diff --git a/.github/workflows/ci_misc.yml b/.github/workflows/ci_misc.yml
index 0cd25de316..dbbce8fdcd 100644
--- a/.github/workflows/ci_misc.yml
+++ b/.github/workflows/ci_misc.yml
@@ -37,6 +37,13 @@ jobs:
             build_type: Release
             test_threads: 1 # snippets create and delete files and some separate tests create/delete the same files
 
+          - name: "Performance clang17 libc++"
+            compiler: "clang-17"
+            build: performance
+            build_type: Release
+            test_threads: 2
+            cxx_flags: "-stdlib=libc++"
+
           - name: "Performance gcc11"
             compiler: "gcc-11"
             build: performance
diff --git a/test/include/seqan3/test/performance/units.hpp b/test/include/seqan3/test/performance/units.hpp
index ed40f0dc46..200c046410 100644
--- a/test/include/seqan3/test/performance/units.hpp
+++ b/test/include/seqan3/test/performance/units.hpp
@@ -38,12 +38,14 @@ inline benchmark::Counter bytes_per_second(size_t bytes)
 template <typename sequences_range_t>
 inline size_t pairwise_cell_updates(sequences_range_t const & sequences_range, [[maybe_unused]] auto && align_cfg)
 {
+    using config_t = std::remove_cvref_t<decltype(align_cfg)>;
+
     auto count_cells = [&](auto && seq1, auto && seq2)
     {
         size_t const columns = std::ranges::size(seq1) + 1;
         size_t const rows = std::ranges::size(seq2) + 1;
 
-        if constexpr (align_cfg.template exists<seqan3::align_cfg::band_fixed_size>())
+        if constexpr (config_t::template exists<seqan3::align_cfg::band_fixed_size>())
         {
             using std::get;
             auto const band_cfg = get<seqan3::align_cfg::band_fixed_size>(align_cfg);
diff --git a/test/performance/io/format_vienna_benchmark.cpp b/test/performance/io/format_vienna_benchmark.cpp
index 38e1189102..3c83c48579 100644
--- a/test/performance/io/format_vienna_benchmark.cpp
+++ b/test/performance/io/format_vienna_benchmark.cpp
@@ -26,18 +26,18 @@
 
 inline constexpr size_t iterations_per_run = 1024;
 
-inline std::string const header{"seq foobar blobber"};
-inline auto const rna_sequence = seqan3::test::generate_sequence<seqan3::rna4>(474, 0, 0);
-auto const sequence = rna_sequence | seqan3::views::to_char | seqan3::ranges::to<std::string>();
+static std::string const header{"seq foobar blobber"};
+static inline auto const rna_sequence = seqan3::test::generate_sequence<seqan3::rna4>(474, 0, 0);
+static auto const sequence = rna_sequence | seqan3::views::to_char | seqan3::ranges::to<std::string>();
 
-inline std::string const structure{"(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
+static std::string const structure{"(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
                                    "(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
                                    "(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
                                    "(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
                                    "(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."
                                    "(((((((..((((........)))).((((.........)))).....(((((.......))))))))))))......."};
 
-inline std::string const vienna_file = []()
+static std::string const vienna_file = []()
 {
     std::string file{};
     for (size_t idx = 0; idx < iterations_per_run; idx++)
diff --git a/test/performance/io/lowlevel_stream_input_benchmark.cpp b/test/performance/io/lowlevel_stream_input_benchmark.cpp
index 5035c41b76..8f00da87dc 100644
--- a/test/performance/io/lowlevel_stream_input_benchmark.cpp
+++ b/test/performance/io/lowlevel_stream_input_benchmark.cpp
@@ -47,41 +47,49 @@ void read_all(benchmark::State & state)
     }
 
     /* start benchmark */
-    char c{}; // prevents optimisation
     if constexpr (id == tag::std_stream_it)
     {
         for (auto _ : state)
         {
+            char c{};
             std::ifstream s{filename, std::ios::binary};
             std::istream_iterator<char> it{s};
             std::istream_iterator<char> e{};
 
             for (; it != e; ++it)
                 c += *it;
+
+            benchmark::DoNotOptimize(c);
         }
     }
     else if constexpr (id == tag::std_streambuf_it)
     {
         for (auto _ : state)
         {
+            char c{};
             std::ifstream s{filename, std::ios::binary};
             std::istreambuf_iterator<char> it{s};
             std::istreambuf_iterator<char> e{};
 
