Bochs/bochs/cpu/avx/avx512_fma.cc
2015-05-02 19:54:48 +00:00

173 lines
11 KiB
C++

/////////////////////////////////////////////////////////////////////////
// $Id$
/////////////////////////////////////////////////////////////////////////
//
// Copyright (c) 2013 Stanislav Shwartsman
// Written by Stanislav Shwartsman [sshwarts at sourceforge net]
//
// This library is free software; you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public
// License as published by the Free Software Foundation; either
// version 2 of the License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA B 02110-1301 USA
//
/////////////////////////////////////////////////////////////////////////
#define NEED_CPU_REG_SHORTCUTS 1
#include "bochs.h"
#include "cpu.h"
#define LOG_THIS BX_CPU_THIS_PTR
#if BX_SUPPORT_EVEX
extern float_status_t mxcsr_to_softfloat_status_word(bx_mxcsr_t mxcsr);
#include "simd_int.h"
#include "simd_pfp.h"
//////////////////////////////
// AVX-512 FMA Instructions //
//////////////////////////////
#define EVEX_FMA_PACKED_SINGLE(HANDLER, func) \
BX_INSF_TYPE BX_CPP_AttrRegparmN(1) BX_CPU_C:: HANDLER (bxInstruction_c *i) \
{ \
BxPackedAvxRegister op1 = BX_READ_AVX_REG(i->src1()); \
BxPackedAvxRegister op2 = BX_READ_AVX_REG(i->src2()); \
BxPackedAvxRegister op3 = BX_READ_AVX_REG(i->src3()); \
unsigned mask = BX_READ_16BIT_OPMASK(i->opmask()); \
unsigned len = i->getVL(); \
\
float_status_t status = mxcsr_to_softfloat_status_word(MXCSR); \
softfloat_status_word_rc_override(status, i); \
\
for (unsigned n=0, tmp_mask = mask; n < len; n++, tmp_mask >>= 4) \
(func)(&op1.vmm128(n), &op2.vmm128(n), &op3.vmm128(n), status, tmp_mask); \
\
check_exceptionsSSE(get_exception_flags(status)); \
\
if (! i->isZeroMasking()) { \
for (unsigned n=0; n < len; n++, mask >>= 4) \
xmm_blendps(&BX_READ_AVX_REG_LANE(i->dst(), n), &op1.vmm128(n), mask); \
\
BX_CLEAR_AVX_REGZ(i->dst(), len); \
} \
else { \
BX_WRITE_AVX_REGZ(i->dst(), op1, len); \
} \
\
BX_NEXT_INSTR(i); \
}
EVEX_FMA_PACKED_SINGLE(VFMADDPS_MASK_VpsHpsWpsR, xmm_fmaddps_mask)
EVEX_FMA_PACKED_SINGLE(VFMADDSUBPS_MASK_VpsHpsWpsR, xmm_fmaddsubps_mask)
EVEX_FMA_PACKED_SINGLE(VFMSUBADDPS_MASK_VpsHpsWpsR, xmm_fmsubaddps_mask)
EVEX_FMA_PACKED_SINGLE(VFMSUBPS_MASK_VpsHpsWpsR, xmm_fmsubps_mask)
EVEX_FMA_PACKED_SINGLE(VFNMADDPS_MASK_VpsHpsWpsR, xmm_fnmaddps_mask)
EVEX_FMA_PACKED_SINGLE(VFNMSUBPS_MASK_VpsHpsWpsR, xmm_fnmsubps_mask)
#define EVEX_FMA_PACKED_DOUBLE(HANDLER, func) \
BX_INSF_TYPE BX_CPP_AttrRegparmN(1) BX_CPU_C:: HANDLER (bxInstruction_c *i) \
{ \
BxPackedAvxRegister op1 = BX_READ_AVX_REG(i->src1()); \
BxPackedAvxRegister op2 = BX_READ_AVX_REG(i->src2()); \
BxPackedAvxRegister op3 = BX_READ_AVX_REG(i->src3()); \
unsigned mask = BX_READ_8BIT_OPMASK(i->opmask()); \
unsigned len = i->getVL(); \
\
float_status_t status = mxcsr_to_softfloat_status_word(MXCSR); \
softfloat_status_word_rc_override(status, i); \
