Benchmarks for recover and zero_poly

src/Makefile

@@ -1,6 +1,6 @@
 TESTS = bls12_381_test das_extension_test c_kzg_util_test fft_common_test fft_fr_test fft_g1_test \
 	fk20_proofs_test kzg_proofs_test poly_test recover_test utility_test zero_poly_test
-BENCH = fft_fr_bench fft_g1_bench
+BENCH = fft_fr_bench fft_g1_bench recover_bench zero_poly_bench
 LIB_SRC = bls12_381.c c_kzg_util.c das_extension.c fft_common.c fft_fr.c fft_g1.c fk20_proofs.c kzg_proofs.c poly.c recover.c utility.c zero_poly.c
 LIB_OBJ = $(LIB_SRC:.c=.o)
 

src/fft_fr_bench.c

@@ -50,6 +50,7 @@ long run_bench(int scale, int max_seconds) {
 
     free(out);
     free(data);
+    free_fft_settings(&fs);
 
     return total_time / nits;
 }

src/fft_g1_bench.c

@@ -50,6 +50,7 @@ long run_bench(int scale, int max_seconds) {
 
     free(out);
     free(data);
+    free_fft_settings(&fs);
 
     return total_time / nits;
 }

src/recover_bench.c

@@ -0,0 +1,107 @@
+/*
+ * Copyright 2021 Benjamin Edgington
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <stdlib.h> // malloc(), free(), atoi()
+#include <stdio.h>  // printf()
+#include <assert.h> // assert()
+#include <unistd.h> // EXIT_SUCCESS/FAILURE
+#include "bench_util.h"
+#include "test_util.h"
+#include "fft_fr.h"
+#include "recover.h"
+
+// Run the benchmark for `max_seconds` and return the time per iteration in nanoseconds.
+long run_bench(int scale, int max_seconds) {
+    timespec_t t0, t1;
+    unsigned long total_time = 0, nits = 0;
+    FFTSettings fs;
+
+    assert(C_KZG_OK == new_fft_settings(&fs, scale));
+
+    // Allocate on the heap to avoid stack overflow for large sizes
+    fr_t *poly = malloc(fs.max_width * sizeof(fr_t));
+    for (int i = 0; i < fs.max_width / 2; i++) {
+        fr_from_uint64(&poly[i], i);
+    }
+    for (int i = fs.max_width / 2; i < fs.max_width; i++) {
+        poly[i] = fr_zero;
+    }
+
+    fr_t *data = malloc(fs.max_width * sizeof(fr_t));
+    assert(C_KZG_OK == fft_fr(data, poly, false, fs.max_width, &fs));
+
+    fr_t *samples = malloc(fs.max_width * sizeof(fr_t));
+    for (int i = 0; i < fs.max_width; i++) {
+        samples[i] = data[i];
+    }
+
+    // randomly zero out half of the points
+    for (int i = 0; i < fs.max_width / 2; i++) {
+        int j = rand() % fs.max_width;
+        while (fr_is_null(&samples[j])) j = rand() % fs.max_width;
+        samples[j] = fr_null;
+    }
+
+    fr_t *recovered = malloc(fs.max_width * sizeof(fr_t));
+    while (total_time < max_seconds * NANO) {
+        clock_gettime(CLOCK_REALTIME, &t0);
+        assert(C_KZG_OK == recover_poly_from_samples(recovered, samples, fs.max_width, &fs));
+        clock_gettime(CLOCK_REALTIME, &t1);
+
+        // Verify the result is correct
+        for (int i = 0; i < fs.max_width; i++) {
+            assert(fr_equal(&data[i], &recovered[i]));
+        }
+
+        nits++;
+        total_time += tdiff(t0, t1);
+    }
+
+    free(recovered);
+    free(samples);
+    free(data);
+    free(poly);
+    free_fft_settings(&fs);
+
+    return total_time / nits;
+}
+
+int main(int argc, char *argv[]) {
+    int nsec = 0;
+
+    switch (argc) {
+    case 1:
+        nsec = NSEC;
+        break;
+    case 2:
+        nsec = atoi(argv[1]);
+        break;
+    default:
+        break;
+    };
+
+    if (nsec == 0) {
+        printf("Usage: %s [test time in seconds > 0]\n", argv[0]);
+        exit(EXIT_FAILURE);
+    }
+
+    printf("*** Benchmarking Recover From Samples, %d second%s per test.\n", nsec, nsec == 1 ? "" : "s");
+    for (int scale = 5; scale <= 15; scale++) {
+        printf("recover/scale_%d %lu ns/op\n", scale, run_bench(scale, nsec));
+    }
+
+    return EXIT_SUCCESS;
+}

src/zero_poly.c

@@ -156,7 +156,7 @@ C_KZG_RET reduce_leaves(fr_t *dst, uint64_t len_dst, fr_t *scratch, uint64_t len
  * @retval C_CZK_ERROR   An internal error occurred
  * @retval C_CZK_MALLOC  Memory allocation failed
  *
- * @todo What is the performance impact of tuning `per_leaf_poly`?
+ * @todo What is the performance impact of tuning `per_leaf_poly` and `reduction factor`?
  */
 C_KZG_RET zero_polynomial_via_multiplication(fr_t *zero_eval, fr_t *zero_poly, uint64_t *zero_poly_len, uint64_t length,
                                              const uint64_t *missing_indices, uint64_t len_missing,

src/zero_poly_bench.c

@@ -0,0 +1,87 @@
+/*
+ * Copyright 2021 Benjamin Edgington
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <stdlib.h> // malloc(), free(), atoi()
+#include <stdio.h>  // printf()
+#include <assert.h> // assert()
+#include <unistd.h> // EXIT_SUCCESS/FAILURE
+#include "bench_util.h"
+#include "test_util.h"
+#include "fft_fr.h"
+#include "zero_poly.h"
+
+// Run the benchmark for `max_seconds` and return the time per iteration in nanoseconds.
+long run_bench(int scale, int max_seconds) {
+    timespec_t t0, t1;
+    unsigned long total_time = 0, nits = 0;
+    FFTSettings fs;
+
+    assert(C_KZG_OK == new_fft_settings(&fs, scale));
+
+    // Allocate on the heap to avoid stack overflow for large sizes
+    uint64_t *missing = malloc(fs.max_width * sizeof(uint64_t));
+    for (int i = 0; i < fs.max_width; i++) {
+        missing[i] = i;
+    }
+    shuffle(missing, fs.max_width);
+
+    fr_t *zero_eval = malloc(fs.max_width * sizeof(fr_t));
+    fr_t *zero_poly = malloc(fs.max_width * sizeof(fr_t));
+    uint64_t zero_poly_len;
+    while (total_time < max_seconds * NANO) {
+        clock_gettime(CLOCK_REALTIME, &t0);
+        // Half missing leaves enough FFT computation space
+        assert(C_KZG_OK == zero_polynomial_via_multiplication(zero_eval, zero_poly, &zero_poly_len, fs.max_width,
+                                                              missing, fs.max_width / 2, &fs));
+        clock_gettime(CLOCK_REALTIME, &t1);
+        nits++;
+        total_time += tdiff(t0, t1);
+    }
+
+    free(zero_poly);
+    free(zero_eval);
+    free(missing);
+    free_fft_settings(&fs);
+
+    return total_time / nits;
+}
+
+int main(int argc, char *argv[]) {
+    int nsec = 0;
+
+    switch (argc) {
+    case 1:
+        nsec = NSEC;
+        break;
+    case 2:
+        nsec = atoi(argv[1]);
+        break;
+    default:
+        break;
+    };
+
+    if (nsec == 0) {
+        printf("Usage: %s [test time in seconds > 0]\n", argv[0]);
+        exit(EXIT_FAILURE);
+    }
+
+    printf("*** Benchmarking Zero Polynomial, %d second%s per test.\n", nsec, nsec == 1 ? "" : "s");
+    for (int scale = 5; scale <= 15; scale++) {
+        printf("zero_poly/scale_%d %lu ns/op\n", scale, run_bench(scale, nsec));
+    }
+
+    return EXIT_SUCCESS;
+}