Use deque instead of vector for code readers so that the iterators and references will be stable (#1851)

* Use deque instead of vector for code readers so that the iterators and references will be stable

* Fix compile error

* Assign the iterator

* Add multithreaded test

* Make threads a multiple of hardware concurrency

* Output on failure

* Add setDevice to try and initialize the context on cuda

* Create context for cuda

* Set context on each thread

* Reduce threads on cuda

* Skip test on cuda

* Try to initialize the primary context on cuda

* Push ctx to the stack as current

* Revert "Push ctx to the stack as current"

This reverts commit 0fb72d9e7f.

* Revert "Try to initialize the primary context on cuda"

This reverts commit 44e3908680.

* updated test for nvidia path

* Add c++11 option for nvcc

Co-authored-by: satyanveshd <53337087+satyanveshd@users.noreply.github.com>


[ROCm/hip commit: e7120dd876]

projects/hip/src/program_state.inl

@@ -26,6 +26,7 @@
 #include <cstddef>
 #include <cstdint>
 #include <cstdio>
+#include <deque>
 #include <memory>
 #include <mutex>
 #include <stdexcept>
@@ -202,7 +203,7 @@ public:
                         std::function<void(hsa_code_object_reader_t*)>>;
     std::pair<
         std::mutex,
-        std::vector<std::pair<std::string, RAII_code_reader>>> code_readers;
+        std::deque<std::pair<std::string, RAII_code_reader>>> code_readers;
 
     program_state_impl() {
         // Create placeholder for each agent for the per-agent members.
@@ -419,8 +420,8 @@ public:
         decltype(code_readers.second)::iterator it;
         {
           std::lock_guard<std::mutex> lck{code_readers.first};
-          it = code_readers.second.emplace(code_readers.second.end(),
-                                           move(file), move(tmp));
+          code_readers.second.emplace_back(move(file), move(tmp));
+          it = std::prev(code_readers.second.end());
         }
 
         auto check_hsa_error = [](hsa_status_t s) {