mirror of
https://github.com/henrydcase/pqc.git
synced 2024-11-27 01:41:40 +00:00
ac2c20045c
* Add McEliece reference implementations * Add Vec implementations of McEliece * Add sse implementations * Add AVX2 implementations * Get rid of stuff not supported by Mac ABI * restrict to two cores * Ditch .data files * Remove .hidden from all .S files * speed up duplicate consistency tests by batching * make cpuinfo more robust * Hope to stabilize macos cpuinfo without ccache * Revert "Hope to stabilize macos cpuinfo without ccache" This reverts commit 6129c3cabe1abbc8b956bc87e902a698e32bf322. * Just hardcode what's available at travis * Fixed-size types in api.h * namespace all header files in mceliece * Ditch operations.h * Get rid of static inline functions * fixup! Ditch operations.h
46 lines
2.1 KiB
C
46 lines
2.1 KiB
C
#ifndef PQCLEAN_MCELIECE6960119F_AVX_VEC256_H
|
|
#define PQCLEAN_MCELIECE6960119F_AVX_VEC256_H
|
|
/*
|
|
This file is for functions related to 256-bit vectors
|
|
including functions for bitsliced field operations
|
|
*/
|
|
|
|
|
|
#include "vec128.h"
|
|
|
|
#include <immintrin.h>
|
|
|
|
typedef __m256i vec256;
|
|
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_set1_16b(uint16_t a);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_setzero(void);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_set4x(uint64_t a0, uint64_t a1, uint64_t a2, uint64_t a3);
|
|
|
|
// Extract requires a literal argument so need to be macros
|
|
#define PQCLEAN_MCELIECE6960119F_AVX_vec256_extract2x(a,i) ((vec128) _mm256_extractf128_si256((vec256) (a),(i)))
|
|
#define PQCLEAN_MCELIECE6960119F_AVX_vec256_extract(a,i) ((uint64_t) _mm256_extract_epi64((vec256) (a),(i)))
|
|
|
|
int PQCLEAN_MCELIECE6960119F_AVX_vec256_testz(vec256 a);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_and(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_or(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_sll_4x(vec256 a, int s);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_srl_4x(vec256 a, int s);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_low(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_high(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_low_2x(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_high_2x(vec256 a, vec256 b);
|
|
vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_or_reduce(const vec256 *a);
|
|
void PQCLEAN_MCELIECE6960119F_AVX_vec256_copy(vec256 *dest, const vec256 *src);
|
|
|
|
/* bitsliced field multiplications */
|
|
void PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(vec256 *h, vec256 *f, const vec256 *g);
|
|
void PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(vec256 * /*out*/, const vec256 * /*in*/);
|
|
void PQCLEAN_MCELIECE6960119F_AVX_vec256_inv(vec256 * /*out*/, const vec256 * /*in*/);
|
|
|
|
extern void PQCLEAN_MCELIECE6960119F_AVX_vec256_maa_asm(vec256 *, vec256 *, const vec256 *);
|
|
extern void PQCLEAN_MCELIECE6960119F_AVX_vec256_ama_asm(vec256 *, vec256 *, const vec256 *);
|
|
|
|
#endif
|
|
|