Merge branch 'stclib:master' into modifiedHEADmodified

7 files changed, 192 insertions, 152 deletions

diff --git a/misc/benchmarks/various/csort_bench.c b/misc/benchmarks/various/csort_bench.c
index d5d7fa7c..793a0503 100644
--- a/misc/benchmarks/various/csort_bench.c
+++ b/misc/benchmarks/various/csort_bench.c
@@ -5,8 +5,12 @@
 #ifdef __cplusplus
   #include <algorithm>
 #endif
-#define i_val int
-#include <stc/algo/csort.h>
+#define NDEBUG
+#define i_type Ints
+#define i_key int
+#define i_more
+#include <stc/cvec.h>
+#include <stc/algo/sort.h>
 
 #define ROTL(d,bits) ((d<<(bits)) | (d>>(8*sizeof(d)-(bits))))
 uint64_t romutrio(uint64_t s[3]) {
@@ -21,14 +25,14 @@ static int cmp_int(const void* a, const void* b) {
     return c_default_cmp((const int*)a, (const int*)b);
 }
 
-void testsort(int *a, int size, const char *desc) {
+void testsort(Ints *a, int size, const char *desc) {
     clock_t t = clock();
 #ifdef __cplusplus
-    printf("std::sort: "); std::sort(a, a + size);
+    printf("std::sort: "); std::sort(a->data, a->data + size);
 #elif defined QSORT
-    printf("qsort: "); qsort(a, size, sizeof *a, cmp_int);
+    printf("qsort: "); qsort(a->data, size, sizeof *a->data, cmp_int);
 #else
-    printf("stc_sort: "); csort_int(a, size);
+    printf("STC sort_n: "); Ints_sort_n(a, size);
 #endif
     t = clock() - t;
 
@@ -41,27 +45,27 @@ int main(int argc, char *argv[]) {
     size_t i, size = argc > 1 ? strtoull(argv[1], NULL, 0) : 10000000;
     uint64_t s[3] = {123456789, 3456789123, 789123456};
 
-    int32_t *a = (int32_t*)malloc(sizeof(*a) * size);
-    if (!a) return -1;
+    Ints a = Ints_with_capacity(size);
 
     for (i = 0; i < size; i++)
-        a[i] = romutrio(s) & (1U << 30) - 1;
-    testsort(a, size, "random");
+        Ints_push(&a, romutrio(s) & (1U << 30) - 1);
+    testsort(&a, size, "random");
     for (i = 0; i < 20; i++)
-        printf(" %d", (int)a[i]);
+        printf(" %d", (int)*Ints_at(&a, i));
     puts("");
     for (i = 0; i < size; i++)
-        a[i] = i;
-    testsort(a, size, "sorted");
+        *Ints_at_mut(&a, i) = i;
+    testsort(&a, size, "sorted");
     for (i = 0; i < size; i++)
-        a[i] = size - i;
-    testsort(a, size, "reverse sorted");
+        *Ints_at_mut(&a, i) = size - i;
+    testsort(&a, size, "reverse sorted");
     for (i = 0; i < size; i++)
-        a[i] = 126735;
-    testsort(a, size, "constant");
+        *Ints_at_mut(&a, i) = 126735;
+    testsort(&a, size, "constant");
     for (i = 0; i < size; i++)
-        a[i] = i + 1;
-    a[size - 1] = 0;
-    testsort(a, size, "rotated");
-    free(a);
+        *Ints_at_mut(&a, i) = i + 1;
+    *Ints_at_mut(&a, size - 1) = 0;
+    testsort(&a, size, "rotated");
+
+    Ints_drop(&a);
 }
diff --git a/misc/benchmarks/various/cspan_bench.c b/misc/benchmarks/various/cspan_bench.c
index 589df13a..bfc0ead3 100644
--- a/misc/benchmarks/various/cspan_bench.c
+++ b/misc/benchmarks/various/cspan_bench.c
@@ -1,4 +1,5 @@
-#define STC_NDEBUG
+// ref: https://stackoverflow.com/questions/74382366/why-is-iterating-over-stdrangesviewsjoin-so-slow
+#define NDEBUG
 #include <stc/cspan.h>
 #include <stdio.h>
 #include <time.h>
@@ -11,9 +12,9 @@ enum {
     ny = 64,
     nz = 64
 };
+// subspan 15x5x10:
 int lx = 15, ly = 10, lz = 5;
-int hx = 20, hy = 15, hz = 15;
-intptr_t n = 1000000;
+int hx = 30, hy = 15, hz = 15;
 
