[Version, Doc] Publish new version.

Author: kaniandr

CHANGELOG.md

@@ -9,3 +9,16 @@ Evaluation of standard C/C++ library files has been fixed.
 ### Version 0.0.3: January 23, 2017
 
 Some activation issues have been fixed.
+
+### Version 0.0.4: February 4, 2021
+
+Provide the following analysis results:
+
+- highlight parallelizable loops as well as parallelization issues,
+- show call graph on request,
+- show a summary of memory accessed in a loop and a summary of memory-related traits,
+- show analysis statistic on request,
+
+Add some source-to-source program transformations.
+
+Implement automatic program parallelization for shared memory (multicore CPU and GPU) with OpenMP and DVMH.

README.md

@@ -1,15 +1,26 @@
 # TSAR Advisor
 
 This extension proposes some useful information about a program to simplify its parallelization.
-It also provide some useful source-to-source transformations for programs as well as automate theire parallelization.
-The advisor currently support C/C++ languages and based on LLVM & Clang projects.
+It also provides some useful source-to-source transformations for programs as well as automates their parallelization.
+The advisor currently supports C/C++ languages and based on LLVM & Clang projects.
 
-To analyse or transform source code:
+> __Attention.__ To run the adviser the TSAR tool must be installed. It must be available in the system PATH or the path to `tsar-server` executable must be configured manually (see `Analysis Server` configuration variable for details). Details about the TSAR tool can be found in [TSAR Wiki](https://github.com/dvm-system/tsar/wiki).
 
-* Open C/C++ file
-* Use context menu or type one of available TSAR commands in the Command Palette.
-* If `TSAR: Analyse file` action is used the description of discovered traits is rendered in the new tab.
-* If some transformation is requested, the original file is changed on success.
-* Some diagnostic messages may be also provided.
+Use the context menu or type one of the available TSAR commands in the Command Palette.
+
+![Start Analysis Session](images/run-advisor.png)
+
+If `TSAR: Analyse file` action is used the summary of discovered traits is rendered in the new tab.
+
+![Analysis Summary](images/summary.png)
+
+Loops that could be parallelized are marked with green. A detailed description of accessed memory is available for each loop.
+
+![Alias Tree](images/alias-tree.png)
+
+You may also choose a function to view its callgraph.
+
+![Call Graph](images/callgraph.png)
+
+If some transformation is requested, the original file is changed on success. Some diagnostic messages may be also provided.
 
-> __Attention.__ To run the adviser the TSAR tool must be installed. It must be available in the system PATH or the path to `tsar-server` executable must be configured manually (see `Analysis Server` configuration variable for details). Details about the TSAR tool can be found in [TSAR Wiki](https://github.com/dvm-system/tsar/wiki).

images/alias-tree.png

@@ -5,6 +5,11 @@
   "version": "0.0.4",
   "publisher": "DVM",
   "homepage": "http://dvm-system.org",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/dvm-system/tsar-advisor"
+  },
+  "icon": "images/logo.png",
   "keywords": [
     "parallel programming",
     "program analysis",

images/callgraph.png

@@ -123,11 +123,12 @@ export function activate(context: vscode.ExtensionContext) {
       title: 'TSAR Parallelization with OpenMP',
       run: '-clang-openmp-parallel'
     },
-    {
+/*    {
       command: 'tsar.parallel.dvmh',
       title: 'TSAR Parallelization with DVMH',
       run: '-clang-experimental-apc-dvmh'
     },
+*/
     {
       command: 'tsar.parallel.dvmhsm',
       title: 'TSAR Shared Memory Parallelization with DVMH',