mirror of
https://github.com/rdolbeau/VexRiscvBPluginGenerator.git
synced 2025-04-18 18:44:42 -04:00
small pbsad test
This commit is contained in:
Romain Dolbeau
parent
3abeaa3e0c
commit
54c93b289f
1 changed files with 178 additions and 0 deletions
178
pbsad.c
Normal file
178
pbsad.c
Normal file
|
|
@ -0,0 +1,178 @@
|
|||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include <stdint.h>
|
||||
|
||||
typedef uint32_t MpegEncContext; /* don't care */
|
||||
typedef ssize_t ptrdiff_t;
|
||||
|
||||
#ifdef __riscv
|
||||
|
||||
typedef uint32_t uint_xlen_t;
|
||||
#define XLEN 32
|
||||
#define FUN2(NAME, ASNAME) \
|
||||
static inline uint_xlen_t NAME(uint_xlen_t rs1, uint_xlen_t rs2) { \
|
||||
uint32_t r; \
|
||||
asm (#ASNAME " reg_%0, reg_%1, reg_%2\n" \
|
||||
: "=r" (r) \
|
||||
: "r" (rs1), "r" (rs2)); \
|
||||
return r; \
|
||||
}
|
||||
#define FUN3R(NAME, ASNAME) \
|
||||
static inline uint_xlen_t NAME(uint_xlen_t rs1, uint_xlen_t rs2, uint_xlen_t rs3) { \
|
||||
uint32_t r = rs3; \
|
||||
asm (#ASNAME " reg_%0, reg_%1, reg_%2\n" \
|
||||
: "+&r" (r) \
|
||||
: "r" (rs1), "r" (rs2)); \
|
||||
return r; \
|
||||
}
|
||||
#define ASM2MACRO(N, O) asm(".macro "#N" rd, rs1, rs2\n" \
|
||||
".word ("#O" | (\\rd << 7) | (\\rs1 << 15) | (\\rs2 << 20))\n" \
|
||||
".endm\n");
|
||||
#define ASM3RMACRO(N, O) ASM2MACRO(N, O)
|
||||
|
||||
asm("#define reg_zero 0\n");
|
||||
asm("#define reg_ra 1\n");
|
||||
asm("#define reg_sp 2\n");
|
||||
asm("#define reg_gp 3\n");
|
||||
asm("#define reg_tp 4\n");
|
||||
asm("#define reg_t0 5\n");
|
||||
asm("#define reg_t1 6\n");
|
||||
asm("#define reg_t2 7\n");
|
||||
asm("#define reg_s0 8\n");
|
||||
asm("#define reg_s1 9\n");
|
||||
asm("#define reg_a0 10\n");
|
||||
asm("#define reg_a1 11\n");
|
||||
asm("#define reg_a2 12\n");
|
||||
asm("#define reg_a3 13\n");
|
||||
asm("#define reg_a4 14\n");
|
||||
asm("#define reg_a5 15\n");
|
||||
asm("#define reg_a6 16\n");
|
||||
asm("#define reg_a7 17\n");
|
||||
asm("#define reg_s2 18\n");
|
||||
asm("#define reg_s3 19\n");
|
||||
asm("#define reg_s4 20\n");
|
||||
asm("#define reg_s5 21\n");
|
||||
asm("#define reg_s6 22\n");
|
||||
asm("#define reg_s7 23\n");
|
||||
asm("#define reg_s8 24\n");
|
||||
asm("#define reg_s9 25\n");
|
||||
asm("#define reg_s10 26\n");
|
||||
asm("#define reg_s11 27\n");
|
||||
asm("#define reg_t3 28\n");
|
||||
asm("#define reg_t4 29\n");
|
||||
asm("#define reg_t5 30\n");
|
||||
asm("#define reg_t6 31\n");
|
||||
|
||||
ASM2MACRO(PBSAD, 0xfc000077)
|
||||
FUN2(__rv__pbsad, PBSAD)
|
||||
ASM3RMACRO(PBSADA, 0xfe000077)
|
||||
FUN3R(__rv__pbsada, PBSADA)
|
||||
|
||||
static inline long long cpucycles_riscv(void) {
|
||||
long long result;
|
||||
#if defined(__riscv_xlen)
|
||||
#if __riscv_xlen == 64
|
||||
asm volatile("rdcycle %0" : "=r" (result));
|
||||
#elif __riscv_xlen == 32
|
||||
unsigned int l, h, h2;
|
||||
asm volatile( "start:\n"
|
||||
"rdcycleh %0\n"
|
||||
"rdcycle %1\n"
|