 // define the contents of two nx x ny x nz arrays in and out
 double Vout[nx * ny * nz];
@@ -21,36 +22,15 @@ double Vin[nx * ny * nz]; //, 1.23;
 
 // define some slice indices for each dimension
 
-static void MDRanges_setup(intptr_t state)
-{
-    double sum = 0;
-    clock_t t = clock();
-
-    for (intptr_t s = 0; s < state; ++s)
-    {
-        MD3 r_in = cspan_md(Vin, nx, ny, nz);
-        MD3 r_out = cspan_md(Vout, nx, ny, nz);
-
-        r_in = cspan_slice(MD3, &r_in, {lx, hx}, {ly, hy}, {lz, hz});
-        r_out = cspan_slice(MD3, &r_out, {lx, hx}, {ly, hy}, {lz, hz});
-        MD3_iter i = MD3_begin(&r_in); // can be iterated "flat".
-        MD3_iter o = MD3_begin(&r_out);
-        sum += Vin[s % nx];
-    }
-    t = clock() - t;
-    printf("setup: %.1f ms, %f\n", 1000.0f * t / CLOCKS_PER_SEC, sum);
-}
-
-static void TraditionalForLoop(intptr_t state)
+static void Traditional_for_loop(intptr_t n)
 {
     clock_t t = clock();
     double sum = 0;
 
-    for (int s = 0; s < state; ++s) {
+    for (int s = 0; s < n; ++s) {
         for (int x = lx; x < hx; ++x) {
             for (int y = ly; y < hy; ++y) {
-                for (int z = lz; z < hz; ++z)
-                {
+                for (int z = lz; z < hz; ++z) {
                     double d = Vin[nz*(ny*x + y) + z];
                     Vout[nz*(ny*x + y) + z] += d;
                     sum += d;
@@ -59,68 +39,65 @@ static void TraditionalForLoop(intptr_t state)
         }
     }
     t = clock() - t;
-    printf("forloop: %.1f ms, %f\n", 1000.0f * t / CLOCKS_PER_SEC, sum);
+    printf("forloop : %.1f ms, %f\n", 1000.0f*t / CLOCKS_PER_SEC, sum);
 }
 
-static void MDRanges_nested_loop(intptr_t state)
+static void MDRanges_loop_over_joined(intptr_t n)
 {
+    clock_t t = clock();
     MD3 r_in = cspan_md(Vin, nx, ny, nz);
     MD3 r_out = cspan_md(Vout, nx, ny, nz);
     r_in = cspan_slice(MD3, &r_in, {lx, hx}, {ly, hy}, {lz, hz});
     r_out = cspan_slice(MD3, &r_out, {lx, hx}, {ly, hy}, {lz, hz});
-
-    // C++23: for (auto [o, i] : std::views::zip(flat(r_out), flat(r_in))) { o = i; }
-    clock_t t = clock();
     double sum = 0;
 
