github workflow llvm16

.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-llvm14"
+	"image": "ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm16"
 
 	// Features to add to the dev container. More info: https://containers.dev/features.
 	// "features": {},

.github/workflows/publish.yml

@@ -14,10 +14,10 @@ jobs:
     runs-on: ubuntu-20.04
 
     steps:
-    - name: Install LLVM-12.0.0
+    - name: Install LLVM-16.0.1
       run: |
         wget https://apt.llvm.org/llvm.sh
-        sudo bash llvm.sh 12
+        sudo bash llvm.sh 16
     - uses: actions/checkout@v2
     - name: Eigen-setup
       run: |

.github/workflows/test.yml

@@ -17,7 +17,7 @@ jobs:
     runs-on: ubuntu-20.04
 
     steps:
-    - name: Install LLVM-16.0.0
+    - name: Install LLVM-16.0.1
       run: |
         wget https://apt.llvm.org/llvm.sh
         sudo bash llvm.sh 16

.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-llvm14
+    container: ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm16
     steps:
       - uses: actions/checkout@v3
 

.github/workflows/wheel.yml

@@ -20,7 +20,7 @@ jobs:
         env:
           CIBW_SKIP: "pp* *-musllinux_*"
           CIBW_ARCHS: "x86_64"
-          CIBW_MANYLINUX_X86_64_IMAGE: "ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm14"
+          CIBW_MANYLINUX_X86_64_IMAGE: "ghcr.io/iith-compilers/manylinux2014-llvm/manylinux2014-llvm:x86-llvm16"
           CIBW_BEFORE_ALL: "bash Manylinux2014_Compliant_Source/pkg/build.sh"
           CIBW_TEST_REQUIRES: pytest
           CIBW_TEST_COMMAND: "pytest {project}/Manylinux2014_Compliant_Source/pkg/tests"

Manylinux2014_Compliant_Source/manylinux-llvm/Dockerfile

@@ -36,7 +36,7 @@ RUN cmake -DCMAKE_BUILD_TYPE=Release -G Ninja -B build \
 	&& cmake --install build --prefix /root/destdir
 
 WORKDIR /root/llvm
-ARG LLVM_VERSION="12.0.1"
+ARG LLVM_VERSION="16.0.1"
 RUN wget -q "https://github.com/llvm/llvm-project/releases/download/llvmorg-${LLVM_VERSION}/llvm-project-${LLVM_VERSION}.src.tar.xz" \
 	&& tar -xf "llvm-project-${LLVM_VERSION}.src.tar.xz"
 WORKDIR /root/llvm/llvm-project-${LLVM_VERSION}.src/build

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-llvm14.txt";
+  string vocab_path = seed_emb_path + "/seedEmbeddingVocab-llvm16.txt";
   // const char*
-  // vocab_path=(seed_emb_path+"/seedEmbeddingVocab-300-llvm14.txt").c_str(); //
+  // vocab_path=(seed_emb_path+"/seedEmbeddingVocab-llvm16.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

Manylinux2014_Compliant_Source/pkg/Manifest.in

@@ -1 +1 @@
-include ./seedEmbeddingVocab-llvm14.txt
+include ./seedEmbeddingVocab-llvm16.txt

Manylinux2014_Compliant_Source/pkg/regen-oracle.sh

@@ -6,7 +6,7 @@ cd src/test-suite
 
 rm -rf oracle
 
-SEED_VERSION="llvm14"
+SEED_VERSION="llvm16"
 SRC_WD="PE-benchmarks"
 DEST_FOLDER_LL="PE-benchmarks-llfiles-${SEED_VERSION}"
 DEST_FOLDER_SYM="oracle/SYM_${SEED_VERSION}_f"
@@ -18,7 +18,7 @@ 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-14 -S -emit-llvm -Xclang -disable-O0-optnone ${d} -o ${DEST_FOLDER_LL}/${oname}.ll &
+	name=$(basename ${d}) && oname=${name%.*} && clang-16 -S -emit-llvm -Xclang -disable-O0-optnone ${d} -o ${DEST_FOLDER_LL}/${oname}.ll &
 done
 wait
 

Manylinux2014_Compliant_Source/readme.md

@@ -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-llvm14.txt" . You can get this from IR2Vec Source.
+	- "seedEmbeddingVocab-llvm16.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.
 

README.md

@@ -132,16 +132,16 @@ Please use `--help` for further details.
 
