rockbox/apps/codecs/lib/asm_arm.h
Nils Wallménius 7afea91560 ARMv6 versions of X(N)PROD31 macros and MULT32 macro. Saves about 1MHz or 3% decoding vorbis on gigabeat S.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26889 a1c6a512-1295-4272-9138-f99709370657
2010-06-17 15:24:23 +00:00

281 lines
7.6 KiB
C

/********************************************************************
* *
* THIS FILE IS PART OF THE OggVorbis 'TREMOR' CODEC SOURCE CODE. *
* *
* USE, DISTRIBUTION AND REPRODUCTION OF THIS LIBRARY SOURCE IS *
* GOVERNED BY A BSD-STYLE SOURCE LICENSE INCLUDED WITH THIS SOURCE *
* IN 'COPYING'. PLEASE READ THESE TERMS BEFORE DISTRIBUTING. *
* *
* THE OggVorbis 'TREMOR' SOURCE CODE IS (C) COPYRIGHT 1994-2002 *
* BY THE Xiph.Org FOUNDATION http://www.xiph.org/ *
* *
********************************************************************
function: arm7 and later wide math functions
********************************************************************/
#ifdef CPU_ARM
#if !defined(_V_WIDE_MATH) && !defined(_LOW_ACCURACY_)
#define _V_WIDE_MATH
#if ARM_ARCH >= 6
static inline int32_t MULT32(int32_t x, int32_t y) {
int32_t hi;
asm volatile("smmul %[hi], %[x], %[y] \n\t"
: [hi] "=&r" (hi)
: [x] "r" (x), [y] "r" (y) );
return(hi);
}
#else
static inline int32_t MULT32(int32_t x, int32_t y) {
int32_t lo, hi;
asm volatile("smull\t%0, %1, %2, %3 \n\t"
: "=&r"(lo),"=&r"(hi)
: "r"(x),"r"(y) );
return(hi);
}
#endif
static inline int32_t MULT31(int32_t x, int32_t y) {
return MULT32(x,y)<<1;
}
static inline int32_t MULT31_SHIFT15(int32_t x, int32_t y) {
int32_t lo,hi;
asm volatile("smull %0, %1, %2, %3\n\t"
"movs %0, %0, lsr #15\n\t"
"adc %1, %0, %1, lsl #17\n\t"
: "=&r"(lo),"=&r"(hi)
: "r"(x),"r"(y)
: "cc" );
return(hi);
}
#define XPROD32(a, b, t, v, x, y) \
{ \
int32_t l; \
asm("smull %0, %1, %3, %5\n\t" \
"rsb %2, %6, #0\n\t" \
"smlal %0, %1, %4, %6\n\t" \
"smull %0, %2, %3, %2\n\t" \
"smlal %0, %2, %4, %5" \
: "=&r" (l), "=&r" (x), "=&r" (y) \
: "r" ((a)), "r" ((b)), "r" ((t)), "r" ((v)) ); \
}
#if ARM_ARCH >= 6
/* These may yield slightly different result from the macros below
because only the high 32 bits of the multiplications are accumulated while
the below macros use a 64 bit accumulator that is truncated to 32 bits.*/
#define XPROD31_R(_a, _b, _t, _v, _x, _y)\
{\
int32_t x1, y1;\
asm("smmul %[x1], %[t], %[a] \n\t"\
"smmul %[y1], %[t], %[b] \n\t"\
"smmla %[x1], %[v], %[b], %[x1] \n\t"\
"smmls %[y1], %[v], %[a], %[y1] \n\t"\
: [x1] "=&r" (x1), [y1] "=&r" (y1)\
: [a] "r" (_a), [b] "r" (_b), [t] "r" (_t), [v] "r" (_v) );\
_x = x1 << 1;\
_y = y1 << 1;\
}
#define XNPROD31_R(_a, _b, _t, _v, _x, _y)\
{\
int32_t x1, y1;\
asm("smmul %[x1], %[t], %[a] \n\t"\
"smmul %[y1], %[t], %[b] \n\t"\
"smmls %[x1], %[v], %[b], %[x1] \n\t"\
"smmla %[y1], %[v], %[a], %[y1] \n\t"\
: [x1] "=&r" (x1), [y1] "=&r" (y1)\
: [a] "r" (_a), [b] "r" (_b), [t] "r" (_t), [v] "r" (_v) );\
_x = x1 << 1;\
_y = y1 << 1;\
}
#else
#define XPROD31_R(_a, _b, _t, _v, _x, _y)\
{\
int32_t x1, y1, l;\
asm("smull %0, %1, %5, %3\n\t"\
"rsb %2, %3, #0\n\t"\
"smlal %0, %1, %6, %4\n\t"\
"smull %0, %2, %6, %2\n\t"\
"smlal %0, %2, %5, %4"\