-    for (intptr_t s = 0; s < state; ++s) {
-        for (int x = 0; x < r_in.shape[0]; ++x) {
-            for (int y = 0; y < r_in.shape[1]; ++y) {
-                for (int z = 0; z < r_in.shape[2]; ++z)
-                {
-                    double d = *cspan_at(&r_in, x, y, z);
-                    *cspan_at(&r_out, x, y, z) += d;
-                    sum += d;
-                }
-            }
+    for (intptr_t s = 0; s < n; ++s) {
+        MD3_iter i = MD3_begin(&r_in);
+        MD3_iter o = MD3_begin(&r_out);
+
+        for (; i.ref; MD3_next(&i), MD3_next(&o))
+        {
+            *o.ref += *i.ref;
+            sum += *i.ref;
         }
     }
     t = clock() - t;
-    printf("nested: %.1f ms, %f\n", 1000.0f * t / CLOCKS_PER_SEC, sum);
+    printf("joined  : %.1f ms, %f\n", 1000.0f*t / CLOCKS_PER_SEC, sum);
 }
 
-static void MDRanges_loop_over_joined(intptr_t state)
+static void MDRanges_nested_loop(intptr_t n)
 {
+    clock_t t = clock();
     MD3 r_in = cspan_md(Vin, nx, ny, nz);
     MD3 r_out = cspan_md(Vout, nx, ny, nz);
     r_in = cspan_slice(MD3, &r_in, {lx, hx}, {ly, hy}, {lz, hz});
     r_out = cspan_slice(MD3, &r_out, {lx, hx}, {ly, hy}, {lz, hz});
-
-    // C++23: for (auto [o, i] : std::views::zip(flat(r_out), flat(r_in))) { o = i; }
     double sum = 0;
-    clock_t t = clock();
-
-    for (intptr_t s = 0; s < state; ++s) {
-        MD3_iter i = MD3_begin(&r_in);
-        MD3_iter o = MD3_begin(&r_out);
 
-        for (; i.ref; MD3_next(&i), MD3_next(&o))
-        {
-            *o.ref += *i.ref;
-            sum += *i.ref;
+    for (intptr_t s = 0; s < n; ++s) {
+        for (int x = 0; x < r_in.shape[0]; ++x) {
+            for (int y = 0; y < r_in.shape[1]; ++y) {
+                for (int z = 0; z < r_in.shape[2]; ++z)
+                {
+                    double d = *cspan_at(&r_in, x,y,z);
+                    *cspan_at(&r_out, x,y,z) += d;
+                    sum += d;
+                }
+            }
         }
     }
     t = clock() - t;
-    printf("joined: %.1f ms, %f\n", 1000.0f * t / CLOCKS_PER_SEC, sum);
+    printf("nested  : %.1f ms, %f\n", 1000.0f*t / CLOCKS_PER_SEC, sum);
 }
 
-int main()
+
+int main(void)
 {
+    intptr_t n = 100000;
     for (int i = 0; i < nx * ny * nz; ++i)
         Vin[i] = i + 1.23;
 
-    MDRanges_setup(n);
-    TraditionalForLoop(n);
-    MDRanges_nested_loop(n);
+    Traditional_for_loop(n);
     MDRanges_loop_over_joined(n);
+    MDRanges_nested_loop(n);
 }
diff --git a/misc/benchmarks/various/prng_bench.cpp b/misc/benchmarks/various/prng_bench.cpp
index 234e3805..45c14d18 100644
--- a/misc/benchmarks/various/prng_bench.cpp
+++ b/misc/benchmarks/various/prng_bench.cpp
@@ -66,7 +66,7 @@ uint32_t pcg32(uint32_t s[2]) {
 }
 
 
-/* xoshiro128+  */
+/* xo(ro)shiro  */
 
 uint64_t xoroshiro128plus(uint64_t s[2]) {
     const uint64_t s0 = s[0];
@@ -80,9 +80,6 @@ uint64_t xoroshiro128plus(uint64_t s[2]) {
     return result;
 }
 
-
-/* xoshiro256**  */
-
 static inline uint64_t xoshiro256starstar(uint64_t s[4]) {
     const uint64_t result = rotl64(s[1] * 5, 7) * 9;
     const uint64_t t = s[1] << 17;
@@ -95,7 +92,7 @@ static inline uint64_t xoshiro256starstar(uint64_t s[4]) {
     return result;
 }
 