 #### Flow-Aware Embeddings
 For all functions
-* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level <p|f>  -class <class-number> <input_ll_file>``
+* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm16.txt -o <output_file> -level <p|f>  -class <class-number> <input_ll_file>``
 
 For a specific function
-* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level f  -class <class-number> -funcName=\<function-name\><input_ll_file>``
+* `` ir2vec -fa -vocab vocabulary/seedEmbeddingVocab-llvm16.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-llvm14.txt -o <output_file> -level <p|f> -class <class-number> <input_ll_file>``
+ * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-llvm16.txt -o <output_file> -level <p|f> -class <class-number> <input_ll_file>``
 For a specific function
- * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-llvm14.txt -o <output_file> -level f -class <class-number> -funcName=\<function-name\> <input_ll_file>``
+ * `` ir2vec -sym -vocab vocabulary/seedEmbeddingVocab-llvm16.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`.
@@ -170,7 +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-llvm14.txt");
+                         "./vocabulary/seedEmbeddingVocab-llvm16.txt");
 
 // Getting Instruction vectors corresponding to the instructions in <LLVM Module>
 auto instVecMap = ir2vec.getInstVecMap();

experiments/Device_Mapping/DevMap.ipynb

@@ -300,7 +300,7 @@
    ],
    "source": [
     "raw_embeddings, fileIndexNum = readEmd_program(\n",
-    "    \"./output/embeddings/Device_Mapping_Symbolic_llvm14.txt\"\n",
+    "    \"./output/embeddings/Device_Mapping_Symbolic_llvm16.txt\"\n",
     ")\n",
     "ir2vec_sym = evaluate(max_depth=10, learning_rate=0.5, n_estimators=70, seed=104)"
    ]
@@ -359,7 +359,7 @@
    ],
    "source": [
     "raw_embeddings, fileIndexNum = readEmd_program(\n",
-    "    \"./output/embeddings/Device_Mapping_FlowAware_llvm14.txt\"\n",
+    "    \"./output/embeddings/Device_Mapping_FlowAware_llvm16.txt\"\n",
     ")\n",
     "ir2vec_fa = evaluate(max_depth=10, learning_rate=0.5, n_estimators=70, seed=104)"
    ]

experiments/README.md

@@ -32,8 +32,8 @@ Device_Mapping
 |   |-- cgo17-nvidia.csv
 |-- output
 |   |-- embeddings
-|   |   |-- Device_Mapping_Symbolic_llvm14.txt
-|   |   |-- Device_Mapping_FlowAware_llvm14.txt
+|   |   |-- Device_Mapping_Symbolic_llvm16.txt
+|   |   |-- Device_Mapping_FlowAware_llvm16.txt
 |-- DevMap.ipynb
 ```
 ## Thread Coarsening
@@ -60,8 +60,8 @@ Thread_Coarsening
 |   |-- pact-2014-runtimes.csv
 |-- output
 |   |-- embeddings
-|   |   |-- Thread_Coarsening_Symbolic_llvm14.txt
-|   |   |-- Thread_Coarsening_FlowAware_llvm14.txt
+|   |   |-- Thread_Coarsening_Symbolic_llvm16.txt
+|   |   |-- Thread_Coarsening_FlowAware_llvm16.txt
 |-- ThreadCoarsening.ipynb
 ```
 ## Out_Of_Vocabulary

experiments/Thread_Coarsening/ThreadCoarsening.ipynb

@@ -324,7 +324,7 @@
    ],
    "source": [
     "raw_embeddings, fileIndex = readEmd_program(\n",
-    "    \"./output/embeddings/Thread_Coarsening_Symbolic_llvm14.txt\"\n",
+    "    \"./output/embeddings/Thread_Coarsening_Symbolic_llvm16.txt\"\n",
     ")\n",
     "ir2vec_sym = evaluate(max_depth=1, learning_rate=0.05, n_estimators=140)"
    ]
@@ -374,7 +374,7 @@
    ],
    "source": [
     "raw_embeddings, fileIndex = readEmd_program(\n",
-    "    \"./output/embeddings/Thread_Coarsening_FlowAware_llvm14.txt\"\n",
+    "    \"./output/embeddings/Thread_Coarsening_FlowAware_llvm16.txt\"\n",
     ")\n",
     "ir2vec_fa = evaluate(max_depth=1, learning_rate=0.05, n_estimators=140)"
    ]

experiments/generate_IR2Vec_embeddings.sh

@@ -19,7 +19,7 @@ fi
 
 BUILD=$(realpath ${BUILD})
 
-Absolute_path_of_RepresentationFile=$(realpath ../vocabulary/seedEmbeddingVocab-llvm14.txt)
+Absolute_path_of_RepresentationFile=$(realpath ../vocabulary/seedEmbeddingVocab-llvm16.txt)
 
 TASK_DIR=$1
 if [ -z "${TASK_DIR}" ]; then
@@ -45,10 +45,10 @@ fi
 
 if [ $EncodingType = "SYM" ]; then
 	PASS="sym"
-	Trans_type="${TASK_DIR}_Symbolic_llvm14"
+	Trans_type="${TASK_DIR}_Symbolic_llvm16"
 elif [ $EncodingType = "FA" ]; then
 	PASS="fa"
-	Trans_type="${TASK_DIR}_FlowAware_llvm14"
+	Trans_type="${TASK_DIR}_FlowAware_llvm16"
 else
 	echo "Please enter SYM or FA for EncodingType"
 	exit

src/CMakeLists.txt

@@ -4,7 +4,7 @@ project(ir2vec VERSION 2.0.1)
 set(IR2VEC_LIB "IR2Vec")
 set(IR2VEC_LIB_STATIC "IR2Vec_Static")
 
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O2")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O2 -std=c++17")
 
 
 set(LT_LLVM_INSTALL_DIR "" CACHE PATH "LLVM installation directory")