Maxwell/Pascal GPU support and crash fix (#1895)

Signed-off-by: Jared Van Bortel <jared@nomic.ai>

gpt4all-backend/llama.cpp-mainline

@@ -1 +1 @@
-Subproject commit d2e3b5fd464aa95d44b3aeb6d532ad7374f7478b
+Subproject commit b0df9e6fad734a4cff48e80b120b72f1f2c39094

gpt4all-chat/chatlistmodel.h

@@ -196,6 +196,7 @@ public:
         m_newChat = nullptr;
         m_serverChat = nullptr;
         m_currentChat = nullptr;
+        for (auto * chat: m_chats) { delete chat; }
         m_chats.clear();
     }
 

gpt4all-chat/main.cpp

@@ -63,9 +63,11 @@ int main(int argc, char *argv[])
     }
 #endif
 
+    int res = app.exec();
+
     // Make sure ChatLLM threads are joined before global destructors run.
     // Otherwise, we can get a heap-use-after-free inside of llama.cpp.
     ChatListModel::globalInstance()->clearChats();
 
-    return app.exec();
+    return res;
 }