ShuriZma
/
dolphin
mirror of https://github.com/dolphin-emu/dolphin.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Did clean up of indirect texture look up and fixed some bugs related to indirect unit. 

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@2954 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
donkopunchstania 2009-04-11 07:05:57 +00:00
parent c43b1c18e5
commit 2b2c2afa3c
2 changed files with 24 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
Source/Core/VideoCommon/Src/PixelShaderGen.cpp
View File

@ -465,10 +465,6 @@ const char *GeneratePixelShader(u32 texture_mask, bool has_zbuffer_target, bool
"float3 tevcoord;\n" "float3 tevcoord;\n"
"float2 wrappedcoord, tempcoord;\n"); "float2 wrappedcoord, tempcoord;\n");
//if (bOutputZ ) WRITE(p, " float depth;\n");
// WRITE(p, "return 1;}\n");
// return PixelShaderMngr::CompilePixelShader(ps, text);
// indirect texture map lookup // indirect texture map lookup
for(u32 i = 0; i < bpmem.genMode.numindstages; ++i) { for(u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
if (nIndirectStagesUsed & (1<<i)) { if (nIndirectStagesUsed & (1<<i)) {
@ -478,21 +474,18 @@ const char *GeneratePixelShader(u32 texture_mask, bool has_zbuffer_target, bool
int texcoord = bpmem.tevindref.getTexCoord(i); int texcoord = bpmem.tevindref.getTexCoord(i);
if (texcoord < numTexgen) { if (texcoord < numTexgen) {
if (texture_mask & (1<<bpmem.tevindref.getTexMap(i))) { WRITE(p, "tempcoord=uv%d.xy * "I_INDTEXSCALE"[%d].%s;\n", texcoord, i/2, (i&1)?"zw":"xy");
// TODO: I removed a superfluous argument, please check that the resulting expression is correct. (mthuurne 2008-08-27) }
WRITE(p, "float2 induv%d=uv%d.xy * "I_INDTEXSCALE"[%d].%s;\n", i, texcoord, i/2, (i&1)?"zw":"xy"); //, bpmem.tevindref.getTexMap(i) else {
WRITE(p, "tempcoord=float2(0.0f,0.0f);\n");
}
char str[16]; if (texture_mask & (1<<bpmem.tevindref.getTexMap(i))) {
sprintf(str, "induv%d", i); WrapNonPow2Tex(p, "tempcoord", bpmem.tevindref.getTexMap(i), texture_mask);
WrapNonPow2Tex(p, str, bpmem.tevindref.getTexMap(i), texture_mask); WRITE(p, "float3 indtex%d=texRECT(samp%d,tempcoord.xy).abg;\n", i, bpmem.tevindref.getTexMap(i));
WRITE(p, "float3 indtex%d=texRECT(samp%d,induv%d.xy).abg;\n", i, bpmem.tevindref.getTexMap(i), i);
} }
else { else {
WRITE(p, "float3 indtex%d=tex2D(samp%d,uv%d.xy*"I_INDTEXSCALE"[%d].%s).abg;\n", i, bpmem.tevindref.getTexMap(i), texcoord, i/2, (i&1)?"zw":"xy"); WRITE(p, "float3 indtex%d=tex2D(samp%d,tempcoord).abg;\n", i, bpmem.tevindref.getTexMap(i));
}
}
else {
WRITE(p, "float3 indtex%d=tex2D(samp%d,float2(0.0f,0.0f));\n", i, bpmem.tevindref.getTexMap(i));
} }
} }
} }
@ -586,7 +579,6 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
if (bHasIndStage) { if (bHasIndStage) {
// perform the indirect op on the incoming regular coordinates using indtex%d as the offset coords // perform the indirect op on the incoming regular coordinates using indtex%d as the offset coords
bHasIndStage = true;
int texmap = bpmem.tevorders[n/2].getEnable(n&1) ? bpmem.tevorders[n/2].getTexMap(n&1) : bpmem.tevindref.getTexMap(bpmem.tevind[n].bt); int texmap = bpmem.tevorders[n/2].getEnable(n&1) ? bpmem.tevorders[n/2].getTexMap(n&1) : bpmem.tevindref.getTexMap(bpmem.tevind[n].bt);
if (bpmem.tevind[n].bs != ITBA_OFF) { if (bpmem.tevind[n].bs != ITBA_OFF) {
@ -622,11 +614,11 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
WRITE(p, "float2 indtevtrans%d = float2(dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d), dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d));\n", WRITE(p, "float2 indtevtrans%d = float2(dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d), dot("I_INDTEXMTX"[%d].xyz, indtevcrd%d));\n",
n, mtxidx, n, mtxidx+1, n); n, mtxidx, n, mtxidx+1, n);
} }
else if (bpmem.tevind[n].mid <= 5 && bHasTexCoord) { // s matrix else if (bpmem.tevind[n].mid <= 7 && bHasTexCoord) { // s matrix
int mtxidx = 2*(bpmem.tevind[n].mid-5); int mtxidx = 2*(bpmem.tevind[n].mid-5);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.xx;\n", n, mtxidx, texcoord, n); WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.xx;\n", n, mtxidx, texcoord, n);
} }
else if (bpmem.tevind[n].mid <= 9 && bHasTexCoord) { // t matrix else if (bpmem.tevind[n].mid <= 11 && bHasTexCoord) { // t matrix
int mtxidx = 2*(bpmem.tevind[n].mid-9); int mtxidx = 2*(bpmem.tevind[n].mid-9);
WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.yy;\n", n, mtxidx, texcoord, n); WRITE(p, "float2 indtevtrans%d = "I_INDTEXMTX"[%d].ww * uv%d.xy * indtevcrd%d.yy;\n", n, mtxidx, texcoord, n);
} }
@ -645,8 +637,8 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
// non pow2 // non pow2
if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) { if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) {
if (bpmem.tevind[n].sw == ITW_0) {
if (bpmem.tevind[n].tw == ITW_0) { if (bpmem.tevind[n].tw == ITW_0) {
if (bpmem.tevind[n].sw == ITW_0) {
// zero out completely // zero out completely
WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n"); WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n");
} }
@ -655,7 +647,7 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
"wrappedcoord.y = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].sw]); "wrappedcoord.y = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].sw]);
} }
} }
else if (bpmem.tevind[n].tw == ITW_0) { else if (bpmem.tevind[n].sw == ITW_0) {
WRITE(p, "wrappedcoord.y = fmod( (uv%d.y+%s)*"I_TEXDIMS"[%d].y*"I_TEXDIMS"[%d].w, %s);\n" WRITE(p, "wrappedcoord.y = fmod( (uv%d.y+%s)*"I_TEXDIMS"[%d].y*"I_TEXDIMS"[%d].w, %s);\n"
"wrappedcoord.x = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].tw]); "wrappedcoord.x = 0;\n", texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, texmap, tevIndWrapStart[bpmem.tevind[n].tw]);
} }
@ -675,8 +667,8 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
// mult by bitdepth / tex dimensions // mult by bitdepth / tex dimensions
if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) { if (bpmem.tevind[n].sw != ITW_OFF || bpmem.tevind[n].tw != ITW_OFF) {
if (bpmem.tevind[n].sw == ITW_0) {
if (bpmem.tevind[n].tw == ITW_0) { if (bpmem.tevind[n].tw == ITW_0) {
if (bpmem.tevind[n].sw == ITW_0) {
// zero out completely // zero out completely
WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n"); WRITE(p, "wrappedcoord = float2(0.0f,0.0f);\n");
} }
@ -685,7 +677,7 @@ static void WriteStage(char *&p, int n, u32 texture_mask)
"wrappedcoord.y = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, tevIndWrapStart[bpmem.tevind[n].sw]); "wrappedcoord.y = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].sw], texmap, tevIndWrapStart[bpmem.tevind[n].sw]);
} }
} }
else if (bpmem.tevind[n].tw == ITW_0) { else if (bpmem.tevind[n].sw == ITW_0) {
WRITE(p, "wrappedcoord.y = "I_TEXDIMS"[%d].y * fmod( uv%d.y+%s, "I_TEXDIMS"[%d].w*%s);\n" WRITE(p, "wrappedcoord.y = "I_TEXDIMS"[%d].y * fmod( uv%d.y+%s, "I_TEXDIMS"[%d].w*%s);\n"
"wrappedcoord.x = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, tevIndWrapStart[bpmem.tevind[n].tw]); "wrappedcoord.x = 0;\n", texmap, texcoord, tevIndWrapStart[bpmem.tevind[n].tw], texmap, tevIndWrapStart[bpmem.tevind[n].tw]);
} }

