_fa_p():
         full_path = str((TEST_SUITE_DIR / file).resolve()).strip()
         output = IR2Vec.generateEmbeddings(full_path, "fa", "p")
         p_vectors.append(output["Program_List"])
+
+    print(TEST_SUITE_DIR)
     p_vectors_oracle = read_p_file(
-        TEST_SUITE_DIR / "oracle" / f"FA_{SEED_VERSION}" / "ir2vec.txt"
+        TEST_SUITE_DIR / "oracle" / f"FA_{SEED_VERSION}_p" / "ir2vec.txt"
     )
     for idx, v in enumerate(p_vectors):
         assert v == pytest.approx(p_vectors_oracle[idx], abs=ABS_ACCURACY)
@@ -69,8 +71,10 @@ def test_sym_p():
         full_path = str((TEST_SUITE_DIR / file).resolve()).strip()
         output = IR2Vec.generateEmbeddings(full_path, "sym", "p")
         p_vectors.append(output["Program_List"])
+
+    print(TEST_SUITE_DIR)
     p_vectors_oracle = read_p_file(
-        TEST_SUITE_DIR / "oracle" / f"SYM_{SEED_VERSION}" / "ir2vec.txt"
+        TEST_SUITE_DIR / "oracle" / f"SYM_{SEED_VERSION}_p" / "ir2vec.txt"
     )
     for idx, v in enumerate(p_vectors):
         assert v == pytest.approx(p_vectors_oracle[idx], abs=ABS_ACCURACY)
@@ -83,6 +87,8 @@ def test_fa_f():
         output = IR2Vec.generateEmbeddings(str(full_path).strip(), "fa", "f")
         for fun, vec in output["Function_Dict"].items():
             f_vecs[full_path.name.strip()][fun.strip()] = vec
+
+    print(TEST_SUITE_DIR)
     f_vecs_oracle = read_f_file(
         TEST_SUITE_DIR / "oracle" / f"FA_{SEED_VERSION}_f" / "ir2vec.txt"
     )
@@ -100,6 +106,8 @@ def test_sym_f():
         output = IR2Vec.generateEmbeddings(str(full_path).strip(), "sym", "f")
         for fun, vec in output["Function_Dict"].items():
             f_vecs[full_path.name.strip()][fun.strip()] = vec
+
+    print(TEST_SUITE_DIR)
     f_vecs_oracle = read_f_file(
         TEST_SUITE_DIR / "oracle" / f"SYM_{SEED_VERSION}_f" / "ir2vec.txt"
     )

diff --git a/Manylinux2014_Compliant_Source/readme.md b/Manylinux2014_Compliant_Source/readme.md
@@ -6,7 +6,7 @@ AUTHORS : SHIKHAR JAIN (IITH COMPILERS) & ANILAVA KUNDU (IITH COMPILERS)
 - In order to build from source you need to initiate a manylinux 2014 docker image and then use this source dir accordingly.
 - While building source it is necessary to have a static library of LLVM that has all other static libs within it.
 - These wheel files generated are specific for python abi versions as reflected by their names . But they will work on old/new *nix OSs.
-- The package is specific for LLVM-12.0 and current IR2Vec main branch (1.1.0 verion)
+- The package is specific for LLVM-14.0 and current IR2Vec main branch (1.1.0 verion)
 
 - Inorder to use source on a MANYLINUX2014 Docker image and build it , there are several dependencies w.r.t static libs and header files .
 	- A static lib containing all the static libs of LLVM which can be found in llvm/buid_dir/lib.
@@ -29,7 +29,7 @@ AUTHORS : SHIKHAR JAIN (IITH COMPILERS) & ANILAVA KUNDU (IITH COMPILERS)
 		```
 	- Dir "llvm" : You can get this from llvm-project/llvm/include.
 	- Dir "llvm-c" : You can get this from llvm-project/llvm/include.
-	- "seedEmbeddingVocab-300-llvm12.txt" . You can get this from IR2Vec Source.
+	- "seedEmbeddingVocab-llvm14.txt" . You can get this from IR2Vec Source.
 
 ### In future we plan to automate wheel generation by integrating needed workflows in CI/CD pipeline. These wheels then will reflect changes in either IR2Vec or in LLVM project.
