diff --git a/task1/main.cpp b/task1/main.cpp
index 0a31d51..0a01ede 100644
--- a/task1/main.cpp
+++ b/task1/main.cpp
@@ -63,9 +63,9 @@ auto main(int argc, char *argv[]) -> int {
         fmt::print("Sorted {} entries within {} ms in sequential\n", dataset_seq.size(), delay_ms.count());
 
 
-        static constexpr int nthreads = 16;
+        const int nthreads = std::thread::hardware_concurrency();
         t1 = std::chrono::high_resolution_clock::now();
-        algo::MergeSort_v2::sort(dataset_par.begin(), dataset_par.end(), [](int32_t a, int32_t b) {
+        algo::MergeSort_mt::sort(dataset_par, [](int32_t a, int32_t b) {
             return (a > b);
         }, nthreads);
         t2 = std::chrono::high_resolution_clock::now();
diff --git a/task1/mergesort.h b/task1/mergesort.h
index ecd942a..c798079 100644
--- a/task1/mergesort.h
+++ b/task1/mergesort.h
@@ -4,6 +4,8 @@
 #include <algorithm>
 #include <thread>
 #include <mutex>
+#include <future>
+#include <ranges>
 
 namespace algo {
 
@@ -84,15 +86,42 @@ namespace algo {
     class MergeSort_v2 {
     private:
 
-        template<typename Iterator, typename Comparator>
+        template<typename Container, typename Comparator>
         static auto
-        mt_merge() {
+        mt_merge(Container left, Container right, Comparator cmp) -> Container {
+            //using Iterator = typename std::iterator_traits<Container>::value_type;
 
+            Container output;
+
+            auto lefti = left.begin();
+            auto righti = right.begin();
+
+            while (lefti < left.end() && righti < right.end()) {
+                if (cmp(*lefti, *righti)) {
+                    output.emplace_back(std::move(*lefti));
+                    lefti++;
+                } else {
+                    output.emplace_back(std::move(*righti));
+                    righti++;
+                }
+            }
+
+            while (lefti < left.end()) {
+                output.emplace_back(std::move(*lefti));
+                lefti++;
+            }
+            while (righti < right.end()) {
+                output.emplace_back(std::move(*righti));
+                righti++;
+            }
+
+            return output;
         }
 
         template<typename Iterator, typename Comparator>
         static auto
-        merge(Iterator start, Iterator middle, Iterator end, Comparator cmp, Iterator output_start, std::recursive_mutex &dataset_guard) -> void {
+        merge(Iterator start, Iterator middle, Iterator end, Comparator cmp, Iterator output_start,
+              std::recursive_mutex &dataset_guard) -> void {
             Iterator start_m = start;
             Iterator begin = output_start;
             Iterator start2 = middle + 1;
@@ -131,6 +160,69 @@ namespace algo {
 
         }
 
+        template<typename Container, typename Iterator, typename Comparator>
+        static auto mt_split(Container &output_vec, Iterator start, Iterator end, Comparator cmp, Iterator output_start,
+                             int &nthreads, std::recursive_mutex &dataset_guard, std::mutex &depth_guard) -> void {
+            Iterator mid = start;
+            Iterator to_start = output_start;
+
+            if (std::distance(start, end) < 1) {
+                return;
+
+            }
+
+            bool rem_threads;
+            {
+                std::lock_guard guard(depth_guard); //RAII guard
+                rem_threads = nthreads > 1;
+            }
+
+            if (rem_threads) {
+                {
+                    std::lock_guard guard(depth_guard); //RAII guard
+                    nthreads -= 2;
+                }
+
+                std::advance(mid, std::distance(start, end) / 2);
+                std::thread t1([&]() {
+                    mt_split(output_vec, start, mid, cmp, output_start, nthreads, dataset_guard, depth_guard);
+                });
+
+                std::advance(output_start, std::distance(start, mid + 1));
+                std::thread t2([&]() {
+                    mt_split(output_vec, mid + 1, end, cmp, output_start, nthreads, dataset_guard, depth_guard);
+                });
+
+
+                //merge everything together starting from the complete beginning
+                t1.join();
+                t2.join();
+                std::vector<int> left;
+                left.assign(start, mid);
+                std::vector<int> right;
+                right.assign(mid + 1, end);
+
+                mt_merge(left, right, cmp);
+            } else {
+                //move mid iterator litterally to the mid
+                std::advance(mid, std::distance(start, end) / 2);
+                //sort the first half within an recursion
+                mt_split(output_vec, start, mid, cmp, output_start, nthreads, dataset_guard, depth_guard);
+
+                //move output iterator
+                std::advance(output_start, std::distance(start, mid + 1));
+                //sort the second half within a recursion
+                mt_split(output_vec, mid + 1, end, cmp, output_start, nthreads, dataset_guard, depth_guard);
+
+                std::vector<int> left;
+                left.assign(start, mid);
+                std::vector<int> right;
+                right.assign(mid + 1, end);
+
+                mt_merge(left, right, cmp);
+            }
+        }
+
         template<typename Container, typename Iterator, typename Comparator>
         static auto ms_split(Container &output_vec, Iterator start, Iterator end, Comparator cmp, Iterator output_start,
                              int &nthreads, std::recursive_mutex &dataset_guard, std::mutex &depth_guard) -> void {
@@ -149,12 +241,16 @@ namespace algo {
                     std::advance(mid, std::distance(start, end) / 2);
                     //sort the first half within an recursion
 