-// wyrand - 2020-12-07
+/* wyrand - 2020-12-07 */
 static inline void _wymum(uint64_t *A, uint64_t *B){
 #if defined(__SIZEOF_INT128__)
     __uint128_t r = *A; r *= *B;
@@ -136,44 +133,44 @@ int main(void)
     for (size_t ti = 0; ti < 2; ti++) {
         init_state(rng.state, 12345123);
         cout << endl << "ROUND " << ti+1 << " ---------" << endl;
-
+/*
         beg = clock();
         for (size_t i = 0; i < N; i++)
-            recipient[i] = romu_trio(rng.state);
+            recipient[i] = sfc32((uint32_t *)rng.state);
         end = clock();
-        cout << "romu_trio:\t"
+        cout << "sfc32:\t\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
 
         beg = clock();
         for (size_t i = 0; i < N; i++)
-            recipient[i] = wyrand64(rng.state);
+            recipient[i] = stc32((uint32_t *)rng.state);
         end = clock();
-        cout << "wyrand64:\t"
+        cout << "stc32:\t\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
 
         beg = clock();
         for (size_t i = 0; i < N; i++)
-            recipient[i] = sfc32((uint32_t *)rng.state);
+            recipient[i] = pcg32((uint32_t *)rng.state);
         end = clock();
-        cout << "sfc32:\t\t"
+        cout << "pcg32:\t\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
-
+*/
         beg = clock();
         for (size_t i = 0; i < N; i++)
-            recipient[i] = stc32((uint32_t *)rng.state);
+            recipient[i] = romu_trio(rng.state);
         end = clock();
-        cout << "stc32:\t\t"
+        cout << "romu_trio:\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
 
         beg = clock();
         for (size_t i = 0; i < N; i++)
-            recipient[i] = pcg32((uint32_t *)rng.state);
+            recipient[i] = wyrand64(rng.state);
         end = clock();
-        cout << "pcg32:\t\t"
+        cout << "wyrand64:\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
 
@@ -189,7 +186,7 @@ int main(void)
         for (size_t i = 0; i < N; i++)
             recipient[i] = crand_u64(&rng);
         end = clock();
-        cout << "stc64:\t\t"
+        cout << "crand64:\t"
              << (float(end - beg) / CLOCKS_PER_SEC)
              << "s: " << recipient[312] << endl;
 
diff --git a/misc/benchmarks/various/rust_cmap.c b/misc/benchmarks/various/rust_cmap.c
index abdb42b0..97047e0b 100644
--- a/misc/benchmarks/various/rust_cmap.c
+++ b/misc/benchmarks/various/rust_cmap.c
@@ -22,7 +22,7 @@ uint64_t romu_trio(uint64_t s[3]) {
    return xp;
 }
 
-int main()
+int main(void)
 {
     cmap_u64 m = {0};
 
diff --git a/misc/benchmarks/various/sso_bench.cpp b/misc/benchmarks/various/sso_bench.cpp
index 993ff1bb..244c1291 100644
--- a/misc/benchmarks/various/sso_bench.cpp
+++ b/misc/benchmarks/various/sso_bench.cpp
@@ -3,65 +3,82 @@
 #include <chrono>
 
 #include <stc/crand.h>
+#define i_static
 #include <stc/cstr.h>
 
 #define i_type StcVec
 #define i_val_str
 #include <stc/cstack.h>
 
-#define i_type StcSet
-#define i_val_str
-#include <stc/csset.h>
-
 #include <vector>
 using StdVec = std::vector<std::string>;
-#include <set>
-using StdSet = std::set<std::string>;
 