||||
"rdcycleh %2\n"
|
||||
"bne %0, %2, start\n"
|
||||
: "=r" (h), "=r" (l), "=r" (h2));
|
||||
|
||||
result = (((unsigned long long)h)<<32) | ((unsigned long long)l);
|
||||
#else
|
||||
#error "unknown __riscv_xlen"
|
||||
#endif
|
||||
#else // __riscv_xlen
|
||||
#error "__riscv_xlen required for RISC-V support"
|
||||
#endif // __riscv_xlen
|
||||
return result;
|
||||
}
|
||||
static inline int pix_abs8_r5vp(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
|
||||
ptrdiff_t stride, int h)
|
||||
{
|
||||
int s = 0, i;
|
||||
|
||||
for (i = 0; i < h; i++) {
|
||||
uint32_t p1 = ((uint32_t*)pix1)[0];
|
||||
uint32_t p2 = ((uint32_t*)pix2)[0];
|
||||
s = __rv__pbsada(p1, p2, s);
|
||||
p1 = ((uint32_t*)pix1)[1];
|
||||
p2 = ((uint32_t*)pix2)[1];
|
||||
s = __rv__pbsada(p1, p2, s);
|
||||
pix1 += stride;
|
||||
pix2 += stride;
|
||||
}
|
||||
return s;
|
||||
}
|
||||
#endif // __riscv
|
||||
|
||||
|
||||
|
||||
/* lifted from FFmpeg */
|
||||
static inline int pix_abs8_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
|
||||
ptrdiff_t stride, int h)
|
||||
{
|
||||
int s = 0, i;
|
||||
|
||||
for (i = 0; i < h; i++) {
|
||||
s += abs(pix1[0] - pix2[0]);
|
||||
s += abs(pix1[1] - pix2[1]);
|
||||
s += abs(pix1[2] - pix2[2]);
|
||||
s += abs(pix1[3] - pix2[3]);
|
||||
s += abs(pix1[4] - pix2[4]);
|
||||
s += abs(pix1[5] - pix2[5]);
|
||||
s += abs(pix1[6] - pix2[6]);
|
||||
s += abs(pix1[7] - pix2[7]);
|
||||
pix1 += stride;
|
||||
pix2 += stride;
|
||||
}
|
||||
return s;
|
||||
}
|
||||
|
||||
int main(int argc, char **argv) {
|
||||
unsigned int seed = 0;
|
||||
uint8_t *pix1, *pix2;
|
||||
uint32_t i, j;
|
||||
|
||||
if (argc > 1)
|
||||
seed = atol(argv[1]);
|
||||
|
||||
srandom(seed);
|
||||
|
||||
pix1 = malloc(8 * 4 * 8 * 4 * sizeof(uint8_t));
|
||||
pix2 = malloc(8 * 4 * 8 * 4 * sizeof(uint8_t));
|
||||
for (i = 0 ; i < 8 * 8 * 4 * 4 ; i ++) {
|
||||
pix1[i] = random() & 0xFF;
|
||||
pix2[i] = random() & 0xFF;
|
||||
}
|
||||
|
||||
for (i = 0 ; i < 4 ; i++) {
|
||||
for (j = 0 ; j < 4 ; j++) {
|
||||
size_t offset = i*8+j*(8*4*8);
|
||||
long long t0 = 0, t1 = 0, t2 = 0, t3 = 0, t4 = 0;
|
||||
#ifdef __riscv
|
||||
t0 = cpucycles_riscv();
|
||||
#endif
|
||||
int res = pix_abs8_c(&i, pix1 + offset, pix2 + offset, 8 * 4, 8);
|
||||
#ifdef __riscv
|
||||
t1 = cpucycles_riscv();
|
||||
int res2 = pix_abs8_r5vp(&i, pix1 + offset, pix2 + offset, 8 * 4, 8);
|
||||
t2 = cpucycles_riscv();
|
||||
int res3 = pix_abs8_c(&i, pix1 + offset, pix2 + offset, 8 * 4, 8);
|
||||
t3 = cpucycles_riscv();
|
||||
int res4 = pix_abs8_r5vp(&i, pix1 + offset, pix2 + offset, 8 * 4, 8);
|
||||
t4 = cpucycles_riscv();
|
||||
#else
|
||||
int res2 = 0, res3 = 0, res4 = 0;
|
||||
#endif
|
||||
printf("%u, %u: %d, %d / %d, %d (%llu, %llu / %llu, %llu)\n", i, j, res, res2, res3, res4, t1-t0, t2-t1, t3-t2, t4-t3);
|
||||
}
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
Loading…
Add table
Reference in a new issue