             for (; it != e; ++it)
                 c += *it;
+
+            benchmark::DoNotOptimize(c);
         }
     }
     else if constexpr (id == tag::seqan3_streambuf_it)
     {
         for (auto _ : state)
         {
+            char c{};
             std::ifstream s{filename, std::ios::binary};
             seqan3::detail::fast_istreambuf_iterator<char> it{*s.rdbuf()};
             std::default_sentinel_t e{};
 
             for (; it != e; ++it)
                 c += *it;
+
+            benchmark::DoNotOptimize(c);
         }
     }
 #ifdef SEQAN3_HAS_SEQAN2
@@ -89,11 +97,14 @@ void read_all(benchmark::State & state)
     {
         for (auto _ : state)
         {
+            char c{};
             std::ifstream s{filename, std::ios::binary};
             auto it = seqan2::Iter<std::ifstream, seqan2::StreamIterator<seqan2::Input>>{s};
 
             for (; !seqan2::atEnd(it); ++it)
                 c += *it;
+
+            benchmark::DoNotOptimize(c);
         }
     }
 #endif
diff --git a/test/performance/range/container_assignment_benchmark.cpp b/test/performance/range/container_assignment_benchmark.cpp
index 7208781a31..e22ad21e64 100644
--- a/test/performance/range/container_assignment_benchmark.cpp
+++ b/test/performance/range/container_assignment_benchmark.cpp
@@ -86,7 +86,7 @@ static constexpr void resize(container_t & container,
 }
 #endif // SEQAN3_HAS_SEQAN2
 
-template <tag id, template <typename> typename container_t, typename alphabet_t>
+template <tag id, template <typename, typename...> typename container_t, typename alphabet_t, typename... args>
 static void assign(benchmark::State & state)
 {
     auto random_sequence = []() constexpr
@@ -99,10 +99,10 @@ static void assign(benchmark::State & state)
 #endif // SEQAN3_HAS_SEQAN2
     }();
 
-    container_t<alphabet_t> from{};
+    container_t<alphabet_t, args...> from{};
     resize(from, vector_size);
     std::copy(std::ranges::begin(random_sequence), std::ranges::end(random_sequence), std::ranges::begin(from));
-    container_t<alphabet_t> to{};
+    container_t<alphabet_t, args...> to{};
     resize(to, vector_size);
     assignment_functor<id> fn{};
 
diff --git a/test/performance/range/container_push_back_benchmark.cpp b/test/performance/range/container_push_back_benchmark.cpp
index 32470ac9b8..b49bbdeada 100644
--- a/test/performance/range/container_push_back_benchmark.cpp
+++ b/test/performance/range/container_push_back_benchmark.cpp
@@ -29,14 +29,14 @@ using small_vec = seqan3::small_vector<t, 10'000>;
 //  push_back
 // ============================================================================
 
-template <template <typename> typename container_t, typename alphabet_t>
+template <template <typename, typename...> typename container_t, typename alphabet_t, typename... args>
 void push_back(benchmark::State & state)
 {
     alphabet_t letter{};
 
     for (auto _ : state)
     {
-        container_t<alphabet_t> container;
+        container_t<alphabet_t, args...> container;
         for (size_t i = 0; i < 10'000; ++i)
             container.push_back(letter);
         benchmark::DoNotOptimize(letter = container.back());
diff --git a/test/performance/range/container_seq_read_benchmark.cpp b/test/performance/range/container_seq_read_benchmark.cpp
index 6a3e41fc5b..aa7d57d978 100644
--- a/test/performance/range/container_seq_read_benchmark.cpp
+++ b/test/performance/range/container_seq_read_benchmark.cpp
@@ -30,17 +30,20 @@ using small_vec = seqan3::small_vector<t, 10'000>;
 //  sequential_read
 // ============================================================================
 
-template <template <typename> typename container_t, typename alphabet_t, bool const_qualified = false>
+template <template <typename, typename...> typename container_t,
+          typename alphabet_t,
+          bool const_qualified = false,
+          typename... args>
 void sequential_read(benchmark::State & state)
 {
-    container_t<alphabet_t> container = []()
+    container_t<alphabet_t, args...> container = []()
     {
         auto container = seqan3::test::generate_sequence<alphabet_t>(10'000, 0, 0);
-        return container_t<alphabet_t>(container.begin(), container.end());
+        return container_t<alphabet_t, args...>(container.begin(), container.end());
     }();
 