 

diff --git a/README.md b/README.md
@@ -25,21 +25,26 @@ Please see [here](https://compilers.cse.iith.ac.in/projects/ir2vec/) for more de
 
 ## Table Of Contents
 - [IR2Vec](#ir2vec)
-  - [LLVM Version Archive](#llvm-version-archive)
-  - [Table Of Contents](#table-of-contents)
-  - [Installation](#installation)
+    - [LLVM Version Archive](#llvm-version-archive)
+    - [Table Of Contents](#table-of-contents)
+    - [Installation](#installation)
     - [Python](#python)
-    - [C++](#cpp)
-        - [Requirements](#requirements)
-        - [Building from Source](#building-from-source)
-  - [Generating program representations](#generating-program-representations)
-    - [Using Binary](#using-binary)
+    - [Cpp](#cpp)
+    - [Requirements](#requirements)
+    - [Building from source](#building-from-source)
+    - [Generating program representations](#generating-program-representations)
+        - [Using Binary](#using-binary)
+            - [Command-Line options](#command-line-options)
+            - [Flow-Aware Embeddings](#flow-aware-embeddings)
+            - [Symbolic Embeddings](#symbolic-embeddings)
     - [Using Libraries](#using-libraries)
     - [Using Python package (IR2Vec-Wheels)](#using-python-package-ir2vec-wheels)
-  - [Experiments](#experiments)
-  - [Citation](#citation)
-  - [Contributions](#contributions)
-  - [License](#license)
+    - [Binaries, Libraries and Wheels - Artifacts](#binaries-libraries-and-wheels---artifacts)
+    - [Experiments](#experiments)
+        - [Note](#note)
+    - [Citation](#citation)
+    - [Contributions](#contributions)
+    - [License](#license)
 
 ## Installation
 
@@ -127,16 +132,16 @@ Please use `--help` for further details.
 
 #### Flow-Aware Embeddings
 For all functions
-* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-300-llvm12.txt -o <output_file> -level <p|f>  -class <class-number> <input_ll_file>``
+* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level <p|f>  -class <class-number> <input_ll_file>``
 
 For a specific function
-* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-300-llvm12.txt -o <output_file> -level f  -class <class-number> -funcName=\<function-name\><input_ll_file>``
+* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level f  -class <class-number> -funcName=\<function-name\><input_ll_file>``
 
 #### Symbolic Embeddings
 For all functions
- * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-300-llvm12.txt -o <output_file> -level <p|f> -class <class-number> <input_ll_file>``
+ * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level <p|f> -class <class-number> <input_ll_file>``
 For a specific function
- * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-300-llvm12.txt -o <output_file> -level f -class <class-number> -funcName=\<function-name\> <input_ll_file>``
+ * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level f -class <class-number> -funcName=\<function-name\> <input_ll_file>``
 
 ## Using Libraries
 The libraries can be installed by passing the installation location to the `CMAKE_INSTALL_PREFIX` flag during `cmake` followed by `make install`.
@@ -165,7 +170,7 @@ The following example snippet shows how to query the exposed vector representati
 // Creating object to generate FlowAware representation
 auto ir2vec =
       IR2Vec::Embeddings(<LLVM Module>, IR2Vec::IR2VecMode::FlowAware,
-                         "./vocabulary/seedEmbeddingVocab-300-llvm12.txt");
+                         "./vocabulary/seedEmbeddingVocab-llvm14.txt");
 
 // Getting Instruction vectors corresponding to the instructions in <LLVM Module>
 auto instVecMap = ir2vec.getInstVecMap();

diff --git a/experiments/Device_Mapping/DevMap.ipynb b/experiments/Device_Mapping/DevMap.ipynb
@@ -285,22 +285,22 @@
       "Accuracy Matrix: IR2Vec Vs. others\n",
       "\n",
       "                 Static Mapping  Grewe et al.   DeepTune     NCC     IR2Vec\n",
-      "AMD Tahiti 7970       58.823529     73.382353  83.676471  82.790  90.284006\n",
-      "NVIDIA GTX 970        56.911765     72.941176  80.294118  81.760  87.144993\n",
-      "Average               57.867647     73.161765  81.985294  82.275  88.714499\n",
