|
PQ-WB-KEM: White-Box ML-KEM-768 1.0
PQ-WB-KEM: Toward a White-Box Construction of ML-KEM-768 with Arithmetic Masking for M2M Communications - Mathematics (MDPI)
|
liboqs ML-KEM-768 KeyGen/Encaps/Decaps wall-clock benchmark; emits the full arm64_benchmark.json / baseline schema to stdout so every field is measured or computed, never transcribed. More...
#include <stdio.h>#include <stdlib.h>#include <stdint.h>#include <string.h>#include <time.h>#include <math.h>#include <oqs/oqs.h>
Include dependency graph for bench_liboqs.c:Go to the source code of this file.
Classes | |
| struct | opstat |
| struct | samples |
| Grow-on-demand sample buffer so the loop is bounded by wall-clock, not a fixed count. More... | |
Macros | |
| #define | BENCH_COMPILER "unknown" |
| #define | BENCH_OPENSSL 0 |
| #define | BENCH_PLATFORM "unknown" |
| #define | BENCH_TARGET "unknown" |
| #define | WARMUP 1000 |
| Warmup iterations discarded before timing. | |
Functions | |
| static void | emit_op (const char *name, opstat o, int last) |
| int | main (int argc, char **argv) |
| static double | now_sec (void) |
| static void | push (samples *s, double v) |
| static opstat | reduce (const samples *s) |
liboqs ML-KEM-768 KeyGen/Encaps/Decaps wall-clock benchmark; emits the full arm64_benchmark.json / baseline schema to stdout so every field is measured or computed, never transcribed.
Each operation runs a fixed wall-clock budget (default 3.0 s, argv[1]) after a warmup, and the per-iteration times are reduced to iteration count, mean and standard deviation. A time budget rather than a fixed iteration count keeps the aarch64/QEMU and x86-64 runs at equal wall-clock per operation, so the iteration count itself reports relative op cost. An unavailable ML-KEM-768 is a hard error: the harness never substitutes a different algorithm or emits a partial result. cpu_extensions is read from liboqs runtime detection; compiler and openssl_enabled are the build facts passed in as macros; oqs_version is the linked library's own string.
Definition in file bench_liboqs.c.
| struct opstat |
Definition at line 56 of file bench_liboqs.c.
Collaboration diagram for opstat:| Class Members | ||
|---|---|---|
| double | mean_us | |
| long | n | |
| double | stdev_us | |
| struct samples |
Grow-on-demand sample buffer so the loop is bounded by wall-clock, not a fixed count.
Definition at line 50 of file bench_liboqs.c.
Collaboration diagram for samples:| Class Members | ||
|---|---|---|
| long | cap | |
| long | n | |
| double * | t | |
| #define BENCH_COMPILER "unknown" |
Definition at line 28 of file bench_liboqs.c.
| #define BENCH_OPENSSL 0 |
Definition at line 37 of file bench_liboqs.c.
| #define BENCH_PLATFORM "unknown" |
Definition at line 34 of file bench_liboqs.c.
| #define BENCH_TARGET "unknown" |
Definition at line 31 of file bench_liboqs.c.
| #define WARMUP 1000 |
Warmup iterations discarded before timing.
Definition at line 41 of file bench_liboqs.c.
|
static |
Definition at line 68 of file bench_liboqs.c.
References opstat::mean_us, opstat::n, and opstat::stdev_us.
Referenced by main().
Here is the caller graph for this function:| int main | ( | int | argc, |
| char ** | argv | ||
| ) |
Definition at line 77 of file bench_liboqs.c.
References BENCH_COMPILER, BENCH_OPENSSL, BENCH_PLATFORM, BENCH_TARGET, emit_op(), now_sec(), push(), reduce(), samples::t, and WARMUP.
Here is the call graph for this function:
|
static |
Definition at line 43 of file bench_liboqs.c.
Referenced by main().
Here is the caller graph for this function:
|
static |
Definition at line 51 of file bench_liboqs.c.
References samples::cap, samples::n, and samples::t.
Referenced by main().
Here is the caller graph for this function:Definition at line 57 of file bench_liboqs.c.
References opstat::mean_us, samples::n, opstat::stdev_us, and samples::t.
Referenced by main().
Here is the caller graph for this function: