zdoom/qzdoom
0
0
Fork 0
mirror of https://github.com/ZDoom/qzdoom.git synced 2024-11-10 14:51:51 +00:00
Code Issues Projects Releases Packages Wiki Activity
qzdoom/src/r_draw.h
Randy Heit dda5ddd3c2 - Ported vlinetallasm4 to AMD64 assembly. Even with the increased number of 
registers AMD64 provides, this routine still needs to be written as self-
  modifying code for maximum performance. The additional registers do allow
  for further optimization over the x86 version by allowing all four pixels
  to be in flight at the same time. The end result is that AMD64 ASM is about
  2.18 times faster than AMD64 C and about 1.06 times faster than x86 ASM.
  (For further comparison, AMD64 C and x86 C are practically the same for
  this function.) Should I port any more assembly to AMD64, mvlineasm4 is the
  most likely candidate, but it's not used enough at this point to bother.
  Also, this may or may not work with Linux at the moment, since it doesn't
  have the eh_handler metadata. Win64 is easier, since I just need to
  structure the function prologue and epilogue properly and use some
  assembler directives/macros to automatically generate the metadata. And
  that brings up another point: You need YASM to assemble the AMD64 code,
  because NASM doesn't support the Win64 metadata directives.
- Added an SSE version of DoBlending. This is strictly C intrinsics.
  VC++ still throws around unneccessary register moves. GCC seems to be
  pretty close to optimal, requiring only about 2 cycles/color. They're
  both faster than my hand-written MMX routine, so I don't need to feel
  bad about not hand-optimizing this for x64 builds.
- Removed an extra instruction from DoBlending_MMX, transposed two
  instructions, and unrolled it once, shaving off about 80 cycles from the
  time required to blend 256 palette entries. Why? Because I tried writing
  a C version of the routine using compiler intrinsics and was appalled by
  all the extra movq's VC++ added to the code. GCC was better, but still
  generated extra instructions. I only wanted a C version because I can't
  use inline assembly with VC++'s x64 compiler, and x64 assembly is a bit
  of a pain. (It's a pain because Linux and Windows have different calling
  conventions, and you need to maintain extra metadata for functions.) So,
  the assembly version stays and the C version stays out.
- Removed all the pixel doubling r_detail modes, since the one platform they
  were intended to assist (486) actually sees very little benefit from them.
- Rewrote CheckMMX in C and renamed it to CheckCPU.
- Fixed: CPUID function 0x80000005 is specified to return detailed L1 cache
  only for AMD processors, so we must not use it on other architectures, or
  we end up overwriting the L1 cache line size with 0 or some other number
  we don't actually understand.


SVN r1134 (trunk)
2008-08-09 03:13:43 +00:00

277 lines
9.1 KiB
C++
Raw Blame History

// Emacs style mode select -*- C++ -*-
//-----------------------------------------------------------------------------
//
// $Id:$
//
// Copyright (C) 1993-1996 by id Software, Inc.
//
// This source is available for distribution and/or modification
// only under the terms of the DOOM Source Code License as
// published by id Software. All rights reserved.
//
// The source is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// FITNESS FOR A PARTICULAR PURPOSE. See the DOOM Source Code License
// for more details.
//
// DESCRIPTION:
// System specific interface stuff.
//
//-----------------------------------------------------------------------------
#ifndef __R_DRAW__
#define __R_DRAW__
#include "r_data.h"
extern "C" int ylookup[MAXHEIGHT];
extern "C" int dc_pitch; // [RH] Distance between rows
extern "C" lighttable_t*dc_colormap;
extern "C" int dc_x;
extern "C" int dc_yl;
extern "C" int dc_yh;
extern "C" fixed_t dc_iscale;
extern "C" fixed_t dc_texturemid;
extern "C" fixed_t dc_texturefrac;
extern "C" int dc_color; // [RH] For flat colors (no texturing)
extern "C" DWORD dc_srccolor;
extern "C" DWORD *dc_srcblend;
extern "C" DWORD *dc_destblend;
// first pixel in a column
extern "C" const BYTE* dc_source;
extern "C" BYTE *dc_dest, *dc_destorg;
extern "C" int dc_count;
extern "C" DWORD vplce[4];
extern "C" DWORD vince[4];
extern "C" BYTE* palookupoffse[4];
extern "C" const BYTE* bufplce[4];
// [RH] Temporary buffer for column drawing
extern "C" BYTE dc_temp[MAXHEIGHT*4];
extern "C" unsigned int dc_tspans[4][MAXHEIGHT];
extern "C" unsigned int *dc_ctspan[4];
extern "C" unsigned int horizspans[4];
// [RH] Pointers to the different column and span drawers...