+      "AMD Tahiti 7970       58.823529     73.382353  83.676471  82.790  91.766467\n",
+      "NVIDIA GTX 970        56.911765     72.941176  80.294118  81.760  86.976048\n",
+      "Average               57.867647     73.161765  81.985294  82.275  89.371257\n",
       "\n",
       "Speedup Matrix: IR2Vec Vs. others\n",
       "\n",
       "                 Static Mapping  Grewe et al.  DeepTune    NCC    IR2Vec\n",
-      "AMD Tahiti 7970             1.0      2.905822  3.335612  3.420  3.471963\n",
-      "NVIDIA GTX 970              1.0      1.264801  1.412222  1.390  1.433372\n",
-      "Average                     1.0      2.085312  2.373917  2.405  2.452667\n"
+      "AMD Tahiti 7970             1.0      2.905822  3.335612  3.420  3.477054\n",
+      "NVIDIA GTX 970              1.0      1.264801  1.412222  1.390  1.453360\n",
+      "Average                     1.0      2.085312  2.373917  2.405  2.465207\n"
      ]
     }
    ],
    "source": [
     "raw_embeddings, fileIndexNum = readEmd_program(\n",
-    "    \"./output/embeddings/Device_Mapping_Symbolic_llvm12.txt\"\n",
+    "    \"./output/embeddings/Device_Mapping_Symbolic_llvm14.txt\"\n",
     ")\n",
     "ir2vec_sym = evaluate(max_depth=10, learning_rate=0.5, n_estimators=70, seed=104)"
    ]
@@ -344,22 +344,22 @@
       "Accuracy Matrix: IR2Vec Vs. others\n",
       "\n",
       "                 Static Mapping  Grewe et al.   DeepTune     NCC     IR2Vec\n",
-      "AMD Tahiti 7970       58.823529     73.382353  83.676471  82.790  92.825112\n",
-      "NVIDIA GTX 970        56.911765     72.941176  80.294118  81.760  89.686099\n",
-      "Average               57.867647     73.161765  81.985294  82.275  91.255605\n",
+      "AMD Tahiti 7970       58.823529     73.382353  83.676471  82.790  91.167665\n",
+      "NVIDIA GTX 970        56.911765     72.941176  80.294118  81.760  88.323353\n",
+      "Average               57.867647     73.161765  81.985294  82.275  89.745509\n",
       "\n",
       "Speedup Matrix: IR2Vec Vs. others\n",
       "\n",
       "                 Static Mapping  Grewe et al.  DeepTune    NCC    IR2Vec\n",
-      "AMD Tahiti 7970             1.0      2.905822  3.335612  3.420  3.510104\n",
-      "NVIDIA GTX 970              1.0      1.264801  1.412222  1.390  1.467221\n",
-      "Average                     1.0      2.085312  2.373917  2.405  2.488663\n"
+      "AMD Tahiti 7970             1.0      2.905822  3.335612  3.420  3.450715\n",
+      "NVIDIA GTX 970              1.0      1.264801  1.412222  1.390  1.445097\n",
+      "Average                     1.0      2.085312  2.373917  2.405  2.447906\n"
      ]
     }
    ],
    "source": [
     "raw_embeddings, fileIndexNum = readEmd_program(\n",
-    "    \"./output/embeddings/Device_Mapping_FlowAware_llvm12.txt\"\n",
+    "    \"./output/embeddings/Device_Mapping_FlowAware_llvm14.txt\"\n",
     ")\n",
     "ir2vec_fa = evaluate(max_depth=10, learning_rate=0.5, n_estimators=70, seed=104)"
    ]
@@ -468,8 +468,8 @@
       "Geometric mean of Grewe et al. 1.22x\n",
       "Geometric mean of DeepTune 1.46x\n",
       "Geometric mean of Inst2Vec 1.40x\n",
-      "Geometric mean of IR2Vec Symbolic 1.559x\n",
-      "Geometric mean of IR2Vec Flow-Aware 1.581x\n"
+      "Geometric mean of IR2Vec Symbolic 1.566x\n",
+      "Geometric mean of IR2Vec Flow-Aware 1.552x\n"
      ]
     }
    ],
@@ -498,8 +498,8 @@
       "Geometric mean of Grewe et al. 1.11x\n",
       "Geometric mean of DeepTune 1.21x\n",
       "Geometric mean of Inst2Vec 1.21x\n",
-      "Geometric mean of IR2Vec Symbolic 1.237x\n",
-      "Geometric mean of IR2Vec Flow-Aware 1.261x\n"
+      "Geometric mean of IR2Vec Symbolic 1.248x\n",
+      "Geometric mean of IR2Vec Flow-Aware 1.258x\n"
      ]
     }
    ],
@@ -528,8 +528,8 @@
       "Geometric mean of Grewe et al. - 1.16x\n",
       "Geometric mean of DeepTune - 1.33x\n",
       "Geometric mean of Inst2Vec - 1.30x\n",
-      "Geometric mean of IR2Vec Symbolic 1.39x\n",
-      "Geometric mean of IR2Vec Flow-Aware 1.41x\n"
+      "Geometric mean of IR2Vec Symbolic 1.40x\n",