-                    std::thread t1([&]() { ms_split(output_vec, start, mid, cmp, output_start, nthreads, dataset_guard, depth_guard); });
+                    std::thread t1([&]() {
+                        ms_split(output_vec, start, mid, cmp, output_start, nthreads, dataset_guard, depth_guard);
+                    });
 
                     //move output iteratoroutput_vec, start, mid, cmp, output_start
                     std::advance(output_start, std::distance(start, mid + 1));
                     //sort the second half within a recursion
-                    std::thread t2([&]() { ms_split(output_vec, mid + 1, end, cmp, output_start, nthreads, dataset_guard, depth_guard); });
+                    std::thread t2([&]() {
+                        ms_split(output_vec, mid + 1, end, cmp, output_start, nthreads, dataset_guard, depth_guard);
+                    });
 
 
                     //merge everything together starting from the complete beginning
@@ -186,8 +282,87 @@ namespace algo {
             std::recursive_mutex dataset_guard;
             std::mutex depth_guard;
 
-            ms_split(temporary_dataset, start, end - 1, cmp, temporary_dataset.begin(), nthreads, dataset_guard, depth_guard);
+            mt_split(temporary_dataset, start, end - 1, cmp, temporary_dataset.begin(), nthreads, dataset_guard,
+                     depth_guard);
         }
     };
 
+    class MergeSort_mt {
+
+        template<typename T, typename Comparator>
+        static auto
+        merge(typename std::vector<T>::iterator mid, std::vector<T> &left, std::vector<T> &right,
+              Comparator cmp, std::mutex &mut) {
+            {
+                std::lock_guard lock(mut);
+
+                auto l = left.begin();
+                auto r = right.begin();
+
+                while (l < left.end() && r < right.end()) {
+                    if (cmp(*l, *r)) {
+                        *mid = *l;
+                        l++;
+                    } else {
+                        *mid = *r;
+                        r++;
+                    }
+                    mid++;
+                }
+                while (l < left.end()) {
+                    *mid = *l;
+                    mid++;
+                    l++;
+                }
+                while (r < right.end()) {
+                    *mid = *r;
+                    mid++;
+                    r++;
+                }
+            }
+        }
+
+        template<typename T, typename Comparator>
+        static auto split(std::vector<T> &data, std::vector<T> &outdata, Comparator cmp, int depth, int &num_threads, std::mutex &mut) {
+            if (data.size() <= 1) {
+                return;
+            }
+
+            auto mid = data.begin();
+            std::advance(mid, std::distance(data.begin(), data.end()) / 2);
+
+
+
+
+            std::vector<T> left(std::make_move_iterator(data.begin()), std::make_move_iterator(mid));
+            std::vector<T> right(std::make_move_iterator(mid), std::make_move_iterator(data.end()));
+
+            if (depth < num_threads) {
+                std::thread left_thread([&]() { split(left, outdata, cmp, depth + 1, num_threads, mut); });
+                std::thread right_thread([&]() { split(right, outdata, cmp, depth + 1, num_threads, mut); });
+
+                left_thread.join();
+                right_thread.join();
+            } else {
+                split(left, outdata, cmp, depth + 1, num_threads, mut);
+                split(right, outdata, cmp, depth + 1, num_threads, mut);
+            }
+
+
+            mid = outdata.begin();
+            std::advance(mid, std::distance(outdata.begin(), outdata.end())/2);
+            merge(mid, left, right, cmp, mut);
+        }
+
+    public:
+        template<typename T, typename Comparator>
+        static auto
+        sort(std::vector<T> &data, Comparator cmp, int num_threads = std::thread::hardware_concurrency()) -> void {
+            std::vector<T> local_result = data;
+            std::mutex local_result_lock;
+
+            split(data, local_result, cmp, 0, num_threads, local_result_lock);
+            data = local_result;
+        }
+    };
 }
\ No newline at end of file