// The span blitting interface.
// Hook in assembler or system specific BLT here.
extern void (*R_DrawColumn)(void);
extern DWORD (STACK_ARGS *dovline1) ();
extern DWORD (STACK_ARGS *doprevline1) ();
#ifdef X64_ASM
#define dovline4 vlinetallasm4
extern "C" void vlinetallasm4();
#else
extern void (STACK_ARGS *dovline4) ();
#endif
extern void setupvline (int);
extern DWORD (STACK_ARGS *domvline1) ();
extern void (STACK_ARGS *domvline4) ();
extern void setupmvline (int);
extern void setuptmvline (int);
// The Spectre/Invisibility effect.
extern void (*R_DrawFuzzColumn)(void);
// [RH] Draw shaded column
extern void (*R_DrawShadedColumn)(void);
// Draw with color translation tables, for player sprite rendering,
// Green/Red/Blue/Indigo shirts.
extern void (*R_DrawTranslatedColumn)(void);
// Span drawing for rows, floor/ceiling. No Sepctre effect needed.
extern void (*R_DrawSpan)(void);
// Span drawing for masked textures.
extern void (*R_DrawSpanMasked)(void);
// Span drawing for translucent textures.
extern void (*R_DrawSpanTranslucent)(void);
// Span drawing for masked, translucent textures.
extern void (*R_DrawSpanMaskedTranslucent)(void);
// [RH] Span blit into an interleaved intermediate buffer
extern void (*R_DrawColumnHoriz)(void);
void R_DrawMaskedColumnHoriz (const BYTE *column, const FTexture::Span *spans);
// [RH] Initialize the above pointers
void R_InitColumnDrawers ();
// [RH] Moves data from the temporary buffer to the screen.
extern "C"
{
void rt_copy1col_c (int hx, int sx, int yl, int yh);
void STACK_ARGS rt_copy4cols_c (int sx, int yl, int yh);
void rt_shaded1col (int hx, int sx, int yl, int yh);
void STACK_ARGS rt_shaded4cols_c (int sx, int yl, int yh);
void STACK_ARGS rt_shaded4cols_asm (int sx, int yl, int yh);
void rt_map1col_c (int hx, int sx, int yl, int yh);
void rt_add1col (int hx, int sx, int yl, int yh);
void rt_addclamp1col (int hx, int sx, int yl, int yh);
void rt_subclamp1col (int hx, int sx, int yl, int yh);
void rt_revsubclamp1col (int hx, int sx, int yl, int yh);
void rt_tlate1col (int hx, int sx, int yl, int yh);
void rt_tlateadd1col (int hx, int sx, int yl, int yh);
void rt_tlateaddclamp1col (int hx, int sx, int yl, int yh);
void rt_tlatesubclamp1col (int hx, int sx, int yl, int yh);
void rt_tlaterevsubclamp1col (int hx, int sx, int yl, int yh);
void STACK_ARGS rt_map4cols_c (int sx, int yl, int yh);
void STACK_ARGS rt_add4cols_c (int sx, int yl, int yh);
void STACK_ARGS rt_addclamp4cols_c (int sx, int yl, int yh);
void STACK_ARGS rt_subclamp4cols (int sx, int yl, int yh);
void STACK_ARGS rt_revsubclamp4cols (int sx, int yl, int yh);
void STACK_ARGS rt_tlate4cols (int sx, int yl, int yh);
void STACK_ARGS rt_tlateadd4cols (int sx, int yl, int yh);
void STACK_ARGS rt_tlateaddclamp4cols (int sx, int yl, int yh);
void STACK_ARGS rt_tlatesubclamp4cols (int sx, int yl, int yh);
void STACK_ARGS rt_tlaterevsubclamp4cols (int sx, int yl, int yh);
void rt_copy1col_asm (int hx, int sx, int yl, int yh);
void rt_map1col_asm (int hx, int sx, int yl, int yh);
void STACK_ARGS rt_copy4cols_asm (int sx, int yl, int yh);
void STACK_ARGS rt_map4cols_asm1 (int sx, int yl, int yh);
void STACK_ARGS rt_map4cols_asm2 (int sx, int yl, int yh);
void STACK_ARGS rt_add4cols_asm (int sx, int yl, int yh);
void STACK_ARGS rt_addclamp4cols_asm (int sx, int yl, int yh);
}
extern void (STACK_ARGS *rt_map4cols)(int sx, int yl, int yh);
#ifdef X86_ASM
#define rt_copy1col rt_copy1col_asm
#define rt_copy4cols rt_copy4cols_asm
#define rt_map1col rt_map1col_asm
#define rt_shaded4cols rt_shaded4cols_asm
#define rt_add4cols rt_add4cols_asm
#define rt_addclamp4cols rt_addclamp4cols_asm
#else
#define rt_copy1col rt_copy1col_c
#define rt_copy4cols rt_copy4cols_c
#define rt_map1col rt_map1col_c
#define rt_shaded4cols rt_shaded4cols_c
#define rt_add4cols rt_add4cols_c
#define rt_addclamp4cols rt_addclamp4cols_c
#endif
void rt_draw4cols (int sx);
// [RH] Preps the temporary horizontal buffer.