\
for (unsigned n=0, tmp_mask = mask; n < len; n++, tmp_mask >>= 2) \
(func)(&op1.vmm128(n), &op2.vmm128(n), &op3.vmm128(n), status, tmp_mask); \
\
check_exceptionsSSE(get_exception_flags(status)); \
\
if (! i->isZeroMasking()) { \
for (unsigned n=0; n < len; n++, mask >>= 2) \
xmm_blendpd(&BX_READ_AVX_REG_LANE(i->dst(), n), &op1.vmm128(n), mask); \
\
BX_CLEAR_AVX_REGZ(i->dst(), len); \
} \
else { \
BX_WRITE_AVX_REGZ(i->dst(), op1, len); \
} \
\
BX_NEXT_INSTR(i); \
}
EVEX_FMA_PACKED_DOUBLE(VFMADDPD_MASK_VpdHpdWpdR, xmm_fmaddpd_mask)
EVEX_FMA_PACKED_DOUBLE(VFMADDSUBPD_MASK_VpdHpdWpdR, xmm_fmaddsubpd_mask)
EVEX_FMA_PACKED_DOUBLE(VFMSUBADDPD_MASK_VpdHpdWpdR, xmm_fmsubaddpd_mask)
EVEX_FMA_PACKED_DOUBLE(VFMSUBPD_MASK_VpdHpdWpdR, xmm_fmsubpd_mask)
EVEX_FMA_PACKED_DOUBLE(VFNMADDPD_MASK_VpdHpdWpdR, xmm_fnmaddpd_mask)
EVEX_FMA_PACKED_DOUBLE(VFNMSUBPD_MASK_VpdHpdWpdR, xmm_fnmsubpd_mask)
#define EVEX_FMA_SCALAR_SINGLE(HANDLER, func) \
BX_INSF_TYPE BX_CPP_AttrRegparmN(1) BX_CPU_C:: HANDLER (bxInstruction_c *i) \
{ \
if (BX_SCALAR_ELEMENT_MASK(i->opmask())) { \
float32 op1 = BX_READ_XMM_REG_LO_DWORD(i->src1()); \
float32 op2 = BX_READ_XMM_REG_LO_DWORD(i->src2()); \
float32 op3 = BX_READ_XMM_REG_LO_DWORD(i->src3()); \
\
float_status_t status = mxcsr_to_softfloat_status_word(MXCSR); \
softfloat_status_word_rc_override(status, i); \
op1 = (func)(op1, op2, op3, status); \
check_exceptionsSSE(get_exception_flags(status)); \
\
BX_WRITE_XMM_REG_LO_DWORD(i->dst(), op1); \
} \
else { \
if (i->isZeroMasking()) \
BX_WRITE_XMM_REG_LO_DWORD(i->dst(), 0); \
} \
\
BX_CLEAR_AVX_HIGH128(i->dst()); \
BX_NEXT_INSTR(i); \
}
EVEX_FMA_SCALAR_SINGLE(VFMADDSS_MASK_VpsHssWssR, float32_fmadd)
EVEX_FMA_SCALAR_SINGLE(VFMSUBSS_MASK_VpsHssWssR, float32_fmsub)
EVEX_FMA_SCALAR_SINGLE(VFNMADDSS_MASK_VpsHssWssR, float32_fnmadd)
EVEX_FMA_SCALAR_SINGLE(VFNMSUBSS_MASK_VpsHssWssR, float32_fnmsub)
#define EVEX_FMA_SCALAR_DOUBLE(HANDLER, func) \
BX_INSF_TYPE BX_CPP_AttrRegparmN(1) BX_CPU_C:: HANDLER (bxInstruction_c *i) \
{ \
if (BX_SCALAR_ELEMENT_MASK(i->opmask())) { \
float64 op1 = BX_READ_XMM_REG_LO_QWORD(i->src1()); \
float64 op2 = BX_READ_XMM_REG_LO_QWORD(i->src2()); \
float64 op3 = BX_READ_XMM_REG_LO_QWORD(i->src3()); \
\
float_status_t status = mxcsr_to_softfloat_status_word(MXCSR); \
softfloat_status_word_rc_override(status, i); \
op1 = (func)(op1, op2, op3, status); \
check_exceptionsSSE(get_exception_flags(status)); \
\
BX_WRITE_XMM_REG_LO_QWORD(i->dst(), op1); \
} \
else { \
if (i->isZeroMasking()) \
BX_WRITE_XMM_REG_LO_QWORD(i->dst(), 0); \
} \
\
BX_CLEAR_AVX_HIGH128(i->dst()); \
BX_NEXT_INSTR(i); \
}
EVEX_FMA_SCALAR_DOUBLE(VFMADDSD_MASK_VpdHsdWsdR, float64_fmadd)
EVEX_FMA_SCALAR_DOUBLE(VFMSUBSD_MASK_VpdHsdWsdR, float64_fmsub)
EVEX_FMA_SCALAR_DOUBLE(VFNMADDSD_MASK_VpdHsdWsdR, float64_fnmadd)
EVEX_FMA_SCALAR_DOUBLE(VFNMSUBSD_MASK_VpdHsdWsdR, float64_fnmsub)
#endif