 
-static const int BENCHMARK_SIZE = 2000000;
-static const int MAX_STRING_SIZE = 50;
+#include <unordered_set>
+#include "../external/ankerl/robin_hood.h"
+
+struct string_hash {
+  using is_transparent = void;
+  [[nodiscard]] size_t operator()(const char *txt) const {
+    return std::hash<std::string_view>{}(txt);
+  }
+  [[nodiscard]] size_t operator()(std::string_view txt) const {
+    return std::hash<std::string_view>{}(txt);
+  }
+  [[nodiscard]] size_t operator()(const std::string &txt) const {
+    return std::hash<std::string>{}(txt);
+  }
+};
+using StdSet = robin_hood::unordered_flat_set<std::string, string_hash, std::equal_to<>>;
+//using StdSet = std::unordered_set<std::string>;
+
+#define i_type StcSet
+#define i_val_str
+//#define i_hash(txtp) std::hash<std::string_view>{}(*txtp)
+#include <stc/cset.h>
+
+
+static const int BENCHMARK_SIZE = 250000;
+static const int MAX_STRING_SIZE = 100;
 static const char CHARS[] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz=+-";
 using time_point = std::chrono::high_resolution_clock::time_point;
 
 
-static inline std::string randomString_STD(int strsize) {
-    std::string s(strsize, 0);
-    char* p = &s[0];
+static inline const char* randomString(int strsize) {
+    static char str[256];
     union { uint64_t u8; uint8_t b[8]; } r;
     for (int i = 0; i < strsize; ++i) {
         if ((i & 7) == 0) r.u8 = crand() & 0x3f3f3f3f3f3f3f3f;
-        p[i] = CHARS[r.b[i & 7]];
+        str[i] = CHARS[r.b[i & 7]];
     }
-    return s;
+    str[strsize] = 0;
+    return str;
 }
 
-static inline cstr randomString_STC(int strsize) {
-    cstr s = cstr_with_size(strsize, 0);
-    char* p = cstr_data(&s);
-    union { uint64_t u8; uint8_t b[8]; } r;
-    for (int i = 0; i < strsize; ++i) {
-        if ((i & 7) == 0) r.u8 = crand() & 0x3f3f3f3f3f3f3f3f;
-        p[i] = CHARS[r.b[i & 7]];
-    }
-    return s;
+
+
+static inline void addRandomString(StdVec& vec, const char* str) {
+    vec.push_back(str);
 }
 
+static inline void addRandomString(StcVec& vec, const char* str) {
+    StcVec_emplace(&vec, str);
+}
 
-void addRandomString(StdVec& vec, int strsize) {
-    vec.push_back(std::move(randomString_STD(strsize)));
+static inline void addRandomString(StdSet& set, const char* str) {
+    set.insert(str);
 }
 
-void addRandomString(StcVec& vec, int strsize) {
-    StcVec_push(&vec, randomString_STC(strsize));
+static inline void addRandomString(StcSet& set, const char* str) {
+    StcSet_emplace(&set, str);
 }
 
-void addRandomString(StdSet& set, int strsize) {
-    set.insert(std::move(randomString_STD(strsize)));
+static inline bool getRandomString(const StdSet& set, const char* str) {
+    return set.find(str) != set.end();
 }
 
-void addRandomString(StcSet& set, int strsize) {
-    StcSet_insert(&set, randomString_STC(strsize));
+static inline bool getRandomString(const StcSet& set, const char* str) {
+    return StcSet_contains(&set, str);
 }
 
 
@@ -70,7 +87,7 @@ int benchmark(C& container, const int n, const int strsize) {
     time_point t1 = std::chrono::high_resolution_clock::now();
 
     for (int i = 0; i < n; i++)
-        addRandomString(container, strsize);
+        addRandomString(container, randomString(strsize));
 
     time_point t2 = std::chrono::high_resolution_clock::now();
     const auto duration = std::chrono::duration_cast<std::chrono::milliseconds>(t2 - t1).count();
@@ -78,9 +95,25 @@ int benchmark(C& container, const int n, const int strsize) {
     return (int)duration;
 }
 