11
Source/Core/VideoCommon/Src/PixelShaderGen.h
View File

@ -40,28 +40,29 @@
#define C_ENVCONST_END (C_FOG+2) #define C_ENVCONST_END (C_FOG+2)
#define C_COLORMATRIX (C_FOG+2) #define C_COLORMATRIX (C_FOG+2)
#define PIXELSHADERUID_MAX_VALUES (5+32+6+11)
class PIXELSHADERUID class PIXELSHADERUID
{ {
public: public:
u32 values[5+32+6+11]; u32 values[PIXELSHADERUID_MAX_VALUES];
u16 tevstages, indstages; u16 tevstages, indstages;
PIXELSHADERUID() { PIXELSHADERUID() {
memset(values, 0, (5+32+6+11) * 4); memset(values, 0, PIXELSHADERUID_MAX_VALUES * 4);
tevstages = indstages = 0; tevstages = indstages = 0;
} }
PIXELSHADERUID(const PIXELSHADERUID& r) PIXELSHADERUID(const PIXELSHADERUID& r)
{ {
tevstages = r.tevstages; tevstages = r.tevstages;
indstages = r.indstages; indstages = r.indstages;
int N = tevstages + indstages + 4; int N = GetNumValues();
_assert_(N <= 5+32+6+11); _assert_(N <= PIXELSHADERUID_MAX_VALUES);
for (int i = 0; i < N; ++i) for (int i = 0; i < N; ++i)
values[i] = r.values[i]; values[i] = r.values[i];
} }
int GetNumValues() const { int GetNumValues() const {
return tevstages + indstages + 3; // numTevStages*3/2+1 return tevstages + indstages + 4;
} }
bool operator <(const PIXELSHADERUID& _Right) const bool operator <(const PIXELSHADERUID& _Right) const
{ {