-    using container_reference_t =
-        std::conditional_t<const_qualified, container_t<alphabet_t> const &, container_t<alphabet_t> &>;
+    using container_reference_t = std::
+        conditional_t<const_qualified, container_t<alphabet_t, args...> const &, container_t<alphabet_t, args...> &>;
 
     container_reference_t container_reference{container};
 
diff --git a/test/performance/range/container_seq_write_benchmark.cpp b/test/performance/range/container_seq_write_benchmark.cpp
index 193ab6965d..f7d9c77663 100644
--- a/test/performance/range/container_seq_write_benchmark.cpp
+++ b/test/performance/range/container_seq_write_benchmark.cpp
@@ -30,13 +30,13 @@ using small_vec = seqan3::small_vector<t, 10'000>;
 //  sequential_write
 // ============================================================================
 
-template <template <typename> typename container_t, typename alphabet_t>
+template <template <typename, typename...> typename container_t, typename alphabet_t, typename... args>
 void sequential_write(benchmark::State & state)
 {
-    container_t<alphabet_t> container = []()
+    container_t<alphabet_t, args...> container = []()
     {
         auto container = seqan3::test::generate_sequence<alphabet_t>(10'000, 0, 0);
-        return container_t<alphabet_t>(container.begin(), container.end());
+        return container_t<alphabet_t, args...>(container.begin(), container.end());
     }();
 
     alphabet_t letter{};
diff --git a/test/performance/search/views/view_minimiser_hash_benchmark.cpp b/test/performance/search/views/view_minimiser_hash_benchmark.cpp
index db4b69f1f5..351a94abc4 100644
--- a/test/performance/search/views/view_minimiser_hash_benchmark.cpp
+++ b/test/performance/search/views/view_minimiser_hash_benchmark.cpp
@@ -98,7 +98,7 @@ void compute_minimisers(benchmark::State & state)
             // Use random seed to randomise order on forward strand.
             auto forward = seq | seqan3::views::kmer_hash(shape)
                          | std::views::transform(
-                               [seed](uint64_t const i)
+                               [](uint64_t const i)
                                {
                                    return i ^ seed;
                                });
@@ -107,7 +107,7 @@ void compute_minimisers(benchmark::State & state)
                          | std::views::reverse             // Reverse order.
                          | seqan3::views::kmer_hash(shape) // Get hash values.
                          | std::views::transform(
-                               [seed](uint64_t const i)
+                               [](uint64_t const i)
                                {
                                    return i ^ seed;
                                })               // Randomise.
@@ -131,7 +131,7 @@ void compute_minimisers(benchmark::State & state)
             while (subrange_end != end(both))
             {
                 ++subrange_end; // Extends the subrange to `w - shape.size()+1`
-                auto h = *std::min_element(subrange_begin, subrange_end);
+                auto h = *std::ranges::min_element(subrange_begin, subrange_end);
 
                 ++subrange_begin; // Move the beginning one forward
                 benchmark::DoNotOptimize(sum += h);
diff --git a/test/performance/utility/simd/views/iota_simd_benchmark.cpp b/test/performance/utility/simd/views/iota_simd_benchmark.cpp
index 1d23458669..13c24152cc 100644
--- a/test/performance/utility/simd/views/iota_simd_benchmark.cpp
+++ b/test/performance/utility/simd/views/iota_simd_benchmark.cpp
@@ -88,7 +88,7 @@ struct for_loop_with_simd_add_benchmark
     void operator()(index_simd_t & count) const
     {
         index_simd_t simd_index{};
-        for (size_t index = 0; index < end_index; ++index, ++simd_index)
+        for (size_t index = 0; index < end_index; ++index, simd_index += 1)
             count += simd_index;
     }
 };
diff --git a/test/seqan3-test.cmake b/test/seqan3-test.cmake
index f19dc33f4a..1d383783bd 100644
--- a/test/seqan3-test.cmake
+++ b/test/seqan3-test.cmake
@@ -80,6 +80,8 @@ endif ()
 if (NOT TARGET seqan3::test::performance)
     add_library (seqan3_test_performance INTERFACE)
     target_link_libraries (seqan3_test_performance INTERFACE "seqan3::test" "benchmark_main" "benchmark")
+    # std::views::join is experimental in libc++
+    target_compile_definitions (seqan3_test_performance INTERFACE _LIBCPP_ENABLE_EXPERIMENTAL)
 
     if (SEQAN3_BENCHMARK_ALIGN_LOOPS)
         target_compile_options (seqan3_test_performance INTERFACE "-falign-loops=32")