void rt_initcols (void);
void R_DrawFogBoundary (int x1, int x2, short *uclip, short *dclip);
#ifdef X86_ASM
extern "C" void R_DrawColumnP_Unrolled (void);
extern "C" void R_DrawColumnHorizP_ASM (void);
extern "C" void R_DrawColumnP_ASM (void);
extern "C" void R_DrawFuzzColumnP_ASM (void);
void R_DrawTranslatedColumnP_C (void);
void R_DrawShadedColumnP_C (void);
extern "C" void R_DrawSpanP_ASM (void);
extern "C" void R_DrawSpanMaskedP_ASM (void);
#else
void R_DrawColumnHorizP_C (void);
void R_DrawColumnP_C (void);
void R_DrawFuzzColumnP_C (void);
void R_DrawTranslatedColumnP_C (void);
void R_DrawShadedColumnP_C (void);
void R_DrawSpanP_C (void);
void R_DrawSpanMaskedP_C (void);
#endif
void R_DrawSpanTranslucentP_C (void);
void R_DrawSpanMaskedTranslucentP_C (void);
void R_DrawTlatedLucentColumnP_C (void);
#define R_DrawTlatedLucentColumn R_DrawTlatedLucentColumnP_C
void R_FillColumnP (void);
void R_FillColumnHorizP (void);
void R_FillSpan (void);
extern "C" int ds_y;
extern "C" int ds_x1;
extern "C" int ds_x2;
extern "C" lighttable_t* ds_colormap;
extern "C" dsfixed_t ds_xfrac;
extern "C" dsfixed_t ds_yfrac;
extern "C" dsfixed_t ds_xstep;
extern "C" dsfixed_t ds_ystep;
extern "C" int ds_xbits;
extern "C" int ds_ybits;
extern "C" fixed_t ds_alpha;
// start of a 64*64 tile image
extern "C" const BYTE* ds_source;
extern "C" int ds_color; // [RH] For flat color (no texturing)
extern BYTE shadetables[NUMCOLORMAPS*16*256];
extern FDynamicColormap ShadeFakeColormap[16];
extern BYTE identitymap[256];
extern BYTE *dc_translation;
// If the view size is not full screen, draws a border around it.
void R_DrawViewBorder (void);
void R_DrawTopBorder (void);
void R_DrawBorder (int x1, int y1, int x2, int y2);
// [RH] Added for muliresolution support
void R_InitFuzzTable (int fuzzoff);
// [RH] Consolidate column drawer selection
enum ESPSResult
{
DontDraw, // not useful to draw this
DoDraw0, // draw this as if r_columnmethod is 0
DoDraw1, // draw this as if r_columnmethod is 1
};
ESPSResult R_SetPatchStyle (FRenderStyle style, fixed_t alpha, int translation, DWORD color);
// Call this after finished drawing the current thing, in case its
// style was STYLE_Shade
void R_FinishSetPatchStyle ();
// transmaskwallscan calls this to find out what column drawers to use
bool R_GetTransMaskDrawers (fixed_t (**tmvline1)(), void (**tmvline4)());
// Retrieve column data for wallscan. Should probably be removed
// to just use the texture's GetColumn() method. It just exists
// for double-layer skies.
const BYTE *R_GetColumn (FTexture *tex, int col);
void wallscan (int x1, int x2, short *uwal, short *dwal, fixed_t *swal, fixed_t *lwal, const BYTE *(*getcol)(FTexture *tex, int col)=R_GetColumn);
// maskwallscan is exactly like wallscan but does not draw anything where the texture is color 0.
void maskwallscan (int x1, int x2, short *uwal, short *dwal, fixed_t *swal, fixed_t *lwal, const BYTE *(*getcol)(FTexture *tex, int col)=R_GetColumn);
// transmaskwallscan is like maskwallscan, but it can also blend to the background
void transmaskwallscan (int x1, int x2, short *uwal, short *dwal, fixed_t *swal, fixed_t *lwal, const BYTE *(*getcol)(FTexture *tex, int col)=R_GetColumn);
#endif
Reference in a new issue View git blame Copy permalink