+template <class C>
+int benchmark_lookup(C& container, const int n, const int strsize) {
+    for (int i = 0; i < n; i++)
+        addRandomString(container, randomString(strsize));
+
+    time_point t1 = std::chrono::high_resolution_clock::now();
+    int found = 0;
+    for (int i = 0; i < n; i++)
+        found += (int)getRandomString(container, randomString(strsize));
+
+    time_point t2 = std::chrono::high_resolution_clock::now();
+    const auto duration = std::chrono::duration_cast<std::chrono::milliseconds>(t2 - t1).count();
+    std::cerr << (strsize ? strsize : 32) << "\t" << duration << '\t' << found;
+    return (int)duration;
+}
 
-int main() {
-    uint64_t seed = 4321;
+#include <time.h>
+int main(void) {
+    uint64_t seed = time(NULL); // 4321;
     int sum, n;
 
     // VECTOR WITH STRINGS
@@ -88,48 +121,75 @@ int main() {
     csrand(seed);
     sum = 0, n = 0;
     std::cerr << "\nstrsize\tmsecs\tstd::vector<std::string>, size=" << BENCHMARK_SIZE << "\n";
-    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 2) {
+    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 4) {
         StdVec vec; vec.reserve(BENCHMARK_SIZE);
         sum += benchmark(vec, BENCHMARK_SIZE, strsize), ++n;
         std::cout << '\t' << vec.front() << '\n';
     }
-    std::cout << "Avg:\t" << sum/n << '\n';
+    std::cout << "Avg:\t" << sum/n << "ms\n";
 
     csrand(seed);
     sum = 0, n = 0;
     std::cerr << "\nstrsize\tmsecs\tcvec<cstr>, size=" << BENCHMARK_SIZE << "\n";
-    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 2) {
+    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 4) {
         StcVec vec = StcVec_with_capacity(BENCHMARK_SIZE);
         sum += benchmark(vec, BENCHMARK_SIZE, strsize), ++n;
         std::cout << '\t' << cstr_str(&vec.data[0]) << '\n';
         StcVec_drop(&vec);
     }
-    std::cout << "Avg:\t" << sum/n << '\n';
+    std::cout << "Avg:\t" << sum/n << "ms\n";
+
+    // INSERT: SORTED SET WITH STRINGS
+
+    csrand(seed);
+    sum = 0, n = 0;
+    std::cerr << "\nstrsize\tmsecs\tinsert: robin_hood::unordered_flat_set<std::string>, size=" << BENCHMARK_SIZE/2 << "\n";
+    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 4) {
+        StdSet set; set.reserve(BENCHMARK_SIZE/2);
+        sum += benchmark(set, BENCHMARK_SIZE/2, strsize), ++n;
+        std::cout << '\t' << *set.begin() << '\n';
+    }
+    std::cout << "Avg:\t" << sum/n << "ms\n";
 
-    // SORTED SET WITH STRINGS
     
     csrand(seed);
     sum = 0, n = 0;
-    std::cerr << "\nstrsize\tmsecs\tstd::set<std::string>, size=" << BENCHMARK_SIZE/16 << "\n";
+    std::cerr << "\nstrsize\tmsecs\tinsert: cset<cstr>, size=" << BENCHMARK_SIZE/2 << "\n";
+    for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 4) {
+        StcSet set = StcSet_with_capacity(BENCHMARK_SIZE/2);
+        sum += benchmark(set, BENCHMARK_SIZE/2, strsize), ++n;
+        std::cout << '\t' << cstr_str(StcSet_begin(&set).ref) << '\n';
+        StcSet_drop(&set);
+    }
+    std::cout << "Avg:\t" << sum/n << "ms\n";
+
+    // LOOKUP: SORTED SET WITH STRINGS
+
+    csrand(seed);
+    sum = 0, n = 0;
+    std::cerr << "\nstrsize\tmsecs\tfind: robin_hood::unordered_flat_set<std::string>, size=" << BENCHMARK_SIZE/2 << "\n";
     for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 2) {
-        StdSet set;
-        sum += benchmark(set, BENCHMARK_SIZE/16, strsize), ++n;
+        StdSet set; set.reserve(BENCHMARK_SIZE/2);
+        sum += benchmark_lookup(set, BENCHMARK_SIZE/2, strsize), ++n;
         std::cout << '\t' << *set.begin() << '\n';
     }
-    std::cout << "Avg:\t" << sum/n << '\n';
+    std::cout << "Avg:\t" << sum/n << "ms\n";
 
