diff options
Diffstat (limited to 'src/gallium/auxiliary/tgsi/tgsi_sse2.c')
| -rw-r--r-- | src/gallium/auxiliary/tgsi/tgsi_sse2.c | 153 |
1 files changed, 137 insertions, 16 deletions
diff --git a/src/gallium/auxiliary/tgsi/tgsi_sse2.c b/src/gallium/auxiliary/tgsi/tgsi_sse2.c index f79170b9d6..f93db18114 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_sse2.c +++ b/src/gallium/auxiliary/tgsi/tgsi_sse2.c @@ -27,12 +27,14 @@ #include "pipe/p_config.h" -#if defined(PIPE_ARCH_X86) && defined(PIPE_ARCH_SSE) +#if defined(PIPE_ARCH_X86) #include "pipe/p_debug.h" #include "pipe/p_shader_tokens.h" #include "util/u_math.h" +#if defined(PIPE_ARCH_SSE) #include "util/u_sse.h" +#endif #include "tgsi/tgsi_parse.h" #include "tgsi/tgsi_util.h" #include "tgsi_exec.h" @@ -72,6 +74,9 @@ #define TEMP_R0 TGSI_EXEC_TEMP_R0 #define TEMP_ADDR TGSI_EXEC_TEMP_ADDR +#define TEMP_EXEC_MASK_I TGSI_EXEC_MASK_I +#define TEMP_EXEC_MASK_C TGSI_EXEC_MASK_C + /** * X86 utility functions. @@ -233,6 +238,9 @@ emit_const( int indirectIndex ) { if (indirect) { + /* 'vec' is the offset from the address register's value. + * We're loading CONST[ADDR+vec] into an xmm register. + */ struct x86_reg r0 = get_input_base(); struct x86_reg r1 = get_output_base(); uint i; @@ -243,18 +251,40 @@ emit_const( x86_push( func, r0 ); x86_push( func, r1 ); + /* + * Loop over the four pixels or vertices in the quad. + * Get the value of the address (offset) register for pixel/vertex[i], + * add it to the src offset and index into the constant buffer. + * Note that we're working on SOA data. + * If any of the pixel/vertex execution channels are unused their + * values will be garbage. It's very important that we don't use + * those garbage values as indexes into the constant buffer since + * that'll cause segfaults. + * The solution is to bitwise-AND the offset with the execution mask + * register whose values are either 0 or ~0. + * The caller must setup the execution mask register to indicate + * which channels are valid/alive before running the shader. + * The execution mask will also figure into loops and conditionals + * someday. + */ for (i = 0; i < QUAD_SIZE; i++) { - x86_lea( func, r0, get_const( vec, chan ) ); + /* r1 = address register[i] */ x86_mov( func, r1, x86_make_disp( get_temp( TEMP_ADDR, CHAN_X ), i * 4 ) ); + /* r0 = execution mask[i] */ + x86_mov( func, r0, x86_make_disp( get_temp( TEMP_EXEC_MASK_I, TEMP_EXEC_MASK_C ), i * 4 ) ); + /* r1 = r1 & r0 */ + x86_and( func, r1, r0 ); + /* r0 = 'vec', the offset */ + x86_lea( func, r0, get_const( vec, chan ) ); - /* Quick hack to multiply by 16 -- need to add SHL to rtasm. + /* Quick hack to multiply r1 by 16 -- need to add SHL to rtasm. */ x86_add( func, r1, r1 ); x86_add( func, r1, r1 ); x86_add( func, r1, r1 ); x86_add( func, r1, r1 ); - x86_add( func, r0, r1 ); + x86_add( func, r0, r1 ); /* r0 = r0 + r1 */ x86_mov( func, r1, x86_deref( r0 ) ); x86_mov( func, x86_make_disp( get_temp( TEMP_R0, CHAN_X ), i * 4 ), r1 ); } @@ -268,6 +298,7 @@ emit_const( get_temp( TEMP_R0, CHAN_X ) ); } else { + /* 'vec' is the index into the src register file, such as TEMP[vec] */ assert( vec >= 0 ); sse_movss( @@ -598,6 +629,9 @@ emit_func_call_dst_src( code ); } + +#if defined(PIPE_ARCH_SSE) + /* * Fast SSE2 implementation of special math functions. */ @@ -649,6 +683,7 @@ exp2f4(__m128 x) return _mm_mul_ps(expipart, expfpart); } + /** * See http://www.devmaster.net/forums/showthread.php?p=43580 */ @@ -691,12 +726,16 @@ log2f4(__m128 x) return _mm_add_ps(logmant, exp); } + static INLINE __m128 powf4(__m128 x, __m128 y) { return exp2f4(_mm_mul_ps(log2f4(x), y)); } +#endif /* PIPE_ARCH_SSE */ + + /** * Low-level instruction translators. @@ -751,13 +790,20 @@ emit_cos( } static void PIPE_CDECL -#if defined(PIPE_CC_GCC) +#if defined(PIPE_CC_GCC) && defined(PIPE_ARCH_SSE) __attribute__((force_align_arg_pointer)) #endif ex24f( float *store ) { +#if defined(PIPE_ARCH_SSE) _mm_store_ps(&store[0], exp2f4( _mm_load_ps(&store[0]) )); +#else + store[0] = util_fast_exp2( store[0] ); + store[1] = util_fast_exp2( store[1] ); + store[2] = util_fast_exp2( store[2] ); + store[3] = util_fast_exp2( store[3] ); +#endif } static void @@ -784,6 +830,17 @@ emit_f2it( make_xmm( xmm ) ); } +static void +emit_i2f( + struct x86_function *func, + unsigned xmm ) +{ + sse2_cvtdq2ps( + func, + make_xmm( xmm ), + make_xmm( xmm ) ); +} + static void PIPE_CDECL flr4f( float *store ) @@ -831,13 +888,20 @@ emit_frc( } static void PIPE_CDECL -#if defined(PIPE_CC_GCC) +#if defined(PIPE_CC_GCC) && defined(PIPE_ARCH_SSE) __attribute__((force_align_arg_pointer)) #endif lg24f( float *store ) { +#if defined(PIPE_ARCH_SSE) _mm_store_ps(&store[0], log2f4( _mm_load_ps(&store[0]) )); +#else + store[0] = util_fast_log2( store[0] ); + store[1] = util_fast_log2( store[1] ); + store[2] = util_fast_log2( store[2] ); + store[3] = util_fast_log2( store[3] ); +#endif } static void @@ -890,19 +954,19 @@ emit_neg( } static void PIPE_CDECL -#if defined(PIPE_CC_GCC) +#if defined(PIPE_CC_GCC) && defined(PIPE_ARCH_SSE) __attribute__((force_align_arg_pointer)) #endif pow4f( float *store ) { -#if 1 +#if defined(PIPE_ARCH_SSE) _mm_store_ps(&store[0], powf4( _mm_load_ps(&store[0]), _mm_load_ps(&store[4]) )); #else - store[0] = powf( store[0], store[4] ); - store[1] = powf( store[1], store[5] ); - store[2] = powf( store[2], store[6] ); - store[3] = powf( store[3], store[7] ); + store[0] = util_fast_pow( store[0], store[4] ); + store[1] = util_fast_pow( store[1], store[5] ); + store[2] = util_fast_pow( store[2], store[6] ); + store[3] = util_fast_pow( store[3], store[7] ); #endif } @@ -1702,7 +1766,18 @@ emit_instruction( case TGSI_OPCODE_DOT2ADD: /* TGSI_OPCODE_DP2A */ - return 0; + FETCH( func, *inst, 0, 0, CHAN_X ); /* xmm0 = src[0].x */ + FETCH( func, *inst, 1, 1, CHAN_X ); /* xmm1 = src[1].x */ + emit_mul( func, 0, 1 ); /* xmm0 = xmm0 * xmm1 */ + FETCH( func, *inst, 1, 0, CHAN_Y ); /* xmm1 = src[0].y */ + FETCH( func, *inst, 2, 1, CHAN_Y ); /* xmm2 = src[1].y */ + emit_mul( func, 1, 2 ); /* xmm1 = xmm1 * xmm2 */ + emit_add( func, 0, 1 ); /* xmm0 = xmm0 + xmm1 */ + FETCH( func, *inst, 1, 2, CHAN_X ); /* xmm1 = src[2].x */ + emit_add( func, 0, 1 ); /* xmm0 = xmm0 + xmm1 */ + FOR_EACH_DST0_ENABLED_CHANNEL( *inst, chan_index ) { + STORE( func, *inst, 0, 0, chan_index ); /* dest[ch] = xmm0 */ + } break; case TGSI_OPCODE_INDEX: @@ -2036,7 +2111,39 @@ emit_instruction( break; case TGSI_OPCODE_NRM: - return 0; + /* fall-through */ + case TGSI_OPCODE_NRM4: + /* 3 or 4-component normalization */ + { + uint dims = (inst->Instruction.Opcode == TGSI_OPCODE_NRM) ? 3 : 4; + /* note: cannot use xmm regs 2/3 here (see emit_rsqrt() above) */ + FETCH( func, *inst, 4, 0, CHAN_X ); /* xmm4 = src[0].x */ + FETCH( func, *inst, 5, 0, CHAN_Y ); /* xmm5 = src[0].y */ + FETCH( func, *inst, 6, 0, CHAN_Z ); /* xmm6 = src[0].z */ + if (dims == 4) { + FETCH( func, *inst, 7, 0, CHAN_W ); /* xmm7 = src[0].w */ + } + emit_MOV( func, 0, 4 ); /* xmm0 = xmm3 */ + emit_mul( func, 0, 4 ); /* xmm0 *= xmm3 */ + emit_MOV( func, 1, 5 ); /* xmm1 = xmm4 */ + emit_mul( func, 1, 5 ); /* xmm1 *= xmm4 */ + emit_add( func, 0, 1 ); /* xmm0 += xmm1 */ + emit_MOV( func, 1, 6 ); /* xmm1 = xmm5 */ + emit_mul( func, 1, 6 ); /* xmm1 *= xmm5 */ + emit_add( func, 0, 1 ); /* xmm0 += xmm1 */ + if (dims == 4) { + emit_MOV( func, 1, 7 ); /* xmm1 = xmm7 */ + emit_mul( func, 1, 7 ); /* xmm1 *= xmm7 */ + emit_add( func, 0, 0 ); /* xmm0 += xmm1 */ + } + emit_rsqrt( func, 1, 0 ); /* xmm1 = 1/sqrt(xmm0) */ + FOR_EACH_DST0_ENABLED_CHANNEL( *inst, chan_index ) { + if (chan_index < dims) { + emit_mul( func, 4+chan_index, 1); /* xmm[4+ch] *= xmm1 */ + STORE( func, *inst, 4+chan_index, 0, chan_index ); + } + } + } break; case TGSI_OPCODE_DIV: @@ -2044,7 +2151,16 @@ emit_instruction( break; case TGSI_OPCODE_DP2: - return 0; + FETCH( func, *inst, 0, 0, CHAN_X ); /* xmm0 = src[0].x */ + FETCH( func, *inst, 1, 1, CHAN_X ); /* xmm1 = src[1].x */ + emit_mul( func, 0, 1 ); /* xmm0 = xmm0 * xmm1 */ + FETCH( func, *inst, 1, 0, CHAN_Y ); /* xmm1 = src[0].y */ + FETCH( func, *inst, 2, 1, CHAN_Y ); /* xmm2 = src[1].y */ + emit_mul( func, 1, 2 ); /* xmm1 = xmm1 * xmm2 */ + emit_add( func, 0, 1 ); /* xmm0 = xmm0 + xmm1 */ + FOR_EACH_DST0_ENABLED_CHANNEL( *inst, chan_index ) { + STORE( func, *inst, 0, 0, chan_index ); /* dest[ch] = xmm0 */ + } break; case TGSI_OPCODE_TXL: @@ -2104,7 +2220,12 @@ emit_instruction( break; case TGSI_OPCODE_TRUNC: - return 0; + FOR_EACH_DST0_ENABLED_CHANNEL( *inst, chan_index ) { + FETCH( func, *inst, 0, 0, chan_index ); + emit_f2it( func, 0 ); + emit_i2f( func, 0 ); + STORE( func, *inst, 0, 0, chan_index ); + } break; case TGSI_OPCODE_SHL: |