+      "Geometric mean of IR2Vec Flow-Aware 1.40x\n"
      ]
     }
    ],
@@ -575,16 +575,16 @@
      "text": [
       "\n",
       "AMD Tahiti 7970\n",
-      " % Increase in SpeedUp over Grewe et al -  29.65\n",
-      " % Increase in SpeedUp over DeepTune -  7.96\n",
-      " % Increase in SpeedUp over Inst2Vec -  12.95\n",
-      " % Increase in SpeedUp over IR2Vec Symbolic -  1.43\n",
+      " % Increase in SpeedUp over Grewe et al -  27.27\n",
+      " % Increase in SpeedUp over DeepTune -  5.98\n",
+      " % Increase in SpeedUp over Inst2Vec -  10.88\n",
+      " % Increase in SpeedUp over IR2Vec Symbolic -  0.86\n",
       "\n",
       "NVIDIA GTX 970\n",
-      " % Increase in SpeedUp over Grewe et al -  13.77\n",
-      " % Increase in SpeedUp over DeepTune -  4.31\n",
-      " % Increase in SpeedUp over Inst2Vec -  3.93\n",
-      " % Increase in SpeedUp over IR2Vec Symbolic -  1.92\n"
+      " % Increase in SpeedUp over Grewe et al -  13.51\n",
+      " % Increase in SpeedUp over DeepTune -  4.07\n",
+      " % Increase in SpeedUp over Inst2Vec -  3.7\n",
+      " % Increase in SpeedUp over IR2Vec Symbolic -  0.8\n"
      ]
     }
    ],
@@ -693,8 +693,8 @@
       "Accuracy of Grewe et al. 70.29%\n",
       "Accuracy of DeepTune 83.24%\n",
       "Accuracy of Inst2Vec 82.35%\n",
-      "Accuracy of IR2Vec Symbolic 90.284%\n",
-      "Accuracy of IR2Vec Flow-Aware 92.825%\n"
+      "Accuracy of IR2Vec Symbolic 91.766%\n",
+      "Accuracy of IR2Vec Flow-Aware 91.168%\n"
      ]
     }
    ],
@@ -721,8 +721,8 @@
       "Accuracy of Grewe et al. 74.56%\n",
       "Accuracy of DeepTune 80.88%\n",
       "Accuracy of Inst2Vec 82.65%\n",
-      "Accuracy of IR2Vec Symbolic 87.145%\n",
-      "Accuracy of IR2Vec Flow-Aware 89.686%\n"
+      "Accuracy of IR2Vec Symbolic 86.976%\n",
+      "Accuracy of IR2Vec Flow-Aware 88.323%\n"
      ]
     }
    ],
@@ -757,11 +757,11 @@
      "text": [
       "\n",
       "AMD Tahiti 7970\n",
-      " % Increase in SpeedUp over Grewe et al -  26.5\n",
-      " % Increase in SpeedUp over DeepTune -  10.93\n",
-      " % Increase in SpeedUp over Inst2Vec -  12.12\n",
-      " % Increase in SpeedUp over Inst2Vec-imm -  5.38\n",
-      " % Increase in SpeedUp over IR2Vec Symbolic -  2.81\n"
+      " % Increase in SpeedUp over Grewe et al -  24.24\n",
+      " % Increase in SpeedUp over DeepTune -  8.95\n",
+      " % Increase in SpeedUp over Inst2Vec -  10.12\n",
+      " % Increase in SpeedUp over Inst2Vec-imm -  3.49\n",
+      " % Increase in SpeedUp over IR2Vec Symbolic -  0.65\n"
      ]
     }
    ],
@@ -801,11 +801,11 @@
      "text": [
       "\n",
       "NVIDIA GTX 970\n",
-      " % Increase in SpeedUp over Grewe et al -  22.96\n",
-      " % Increase in SpeedUp over DeepTune -  11.7\n",
-      " % Increase in SpeedUp over Inst2Vec -  9.69\n",
-      " % Increase in SpeedUp over Inst2Vec -  3.54\n",
-      " % Increase in SpeedUp over IR2Vec Symbolic -  2.92\n"
+      " % Increase in SpeedUp over Grewe et al -  21.09\n",
+      " % Increase in SpeedUp over DeepTune -  10.0\n",
+      " % Increase in SpeedUp over Inst2Vec -  8.03\n",
+      " % Increase in SpeedUp over Inst2Vec -  1.97\n",
+      " % Increase in SpeedUp over IR2Vec Symbolic -  1.55\n"
      ]
     }
    ],
@@ -845,10 +845,10 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      " % Increase in SpeedUp over DeepTune -  11.3\n",
-      " % Increase in SpeedUp over Inst2Vec -  10.92\n",
-      " % Increase in SpeedUp over Inst2Vec -  4.47\n",
-      " % Increase in SpeedUp over IR2Vec Symbolic -  2.86\n"
+      " % Increase in SpeedUp over DeepTune -  9.46\n",
+      " % Increase in SpeedUp over Inst2Vec -  9.08\n",
+      " % Increase in SpeedUp over Inst2Vec -  2.74\n",
+      " % Increase in SpeedUp over IR2Vec Symbolic -  0.42\n"
      ]
     }
    ],
@@ -871,7 +871,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python 3 (ipykernel)",
    "language": "python",
    "name": "python3"
   },
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		include ./seedEmbeddingVocab-300-llvm12.txt
		include ./seedEmbeddingVocab-llvm14.txt