     csrand(seed);
     sum = 0, n = 0;
-    std::cerr << "\nstrsize\tmsecs\tcsset<cstr>, size=" << BENCHMARK_SIZE/16 << "\n";
+    std::cerr << "\nstrsize\tmsecs\tfind: cset<cstr>, size=" << BENCHMARK_SIZE/2 << "\n";
     for (int strsize = 1; strsize <= MAX_STRING_SIZE; strsize += 2) {
-        StcSet set = StcSet_with_capacity(BENCHMARK_SIZE/16);
-        sum += benchmark(set, BENCHMARK_SIZE/16, strsize), ++n;
-        std::cout << '\t' << cstr_str(StcSet_front(&set)) << '\n';
+        StcSet set = StcSet_with_capacity(BENCHMARK_SIZE/2);
+        sum += benchmark_lookup(set, BENCHMARK_SIZE/2, strsize), ++n;
+        std::cout << '\t' << cstr_str(StcSet_begin(&set).ref) << '\n';
         StcSet_drop(&set);
     }
-    std::cout << "Avg:\t" << sum/n << '\n';
+    std::cout << "Avg:\t" << sum/n << "ms\n";
+
 
     std::cerr << "sizeof(std::string) : " << sizeof(std::string) << std::endl
-              << "sizeof(cstr)        : " << sizeof(cstr) << std::endl;
+              << "sizeof(cstr)        : " << sizeof(cstr) << std::endl
+              << "sizeof(StdSet)      : " << sizeof(StdSet) << std::endl
+              << "sizeof(StcSet)      : " << sizeof(StcSet) << std::endl;
     return 0;
 }
diff --git a/misc/benchmarks/various/string_bench_STC.cpp b/misc/benchmarks/various/string_bench_STC.cpp
index ae8e4c38..a5dfd901 100644
--- a/misc/benchmarks/various/string_bench_STC.cpp
+++ b/misc/benchmarks/various/string_bench_STC.cpp
@@ -4,10 +4,11 @@
 #include <iostream>
 #include <iomanip>
 #include <chrono>
-#define i_static
+#define i_implement
 #include <stc/cstr.h>   // string
-#define i_static
+#define i_implement
 #include <stc/csview.h> // string_view
+#include <stc/algo/raii.h>
 
 #define i_key_str
 #include <stc/cvec.h>   // vec of cstr with const char* lookup
@@ -183,7 +184,7 @@ void benchmark(
 //const size_t MAX_LOOP = 1000000;
 const size_t MAX_LOOP = 2000;
 
-int main()
+int main(void)
 {
     c_auto (cvec_str, vec_string)
     c_auto (cvec_sv, vec_stringview)
diff --git a/misc/benchmarks/various/string_bench_STD.cpp b/misc/benchmarks/various/string_bench_STD.cpp
index 8bb87937..153ac02f 100644
--- a/misc/benchmarks/various/string_bench_STD.cpp
+++ b/misc/benchmarks/various/string_bench_STD.cpp
@@ -12,6 +12,7 @@
 #include <unordered_map>
 #define i_static
 #include <stc/cstr.h>
+#include <stc/algo/raii.h>
 
 std::vector<std::string> read_file(const char* name)
 {
@@ -193,7 +194,7 @@ void benchmark(
 //const size_t MAX_LOOP = 1000000;
 const size_t MAX_LOOP = 2000;
 
-int main()
+int main(void)
 {
     std::vector<std::string> vec_shortstr;
     std::vector<std::string_view> vec_shortstrview;