: "=&r" (l), "=&r" (x1), "=&r" (y1)\
: "r" (_a), "r" (_b), "r" (_t), "r" (_v) );\
_x = x1 << 1;\
_y = y1 << 1;\
}
#define XNPROD31_R(_a, _b, _t, _v, _x, _y)\
{\
int32_t x1, y1, l;\
asm("smull %0, %1, %5, %3\n\t"\
"rsb %2, %4, #0\n\t"\
"smlal %0, %1, %6, %2\n\t"\
"smull %0, %2, %5, %4\n\t"\
"smlal %0, %2, %6, %3"\
: "=&r" (l), "=&r" (x1), "=&r" (y1)\
: "r" (_a), "r" (_b), "r" (_t), "r" (_v) );\
_x = x1 << 1;\
_y = y1 << 1;\
}
#endif
static inline void XPROD31(int32_t a, int32_t b,
int32_t t, int32_t v,
int32_t *x, int32_t *y)
{
int32_t _x1, _y1;
XPROD31_R(a, b, t, v, _x1, _y1);
*x = _x1;
*y = _y1;
}
static inline void XNPROD31(int32_t a, int32_t b,
int32_t t, int32_t v,
int32_t *x, int32_t *y)
{
int32_t _x1, _y1;
XNPROD31_R(a, b, t, v, _x1, _y1);
*x = _x1;
*y = _y1;
}
#ifndef _V_VECT_OPS
#define _V_VECT_OPS
/* asm versions of vector operations for block.c, window.c */
static inline
void vect_add(int32_t *x, int32_t *y, int n)
{
while (n>=4) {
asm volatile ("ldmia %[x], {r0, r1, r2, r3};"
"ldmia %[y]!, {r4, r5, r6, r7};"
"add r0, r0, r4;"
"add r1, r1, r5;"
"add r2, r2, r6;"
"add r3, r3, r7;"
"stmia %[x]!, {r0, r1, r2, r3};"
: [x] "+r" (x), [y] "+r" (y)
: : "r0", "r1", "r2", "r3",
"r4", "r5", "r6", "r7",
"memory");
n -= 4;
}
/* add final elements */
while (n>0) {
*x++ += *y++;
n--;
}
}
static inline
void vect_copy(int32_t *x, int32_t *y, int n)
{
while (n>=4) {
asm volatile ("ldmia %[y]!, {r0, r1, r2, r3};"
"stmia %[x]!, {r0, r1, r2, r3};"
: [x] "+r" (x), [y] "+r" (y)
: : "r0", "r1", "r2", "r3",
"memory");
n -= 4;
}
/* copy final elements */
while (n>0) {
*x++ = *y++;
n--;
}
}
static inline
void vect_mult_fw(int32_t *data, int32_t *window, int n)
{
while (n>=4) {
asm volatile ("ldmia %[d], {r0, r1, r2, r3};"
"ldmia %[w]!, {r4, r5, r6, r7};"
"smull r8, r9, r0, r4;"
"mov r0, r9, lsl #1;"
"smull r8, r9, r1, r5;"
"mov r1, r9, lsl #1;"
"smull r8, r9, r2, r6;"
"mov r2, r9, lsl #1;"
"smull r8, r9, r3, r7;"
"mov r3, r9, lsl #1;"
"stmia %[d]!, {r0, r1, r2, r3};"
: [d] "+r" (data), [w] "+r" (window)
: : "r0", "r1", "r2", "r3",
"r4", "r5", "r6", "r7", "r8", "r9",
"memory" );
n -= 4;
}
while(n>0) {
*data = MULT31(*data, *window);
data++;
window++;
n--;
}
}
static inline
void vect_mult_bw(int32_t *data, int32_t *window, int n)
{
while (n>=4) {
asm volatile ("ldmia %[d], {r0, r1, r2, r3};"
"ldmda %[w]!, {r4, r5, r6, r7};"
"smull r8, r9, r0, r7;"
"mov r0, r9, lsl #1;"
"smull r8, r9, r1, r6;"
"mov r1, r9, lsl #1;"
"smull r8, r9, r2, r5;"
"mov r2, r9, lsl #1;"
"smull r8, r9, r3, r4;"
"mov r3, r9, lsl #1;"
"stmia %[d]!, {r0, r1, r2, r3};"
: [d] "+r" (data), [w] "+r" (window)
: : "r0", "r1", "r2", "r3",
"r4", "r5", "r6", "r7", "r8", "r9",
"memory" );
n -= 4;
}
while(n>0) {
*data = MULT31(*data, *window);
data++;
window--;
n--;
}
}
#endif
#endif
/* not used anymore */
/*
#ifndef _V_CLIP_MATH
#define _V_CLIP_MATH
static inline int32_t CLIP_TO_15(int32_t x) {
int tmp;
asm volatile("subs %1, %0, #32768\n\t"
"movpl %0, #0x7f00\n\t"
"orrpl %0, %0, #0xff\n"
"adds %1, %0, #32768\n\t"
"movmi %0, #0x8000"
: "+r"(x),"=r"(tmp)
:
: "cc");
return(x);
}
#endif
*/
#ifndef _V_LSP_MATH_ASM
#define _V_LSP_MATH_ASM
#endif
#endif